BR102017006400A2 - METHOD AND APPARATUS FOR COMPENSATION FOR LOSS OF FRAMEWORK - Google Patents

Description

Report of the Invention Patent for "FRAMEWORK COMPENSATION PROCESSING METHOD AND APPARATUS".

TECHNICAL FIELD

[001] Embodiments of the present invention relate to communication technologies and in particular to a frame loss compensation processing method and apparatus. BACKGROUND

[002] In a voice service, problems such as loss of a voice packet and error in a voice packet often occur in a low coverage scenario, an interfering scenario, and a high speed moving scenario. This inevitably causes an unsatisfactory user experience due to midrange, noise or the like.

An existing method of frame loss compensation is as follows: A data flow analysis is performed on a decoder to verify that a current frame is a lost frame. If the current frame is a lost frame, a lost current frame parameter is estimated, a frequency spectrum parameter and a lost frame excitation signal are retrieved according to the current lost frame parameter and a one frame parameter. history, and a lost frame signal is also obtained according to the frequency spectrum parameter and the excitation signal; or if the current frame is a normal frame, a parameter of the current frame is obtained by decoding, and if the current frame is a normal frame and a previous frame is a lost frame, the current frame parameter is corrected according to a previous frame parameter, a frequency spectrum parameter, and an excitation signal for the current frame are obtained according to a corrected parameter, and a current frame signal is synthesized according to the frequency spectrum parameter and the signal. of excitement. The foregoing parameter includes at least one of the parameters, such as a signal type, signal energy, and signal phase.

Because the lost frame parameter is not precisely estimated with the previous method, the quality of audio decoding cannot be guaranteed.

SUMMARY

[005] The embodiments of the present invention provide a frame loss compensation processing method and apparatus for improving the parameter estimation accuracy of a lost frame and improving the signal decoding quality.

A first aspect of the present invention provides a frame loss compensation processing method. First, a check is made to check if a frame is a lost frame using a lost frame flag bit. When the ith frame is a lost frame, a frequency spectrum parameter, a pitch period, and a gain from the ith frame are estimated according to at least one of an inter-frame relationship between the first N frames of frame i. It is either the nth frame or an intraframe relationship between the first N frames of the nth frame. An algebraic codebook of the nth frame is obtained. An excitation signal from the nth frame is generated according to the pitch period and the gain for the nth frame which is obtained by estimation and the algebraic codebook obtained from the nth frame. A signal from the ith frame is also synthesized according to the frequency spectrum parameter referring to the ith frame and which is obtained by estimation and the generated excitation signal from the frame. The interframe relationship between the first N frames includes at least one of the correlation between the first N frames or the energy stability between the first N frames, and the intra-frame relationship between the first N frames includes at least one of the correlation between the first N frames. intersubquads between the first N frames or the energy stability of the intersubquads between the first N frames. The correlation between the signals and the energy stability between the signals is considered, so as to obtain a more accurate nth frame parameter. by estimating and improving the quality of voice signal decoding.

In a possible way of implementing the first aspect, the frequency spectrum parameter of the nth frame is obtained by estimating according to the inter-frame relationship between the first N frames of the nth frame and can be Specifically obtained by estimation as follows: First, determine the weight of a frequency spectrum parameter of a nth frame (i-1) and the weight of a predefined frequency spectrum parameter of the nth frame with the correlation between the first frames N of table i, and then perform a weighting operation on the frequency spectrum parameter of table (i-1) and the default frequency spectrum parameter of table i nth according to the weight of the frequency spectrum parameter of table (i-1) 'nth and the weight of the preset frequency spectrum parameter of table Γ es, mo, to obtain the frequency spectrum parameter ia of the frame is very good.

When the correlation between the first N frames of the nth frame includes a value relationship between a second boundary and a spectrum slope parameter of a nth frame (i-1) signal, a value relationship between a first bound and a normalized frame signal (i-1) "nth value" and a value relationship between a third boundary and a pitch (i-1) signal period deviation from the nth frame, determining the weight of a frequency spectrum parameter of a nth frame (i-1) and the weight of a predefined frequency spectrum parameter of the nth frame according to the correlation between the first N frames of table í specifically means: if the signal from frame (i-1) "es" corresponds to at least one of a first condition, a second condition, and a third condition, determine that the weight of the frequency spectrum parameter of (i-1) ~ this frame is a first weight and the The weight of the default frequency spectrum parameter of the ith frame is a second weight, when the first weight is greater than the second weight, the first condition is: the normalized autocorrelation value of the frame (i-1) signal is greater than the first limit, the second condition is: the signal slope parameter of frame (i-1) 's is higher than the second limit, and the third condition is: the deviation of the signal's pitch period of table (i-1) "th is less than the third limit; or if the signal from table (jl) "nth does not correspond to a first condition, a second condition, or a third condition, determines that the weight of the frequency spectrum parameter of (i-1) qUacjro is a second weight, and the weight of the preset frequency spectrum parameter of table i is a first weight.

[009] In a possible way of implementing the first aspect, the pitch period of the nth frame is obtained by estimation according to the correlation between the first N frames of the nth frame and the correlation of intersubquads between the two. first frames N of the summary frame the correlation includes a relation of values between a fifth bound and a normalized autocorrelation value of a signal of a (i-2) -th frame, a relation of values between a fourth bound and a deviation of one. pitch pitch of the (i-2) -th frame signal and a relationship of values between the fourth boundary and a deviation of a pitch period from a signal of one (i-1) 'the nth frame. Correspondingly, the pitch period of the nth frame is obtained by estimation as follows: if the deviation of the pitch period of the nth frame (i-1) signal is less than the fourth threshold, determine a deviation value of the frame signal pitch period (i-1) 'nth according to the frame signal pitch period (i-1)' nth and determine a frame signal pitch period i 'nth of according to the deviation value of the frame signal pitch period (i-1) -th and the frame signal pitch period (i-1) -th, where the frame signal pitch period of the nth includes a pitch period of each frame subframe i and the value of the frame signal pitch period deviation (i-1) is an average value of the differences between the pitch periods of all adjacent subframes of the frame. table (i-1) "nth; or if the deviation of the pitch (i-1) 's signal pitch period is greater than or equal to the fourth threshold, the normalized autocorrelation value of the (i-2)' s signal is greater than the fifth threshold and If the pitch offset of the (i-2) -th frame signal is less than the fourth threshold, determine a shift value of the pitch signal period of the (i-2) -th frame and the frame signal ( i-1) - nth according to the frame signal pitch period (i-2) "nth and the frame signal pitch period (i-1)" nth, and determine a pitch signal period of the frame i "esimode according to the pitch pitch of the nth frame (i-1) and the deviation value of the pitch signal period of the nth frame (i-2) and the frame signal (i-1 ) 'esimo.

In one embodiment, the pitch period shift value Pv of the frame (i-1) signal can be determined according to the following formula: where p (is a period of pitch of a subframe j "tenth of frame (i-1)" esimo ej — 0, 1, 2, 3. Correspondingly, the pitch period of the signal of the i'th frame is determined according to the following formula: [ 0012], where p (is a pitch period of a third frame subframe (i-1) "nth, pv is the deviation of the frame signal pitch period (i-1, and PcuÁJ) is a pitch period of a subframe already within the nth frame.

In another embodiment, the deviation value of the pitch signal period pv from frame (i-2) 'eso and frame signal (i-1)' eso may be determined according to the following formula : Where p (ί (4Ώ) is a pitch period of a nth subframe of frame (i-2) -this ^ p {\ n) ^ a pitch period of a nth subframe from table (i-1) 'th, m = 2, 3 and n = 0, 1.

Correspondingly, the pitch period of the nth frame signal is determined according to the following formula: where p (is a pitch period of a third frame subframe (i-1)) -this, pv is the pitch period deviation value of the frame signal (i-2) -th and the frame signal (i-1) "esimo ie ρ ^ (χ) § A pitch period of a subframe * 'of the i'th frame.

In a possible way of implementing the first aspect, the gain of the first frame is obtained by estimation according to the correlation between the first N frames of the first frame and the energy stability between the first frames. N of the nth frame, and the nth frame gain includes an adaptive codebook gain and an algebraic codebook gain. Resume frame gain is specifically obtained by estimation as follows: first, determine the adaptive codebook gain of the ith frame according to a adaptive codebook gain of a frame (i-1) " or a predefined fixed value, the correlation of the frame (i-1) is the sequence number of the frame ies, in multiple consecutive lost frames; then determining the weight of a frame (i-1) algebraic codebook gain and the weight of a voice frame detection (VAD) frame gain according to the frame's energy stability (1-1) th; and finally, perform a weighting operation on the algebraic codebook gain of frame (i-1) "and the VAD frame gain according to the weight of the algebraic codebook gain (i-1) and the weight of the VAD frame gain to obtain the algebraic codebook gain of the ith frame. Optionally, more stable frame energy (i-1) indicates a greater weight of the (i-1) 'algebraic codebook, or the weight of the VAD frame gain increases correspondingly as the number of consecutive lost frames increases.

Optionally, before performing a weighting operation on the algebraic codebook gain of frame (i-1) "and the VAD frame gain according to the weight of the algebraic codebook gain of the Table (i-1) 's and the weight of the VAD frame gain, to obtain the algebraic codebook gain of frame i', a first correction factor can also be determined according to a coding rate. and decoding, and the gain of the algebraic codebook of the frame (i-1 is corrected using the first correction factor.

In a possible way of implementing the first aspect, the algebraic codebook of table i can be obtained as follows: by obtaining the algebraic codebook of table i by estimation according to random noise, or by determining the algebraic codebook of the nth frame according to the algebraic codebook of the first frames N of the nth frame.

[0020] In a possible way of implementing the first aspect, prior to generating an excitation signal from frame i "is according to the pitch period and gain referring to frame i" and which are obtained by estimation and From the algebraic codebook obtained from the nth frame, the weight of a contribution from the algebraic codebook from the nth frame must also be determined according to any of a pitch period deviation of a frame (M) nth, a correlation of a nth frame signal (i-1), a nth frame spectrum slope value value (i-1), or a nth frame crossover rate (M) , or the weight of a contribution from the algebraic codebook of the nth frame is determined by performing a weighting operation on any combination of a deviation from a pitch (i-1) pitch period, a correlation of a signal from frame (i-1) ~, a rate value of i inclination of frame spectrum (i-1) "nth or a zero crossing rate of frame (i-1)" nth. When the excitation signal from the ith frame is generated, the contribution of the ith frame algebraic codebook is first determined according to a product obtained by multiplying the ith frame algebraic codebook by the gain of the algebraic codebook of frame i; a contribution from the adaptive codebook of table i, m and i, is determined according to a product obtained by multiplying the adaptive codebook of the parent frame by the gain of the adaptive codebook of the nth frame, and then a weighting operation is performed on the contribution of the algebraic codebook of table i and the contribution of the adaptive codebook of table i according to the weight of the contribution of the algebraic codebook of table i and the weight of the contribution of the adaptive codebook of the ith frame to determine the excitation signal of the ith frame, where the weight of the adaptive codebook is 1.

[0021] In a possible way of implementing the first aspect »when the i th frame is a normal frame» the frequency spectrum parameter »the pitch period» the i th frame algebraic gain and codebook are obtained by decoding according to a received bitstream and then »the excitation signal from frame i" and an updated excitation signal from frame i "are generated according to the pitch period» o gain and the algebraic codebook referring to the nth frame and which are obtained by decoding. If a nth frame (i-1) or a nth frame (i-2) is a lost frame, the need to Correcting at least one of the frequency spectrum parameter, the excitation signal or the updated status excitation signal of Table I must also be determined according to at least one of the inter-frame relations or the intra-frame relations between and the nth frame and the first N frames of the nth frame. The interframe relationship includes at least one of the correlation between the nth frame and the first N frames of the frame, or the energy stability between the parent frame and the first N frames of the nth frame, and the ratio of Intraframe includes at least one of the intersubramer correlation between the nth frame and the first N frames of frame i or the energy stability of the intersubquads between the i frame and the first N frames of frame i .

[0022] When the need to correct at least one of the frequency spectrum parameter, the excitation signal or the updated status excitation signal of the twelfth frame is determined, at least one of the frequency spectrum parameter, the excitation signal or the updated status excitation signal of the ith frame is corrected according to at least one of the interframe relations or the intra-frame relations between the ith frame and the first N frames of the ith frame ; and the ith frame signal is synthesized according to a correction result of at least one of the frequency spectrum parameter, the excitation signal or the updated ith frame excitation signal. When the non-correction of the frequency spectrum parameter, the excitation signal or the updated status excitation signal of the nth frame is determined, the nth frame signal is synthesized according to the frequency spectrum parameter, the excitation signal and the updated status excitation signal from the nth frame.At least one of the frequency spectrum parameter, the excitation signal or the updated status excitation signal from the nth frame is corrected so that An easy transition of both global power between adjacent frames and power within the same frequency band can be deployed.

In a possible way of implementing the first aspect, the need to correct the frequency spectrum parameter of the ith frame can be determined according to the correlation of the ith frame. When the need for correction of the frequency spectrum parameter of the nth frame is determined, the frequency spectrum parameter of the nth frame is corrected according to the frequency spectrum parameter of the nth frame and a frequency parameter. (i-1) 's frequency frame, or the frequency spectrum parameter of the nth frame is corrected according to the frequency spectrum parameter of the nth frame and a predefined frequency spectrum parameter of the frame i "esimo. The correlation of the summary frame includes a ratio of values between a sixth limit and one of the two frequency spectrum parameters that corresponds to an index of a minimum value of the difference between adjacent frequency spectrum parameters of the nth frame, a ratio values between a seventh limit and the minimum value of the difference between the adjacent frequency spectrum parameters of table i and a ratio of values between an eighth limit and the minimum value index of the difference between the adjacent frequency spectrum parameters frame frequency is very high.

When the need to correct the frequency spectrum parameter of frame i is determined, the difference between the adjacent frequency spectrum parameters of frame i is first determined, where each difference corresponds to an index and the frequency spectrum parameter includes either an immitance spectral frequency (ISF) or a line spectral frequency (LSF), then whether the difference between the adjacent frequency spectrum parameters of table i corresponds to at least one of one fourth condition or a fifth condition is determined. The fourth condition includes; one of the two frequency spectrum parameters corresponding to the minimum value index of the difference between the adjacent frequency spectrum parameters of the fiftest table is less than the sixth limit. The fifth condition includes: an index value of the minimum value of the difference between adjacent frequency spectrum parameters of the nth frame is less than the eighth limit, and the minimum difference is less than the seventh limit. If the difference between the adjacent frequency spectrum parameters of frame i 'corresponds to at least one of the fourth condition or the fifth condition, the correction of the frequency spectrum parameter of the parent frame is determined, or if the difference between the adjacent frequency spectrum parameters of the resume frame do not match the fourth condition or the fifth condition, the non-correction of the frequency spectrum parameter of the nth frame is determined [0025] When correction is performed, a corrected frequency spectrum parameter Frequency of the nth frame is determined according to a weighting operation performed on the nth frame frequency spectrum parameter (i-1) and the nth frame frequency spectrum parameter; or a spectrum corrected parameter The frequency frame of the nth frame is determined according to a weighting operation performed on the frequency spectrum parameter. ia of table i "and the default frequency spectrum parameter of table i".

In a possible way of implementing the first aspect, the need to correct the frequency spectrum parameter of the nth frame can be determined according to the correlation between the nth frame and the (i-1) frame. 'esimo. When the need for correction of the frequency spectrum parameter of the i th frame is determined, the frequency spectrum parameter of the i th frame is corrected according to the frequency spectrum parameter of the i th frame and a frequency parameter. frame frequency spectrum (i-1), or the frame frequency spectrum parameter is corrected according to the frame frequency spectrum parameter and a preset frequency spectrum parameter of the very nice picture. The correlation between the nth frame and the (i-1) table includes a relationship of values between a ninth limit and the sum of the differences between the frequency spectrum parameters that correspond to some or all of the same indices. of (i-1) is the nth frame and the nth frame.

When the need to correct the frequency spectrum parameter of the nth frame is determined, the difference between the adjacent frequency spectrum parameters of the nth frame is first determined, where each difference corresponds to an index and the frequency spectrum parameter includes an immittance spectral frequency (ISF) or line spectral frequency (LSF), then it is determined whether the frequency frame parameter of the nth frame and the frequency spectrum parameter of the frame (i-1) 'nth corresponds to a sixth condition, where the sixth condition includes: the sum of the differences between the frequency spectrum parameters that correspond to some or all of the equal indices of table (i-1) -th and table i "nth is greater than the ninth limit; and if the frequency spectrum parameter of the nth frame and the frequency spectrum parameter of (jl) "seventh frame correspond to the sixth condition, the correction of the frequency spectrum parameter of the nth frame is determined, or if the i th frame frequency spectrum parameter and i th frame frequency spectrum parameter i-1 do not match the sixth condition, non-correction of i th frame frequency spectrum parameter is determined [0028 ] When correction is performed, a frequency frame corrected frequency parameter is determined according to a weighting operation performed on the frame frequency spectrum parameter (i-1) and the frequency spectrum parameter. frequency of frame i "; or a frequency spectrum corrected parameter of table i" is determined according to a weighting operation performed on frequency spectrum parameter d Frame I and the Default Frequency Spectrum Parameter of Frame I In a possible way of implementing the first aspect, the need to correct the excitation signal of Frame I can be determined according to the correlation between the summary frame and the (i-1) frame and the energy stability between the tenth frame and the (i-1) frame. When the need for correction of the ith frame excitation signal is determined, the ith frame excitation signal is corrected according to the energy stability between the jth frame and the ith frame (i-1). Specifically, a pre-synthesized signal from the ith frame is first determined according to the excitation signal from the ith frame and the frequency spectrum parameter of the ith frame.

Next, it is determined whether an absolute value of the difference between the power of the pre-synthesized signal of the nth frame and the energy of a synthesized signal of the nth frame is greater than one tenth limit. If the absolute value of the difference between the pre-synthesized signal energy of frame i and the synthesized signal energy of frame i-1 is greater than the tenth limit, the correction of the excitation signal of the table i, or if the absolute value of the difference between the pre-synthesized signal energy of table i and the synthesized signal energy of table (i-1) is less than or equal to the tenth limit is determined non-correction of the i-th frame excitation signal Alternatively, it is determined whether a relationship between the power of the i-th frame pre-synthesized signal and the energy of a (i-1) synthesized signal ~ th frame is greater than an eleventh limit, where the eleventh limit is greater 1. If the ratio of the pre-synthesized signal energy of frame i to the synthesized signal energy of frame i-1 is greater than the eleventh limit, the correction of the excitation of frame i ', or if the relationship between the power of the synthesized signal of frame i'th and the power of synthesized signal of frame (i-1)' is less than or equal to the eleventh threshold, is determined not correcting the excitation signal of the resime frame.

Alternatively, it is determined whether a relationship between the energy of a pre-synthesized signal from frame (i-1) -th and the energy of a synthesized signal from frame i-1 is less than a twelfth limit, where the twelfth limit is less than 1. If the ratio of the pre-synthesized signal energy of frame (i-1) 's to the synthesized signal energy of frame i's is less than the twelfth limit, it is the correction of the excitation signal of frame i "is determined, or if the ratio between the power of the synthesized signal of frame (i-1)" and the power of the synthesized signal of frame i is greater than or equal to twelfth limit, the non-correction of the excitation signal of Table I is determined.

[0033] When correction is performed, a second correction factor is determined according to the energy stability between the nth frame and the nth frame (−1), where the second correction factor is less than 1. and then the excitation signal from the nth frame is multiplied by the second correction factor to obtain a corrected excitation signal from the nth frame. Optionally, the second correction factor is a relationship between the frame (i-1) 'energy and frame i' energy, or the second correction factor is a relationship between the energy of the same number of subframes of frame (i-1) '' and frame i “Very good.

In a possible way of implementing the first aspect, the need to correct the frame excitation signal may be determined according to the correlation of a frame (i-1) signal. need for correction of the frame excitation signal i, the frame excitation signal is corrected according to the energy stability between the frame i and the frame (i-1. table (i-1) -sim ° jnc | uj a ratio of values between a thirteenth limit and a signal correlation value of table (i-1) "nth, and a ratio of values between a fourteenth and a deviation of a pitch period of the signal from frame (i-1) ~ es, mo.

When the need to correct the excitation signal from frame i 'is determined, it is also determined whether the signal from (i-1)' frame meets a seventh condition. If the signal from (i-1) 'If the seventh frame corresponds to the seventh condition, the correction of the excitation signal of the seventh frame is determined, or if the signal of the frame (i-1)' and the seventh condition does not correspond to the correction of the excitation signal The seventh condition is: (i-1) - the seventh frame is a lost frame, a signal correlation value of frame (i-1) 'is higher than the thirteenth limit, and the The pitch pitch deviation of the (i-1) 's frame is less than the fourteenth.

When the correction is performed, a third correction factor is first determined according to the energy stability between the nth frame and the nth frame (i-1), where the third correction factor is less than 1; and thereafter, the excitation signal from frame i is multiplied by the third correction factor to obtain a corrected excitation signal from frame i.

In a possible way of implementing the first aspect, the need to correct the excitation signal of the nth frame can be determined according to the correlation between the nth frame signal and a (i-) frame signal. 1. When the need to correct the excitation signal of the nth frame is determined, the excitation signal of the nth frame is corrected according to the energy stability between the nth frame and the frame ( I-1) -this. The correlation between the i th frame signal and the i th frame signal includes a ratio of values between a thirteenth bound and a th frame i correlation value, and a relation of values between a fourteenth and a deviation of a pitch period of the signal from frame i.

When the need to correct the excitation signal of the nth frame is determined, it is determined whether the signal from the (i-1) "frame" and the signal from the nth frame correspond to an eighth condition. (i-1) 'nth frame signal and the i'th frame signal correspond to the eighth condition, the correction of the excitation signal of the nth frame is determined, or if the (i-1)' nth frame signal and the signal of the ith frame does not correspond to the eighth condition, the non-correction of the excitation signal of the ith frame is determined. The eighth condition includes: (i-1) -th frame is a lost frame, a signal correlation value of frame (i-1) "nth is greater than the predefined thirteenth threshold, and the pitch period offset of the i'esimo frame signal is less than the fourteenth default limit.

When the correction is performed, a third correction factor is first determined according to the energy stability between the nth frame and the nth frame (i-1), where the third correction factor is less than 1, and then the excitation signal from the nth frame is multiplied by the third correction factor to obtain a corrected excitation signal from the nth frame. Optionally, the third correction factor is a relationship between the energy of frame (i-1) -th and the energy of frame i-th, or the third correction factor is a relationship between the energy of the same number of subframes of frame (i-1) -th and frame great.

In a possible way of implementing the first aspect, the need to correct the frame excitation signal i 'can be determined according to the correlation between a frame i (1)' signal and a frame signal. When the need for correction of the frame excitation signal res is determined, m ° the frame excitation signal is corrected according to the energy stability between the frame resume and the frame. The correlation between the frame (i-1) signal and the frame (i-2) signal includes: a relationship of values between a thirteenth limit and a value of frame signal correlation (i-2) "is fine, and if a frame excitation signal (i-1) is corrected.

When the need to correct the nth frame excitation signal is determined, it is determined whether the nth frame signal (i-2) and the nth frame signal (i-1) correspond to a ninth If the signal from the (i-2) "seventh frame and the signal from (i-1)" the seventh frame corresponds to the ninth condition, the correction of the excitation signal from the nth frame is determined, or if the signal from the (i-2) 'nth frame and the sign of (i-1) -th frame do not match the ninth condition, non-correction of the excitation signal of frame nth is determined.The ninth condition includes: the frame (i -2) ~ th is a lost frame, a signal correlation value of frame (i-2) 'th is greater than the thirteenth limit and frame excitation signal (i-1)' is corrected.

When correction is performed, a fourth correction factor is determined according to the energy stability between frame I and frame (i-1), where the fourth correction factor is lower. that 1; and the excitation signal from frame i is multiplied by the fourth correction factor to obtain a corrected excitation signal from frame i.

In a possible way of implementing the first aspect, the need to correct the ith frame excitation signal can be determined according to the correlation between a "th" frame (i-1) signal and a signal from the ith frame. When the need for correction of the excitation signal of the nth frame is determined, the excitation signal of the nth frame is corrected according to the energy stability between the parent and the nth frame. The correlation between the frame (i-1) signal and the frame (i-2) signal includes a relationship of values between a thirteenth limit and a correlation value. of the frame signal (i — 2) "nth, and a relation of values between a fifteenth limit and an algebraic codebook contribution of a frame excitation signal (i-1)" th.

When the need to correct the nth frame excitation signal is determined, it is determined whether the nth frame signal (i-2) and the nth frame signal (i-1) correspond to one tenth. If the signal from the (i-2) "tenth frame and the signal from frame (i-1)" th corresponds to the tenth condition, the correction of the excitement signal from the nth frame is determined, or if the signal from the (i-2) 'The tenth frame and the sign of the (i-1 seventh frame do not match the tenth condition, the non-correction of the excitation signal of the i'th frame is determined. The tenth condition includes: o (i-2) "nth frame is a lost frame, a frame signal correlation value (i-2) 'nth is greater than the thirteenth limit and the algebraic codebook contribution of the frame excitation signal (i-1)' nth is less than the fifteenth limit.

When correction is performed, a fourth correction factor is determined according to the energy stability between the nth frame and the nth frame (i-1), where the fourth correction factor is less than 1. and the excitation signal from frame i is multiplied by the fourth correction factor to obtain a corrected excitation signal from frame i.

[0046] In a possible way of implementing the first aspect, the need to correct the updated status excitation signal of the nth frame can be determined according to the correlation between a "nth and nth" frame signal (i-1). When the necessity of correcting the excitation signal with updated status of the nth frame is determined, the excitation signal with updated status of the nth frame is corrected according to the power stability between the table frame and the frame (i-1) - th, correlation between the signal frame (i-1) - th and the signal of the table frame includes: the correlation between the frame (i-1) - th and the frame and if an excitation signal from the (i-1) 'seventh frame is corrected.

When the need to correct the updated status excitation signal of the nth frame is determined, it is determined whether the nth frame signal and the (i-1) 'nth frame signal correspond to one eleventh. condition. If the signal from frame i and the signal from (i-1) 's | frame correspond to the eleventh condition, the correction of the excitation signal with updated status from frame i' is determined, or if the signal from frame i and the sign in frame (i-1) "nth does not match the eleventh condition, the non-correction of the excitation signal with updated status of frame i" is determined. The eleventh condition includes: table i "or the frame (i-1) - is a highly correlated frame and the excitation signal of frame (i-1)" is corrected.

[0048] When correction is performed, a fifth correction factor is determined according to the energy stability between the nth frame and the nth frame (i-1), where the fifth correction factor is less than 1. and the updated status excitation signal from frame i "es, multiplied by the fifth correction factor to obtain a current status corrected excitation signal from the nth frame.

In a possible way of implementing the first aspect, when frame i "is a normal frame, the method also includes: processing a decoded signal from frame i" to obtain a correlation value of the decoded signal from the nth frame; determine the correlation of a frame i "signal according to any or any combination of a frame i decoded signal correlation value", a relationship of values between the pitch periods of all subframes in frame i " nth, a spectrum slope value of the nth frame or zero crossing rate of the ninth frame; determining the energy of frame i "s according to the decoded signal of frame i" s; determining the energy stability between the energy of the nth frame and the energy of a nth frame (i-1) according to the energy of the nth frame and the energy of the nth frame (i-1); determining the energy of each subframe of the ith frame according to the decoded signal of the resume frame; and determining the energy stability between the subframe of the nth frame according to the energy of each subframe of the nth frame. To estimate or correct a parameter of a frame (i + 1), the correlation of the signal from the frame i, the energy stability between the subframes of the frame i, and the energy stability between the energy of the frame i and the energy of table (i-1) is determined.

A second aspect of the present invention provides a frame loss compensation processing apparatus. The apparatus includes a lost frame determination module, an estimation module, a obtaining module, a generation module and a signal synthesis module. The lost frame determination module is configured to determine, using a lost frame flag bit, whether a nth frame is a lost frame. The estimation module is set to: when the nth frame is a lost frame, estimating a frequency spectrum parameter, a pitch period, and an i-frame gain according to at least one of an inter-frame relationship between the first n-frames of the i-frame or an intra-frame relationship between the first frames N of table i " The obtainment module is configured to obtain an algebraic codebook from frame i. The generation module is configured to generate an excitation signal from frame i according to the pitch period and gain relative to the Table i "and obtained by the estimation module by means of estimation and the algebraic codebook referring to table i and obtained by the obtainment module. The signal synthesis module is configured to synthesize a signal from the nth frame according to the frequency spectrum parameter for the nth frame and which is obtained by the estimation module by means of the estimation and the referring excitation signal. to the i'th frame and which is generated by the generation module. The interframe relationship between the first N frames includes at least one of the correlation between the first N frames or the energy stability between the first N frames, and the intra-frame relationship between the first N frames includes at least one of the correlation between the first N frames. intersubquads between the first N frames or the energy stability of intersubquads between the first N frames so as to obtain a more accurate parameter of the ith frame by estimating and improving the speech signal decoding quality.

[0051] In a possible way of implementing the second aspect, the frequency spectrum parameter of the ith frame is obtained by the estimation module by estimating according to the inter-frame relationship between the first frames N of frame res O The estimation module is specifically configured to: determine the weight of a frequency frame parameter of the nth frame (i-1) and the weight of a preset frequency spectrum parameter of the nth frame according to the correlation between the first N frames of table i "e; and perform a weighting operation on the frequency spectrum parameter of table (i-1)" e and the default frequency spectrum parameter of table i "according to the weight of the parameter Frequency Spectrum (i-1) "is the weight and the weight of the default frequency spectrum parameter of Frame i for obtaining the quad frequency spectrum parameter In a possible way of implementing the second aspect, the correlation between the first N frames of frame I includes a relation of values between a second boundary and a spectrum slope parameter of a frame signal. (i-1) "nth, a relation of values between a first boundary and a normalized signal autocorrelation value of the frame (i-1)" nth and a relation of values between a third boundary and a deviation from a pitch period of the signal from frame (i-1) " Correspondingly, the estimation module is specifically configured to: if the signal from frame (i-1) 'corresponds to at least one of a first condition, a second condition, and a third condition, determine that the weight of the parameter of frequency spectrum of table (i-1) is the first weight and the weight of the default frequency spectrum parameter of table i is a second weight; or if the signal from (i-1) "nth frame does not correspond to a first condition, a second condition, or a third condition, determines that the weight of the frequency (i-1)" nth frame frequency spectrum parameter is a second weight and that the weight of the default frequency spectrum parameter of table i is the first weight. The first weight is greater than the second weight. The first condition is: the normalized autocorrelation value of the frame (i-1) signal is higher than the first limit, the second condition is: the signal slope parameter of the frame (i-1) "esimo" is greater than the second limit and the third condition is: the deviation of the pitch period of the (i-1) signal is less than the third limit.

[0053] In a possible way of implementing the second aspect, the pitch period of frame i "is obtained by the estimation module by estimating according to the correlation between the first frames N of frame i" and the correlation intersubframes between the first N frames of frame i "s. The correlation includes a relation of values between a fifth boundary and a normalized autocorrelation value of a signal from a frame (i-2)", a relation of values between a fourth boundary and a deviation of a pitch frame signal period (i-2), and a relation of values between the fourth boundary and a pitch deviation of a frame signal (j ~ 1) Correspondingly, the estimation module is specifically set to: if the pitch period deviation of the frame (i-1) signal is less than the fourth threshold, determine a period deviation value pitch of the frame signal (i-1) _th of a according to the pitch signal period of frame (i-1), and determine a pitch signal period of frame i and y according to the deviation value of the frame signal pitch period (i-1) 'nth is the frame signal pitch period (jl)' is the nth frame, where the nth frame signal pitch period includes a pitch period of each subframe of the ions frame, m and the deviation value of the nth frame. pitch period of the (i-1) -th signal is an average value of the differences between the pitch periods of all adjacent subframes of the (i-1) 'nth frame; or if the pitch (i-1) signal pitch deviation "s is greater than or equal to the fourth threshold, the normalized autocorrelation value of the (i-2) s signal is greater than the fifth threshold, and the deviation of the pitch period of the (i-2) -th frame signal is less than the fourth threshold, determine a deviation value of the pitch signal period of the (i-2) -th frame and the frame signal (i-1) -th in accordance with the frame signal pitch period (i-2) -th and the frame-signal pitch period (i-1) -th, and determine a signal pitch period res according to the pitch signal period of the frame (i-1) -th and the offset value of the pitch signal period of the frame (i-2) -th and the signal of the frame (i-1) -1) ~ th.

In one embodiment, the estimation module determines the pitch period shift value of the frame (i-1) signal according to the following formula: [0056], where P d) is a pitch period of a nth subframe of the (i-1) 'seventh frame ej = 0, 1,2, 3.

[0057] The estimation module determines the pitch period of the i'esimode frame signal according to the following formula: [0058], where P ® is a pitch period of a third subframe (i-1) 'nth frame P 'is the pitch shift value of the signal frame (i-1)' sth and PcirU) is a pitch period of a jth frame 's subframe.

In another embodiment, the estimation module determines the pitch period shift value P ^ of the nth frame signal (i-2) and the nth frame signal (i-1) according to [0060], where P is a pitch period of a nth subframe of the nth frame, p (l) (n) is a pitch period of a nth subframe of the nth table (i-1) ”th, m = 2,3, and n = 0,1.

[0061] The estimation module determines the pitch period of the i'esimode frame signal according to the following formula: [0062], where F x J is a pitch period of a third subframe of the frame (i-1) - esinro pv ^ Q | The deviation of the frame signal pitch period (i — 2) "isth and the frame signal (i-1)" is th and is a pitch period of a frame subframe i "esimo.

In a possible way of implementing the second aspect, the gain of the resume frame is obtained by the estimation module by estimating according to the correlation between the first N frames of the ith frame and the energy stability between the two. first frames N of the ith frame, and the gain of the ith frame includes an adaptive codebook gain and an algebraic codebook gain. The estimation module is specifically configured to: first determine the adaptive codebook gain of the ith frame according to an adaptive codebook gain of one (i-1) quadro frame or a predefined fixed value, the frame correlation (i-1) "nth and a sequence number of frame i" nth in multiple consecutive lost frames; then determine the weight of an algebraic codebook gain of frame (i-1) -th and the weight of a gain of a VAD frame according to the energy stability of frame (i-1) -th ; and finally, perform a weighting operation on the algebraic codebook gain of (table i-1) and the VAD frame gain according to the weight of the table algebraic codebook gain (i-1) ) 's is the weight of the VAD frame gain to obtain the algebraic codebook gain of the ith frame. Optionally, more stable frame (i-1) energy indicates a higher weight of frame algebraic codebook gain (i-1) "o, or the weight of the VAD frame gain correspondingly increases as the number of consecutive lost frames increases.

Optionally, before performing a weighting operation on the algebraic codebook gain of frame (i-1) "and the VAD frame gain according to the weight of the algebraic codebook gain of the table (i-1) "and the weight of the VAD frame gain, to obtain the algebraic codebook gain of the nth frame, the estimation module is also configured to: determine a first correction factor of according to a coding and decoding rate; and correct the gain of the algebraic codebook of table (i-1) 'esimo using the first correction factor.

[0065] In a possible way of implementing the second aspect, the obtaining module may obtain the algebraic codebook as follows: obtaining the algebraic codebook from the nth frame by estimation according to random noise; or determine the algebraic codebook of the nth frame according to the algebraic codebook of the first frames N of the nth frame [0066] In a possible way of implementing the second aspect, the retrieval module is also configured to: to determine the weight of a contribution from the algebraic codebook of frame ies, according to any one of a deviation of a pith period from a frame (M), or a correlation of a frame signal (i- 1) 'nth, a tilt spectrum value of the frame (i-1)' nth or a zero crossing rate of the frame (i-1) 's, or determine the weight of a book contribution algebraic codes of the nth frame by performing a weighting operation on any combination of a pitch period deviation of a nth frame (i-1), a correlation of a nth frame signal, a value tilt spectrum rate of the frame (i-1) ~ zero crossing of table (i-1) "nth; and performing an interpolation operation on an excitation signal with updated frame (i-1) status to determine a frame adaptive codebook. The generation module is specifically configured to: determine the contribution of table one algebraic codebook according to a product obtained by multiplying table one algebraic codebook by the gain of table algebraic codebook determine a contribution of the adaptive codebook of the ith frame according to a product obtained by multiplying the adaptive codebook of the i th frame by the gain of the adaptive codebook of the i th frame and carry out a weighting on the contribution of the algebraic codebook of table i and the contribution of the adaptive codebook of table i according to the weight of the contribution of the algebraic codebook of table i and e a weight of the contribution of the adaptive codebook of frame ies, mo, to determine the excitation signal of the nth frame, where a weight of adaptive codebook is 1.

In a possible way of implementing the second aspect, if the nth frame is a normal frame, the apparatus also includes: a decoding module, an evaluation module and a correction module. The decoding module is configured to obtain the frequency spectrum parameter, pitch period, gain, and algebraic codebook of the nth frame by decoding according to a received bit stream. The generation module is also configured to generate the frame excitation signal and an updated frame excitation status signal according to the pitch period, gain, and frame algebraic codebook nth and which are obtained by the decoding module by means of decoding. The evaluation module is configured to: when a single frame (jl) or a frame (i-2) is a lost frame, determine according to at least one of the interframe relationships or intraframe relationships between the the nth frame and the first N frames of the nth frame, the need to correct at least one of the frequency spectrum parameter, the excitation signal or the updated status excitation signal of the tenth frame. set to: when the evaluation module determines that it must correct at least one of the frequency spectrum parameter, the excitation signal or the updated status excitation signal of the nth frame, correct at least one of the spectrum parameter frequency, the excitation signal or the excitation signal with the updated status of the nth frame according to at least one of the interframe relationships or the intraframe frame i'esimoe the first frames N of frame i'esimo.

[0068] The signal synthesis module is also set to; synthesizing the frame i signal according to a correction result performed by the correction module on at least one of the frequency spectrum parameter, the excitation signal or the updated frame excitation signal; or when the evaluation module determines the non-correction of the frequency spectrum parameter, the excitation signal or the updated status excitation signal of the nth frame, synthesizes the nth frame signal according to the frequency spectrum parameter. frequency, the excitation signal and the excitation signal with updated status of the ith frame. The interframe relationship includes at least one of the correlation between the ith frame and the first N frames of the ith frame or the stability. between frame i and the first N frames of frame i, and the intraframe relationship includes at least one of the intersubject correlation between frame i and the first N frames of the ith frame or the energy stability of the intersubramers between the ith frame and the first N frames of the resize frame. At least one of the frequency spectrum parameter, excitation signal or excitation signal with updated status of the ith frame is corrected so that an easy transition of both global energy between adjacent frames and energy in the same band frequency can be deployed.

In a possible way of implementing the second aspect, the evaluation module is configured to determine, according to the correlation of the ith frame, the need to correct the frequency spectrum parameter of the ith frame. When the evaluation module determines that it must correct the frequency spectrum parameter of the nth frame, the correction module is configured to: correct the frequency spectrum parameter of the nth frame according to the frequency spectrum parameter Table i "and a frequency spectrum parameter of table (i-1) ~ eso ° or correct the frequency spectrum parameter of table eΓ ° according to the frequency spectrum parameter of table i and a default frequency spectrum parameter of table mo es, mo The correlation of the nth frame includes a ratio of values between a sixth limit and one of the two frequency spectrum parameters that corresponds to an index of a minimum value of the difference between the adjacent frequency spectrum parameters of the nth frame, a relationship of values between a seventh threshold and the minimum value of the difference between the adjacent frequency spectrum parameters of the nth frame, and a ratio of values between an eighth limit and the index of the minimum value of the difference between the adjacent frequency spectrum parameters of the nth frame [0070] Correspondingly, the evaluation module is specifically set to: first determine the difference between the adjacent frequency spectrum parameters of the frame quadro, where each difference corresponds to an index and the frequency spectrum parameter includes an immittance spectral frequency (1SF) or a line spectral frequency ( LSF); then verifying whether the difference between the adjacent frequency spectrum parameters of the parent frame corresponds to at least one of a fourth condition or a fifth condition; and if the difference between the adjacent frequency spectrum parameter of the parent frame corresponds to at least one of the fourth or fifth condition, determine the correction of the frequency spectrum parameter of the nth frame, or if the difference between the Adjacent frequency spectrum parameters of the nth frame do not match the fourth or fifth condition, determine non-correction of the frequency spectrum parameter of the nth frame. The fourth condition includes: one of the two frequency spectrum parameters which corresponds to the minimum value index of the difference between the adjacent frequency spectrum parameters of the fiftest table is that the sixth limit and the fifth condition includes: an index value of the minimum value of the difference between the adjacent frequency spectrum parameters of the nth frame is less than the eighth limit, and the minimum difference is less than the seventh limit.

The correction module is specifically configured to: determine a corrected frequency spectrum parameter of the nth frame according to a weighting operation performed on the frequency spectrum parameter of (i-1) and the nth frame. frequency spectrum parameter of the ith frame; or determining a corrected frequency spectrum parameter of the nth frame according to a weighting operation performed on the frequency spectrum parameter of the nth frame and the default frequency spectrum parameter of the nth frame.

In a possible way of implementing the second aspect, the evaluation module is configured to determine, according to the correlation between the nth frame and the nth frame, the need to correct the parameter. frequency spectrum of the frame res, mo. When the evaluation module determines that it must correct the frequency spectrum parameter of frame i, the correction module is configured to: correct the frequency spectrum parameter of frame i according to the frequency spectrum parameter of the nth frame and a frequency spectrum parameter of the nth frame (i-1), or correct the frequency spectrum parameter of the nth frame according to the frequency spectrum parameter of the nth frame and a preset frequency spectrum parameter of the ith frame. The correlation between Table I and Table I-1 includes a relation of values between a ninth limit and the sum of the differences between the frequency spectrum parameters that correspond to some or all of the same table indices. (i-1) "th and frame i" th.

Correspondingly, the evaluation module is specifically configured to: first determine the difference between adjacent frequency spectrum parameters of table i, where each difference corresponds to an index and the frequency spectrum parameter includes a immittance spectral frequency (ISF) or line spectral frequency (LSF), then check whether the frequency spectrum parameter of table i "nth and the frequency spectrum parameter of frame (i-1)" nth correspond to a sixth condition, and if the frame frequency spectrum parameter and the frame frequency spectrum parameter (i-1) are the sixth condition, determine the correction of the frame frequency spectrum parameter i "nth, or if the frequency spectrum parameter of table i" nth and the frequency spectrum parameter of table (jl) "nth do not match the sixth condition, detains terminate the non-correction of the frequency spectrum parameter of Table I. The sixth condition includes: the sum of the differences between the frequency spectrum parameters that correspond to some or all of the equal indices of the nth frame and the nth frame is greater than the ninth limit.

The correction module is specifically configured to: determine a corrected frequency spectrum parameter of the nth frame according to a weighting operation performed on the frequency spectrum parameter of the (i — 1) frame and the frequency spectrum parameter of the ith frame; or determining a corrected frequency spectrum parameter from table i according to a weighting operation performed on the frequency spectrum parameter from table i and the default frequency spectrum parameter from table i.

In a possible way of implementing the second aspect, the evaluation module is configured to determine, according to the correlation between the nth frame and the nth frame (i-1) and the energy stability between the frame i and frame (i-1) "is the need to correct the excitation signal from frame i. When the evaluation module determines that it must correct the excitation signal from frame i, the module The correction signal is configured to correct the excitation signal of the nth frame according to the energy stability between the nth frame and the nth frame (i-1).

The evaluation module is specifically configured to: first determine a pre-synthesized frame signal in accordance with the frame excitation signal and frame frequency spectrum parameter; and then determine whether an absolute value of the difference between the power of the i-th frame pre-synthesized signal and the power of a th-frame (i-1) synthesized signal is greater than one tenth threshold, and whether the absolute value of the difference between the preset synthesized signal energy of the i th frame and the synthesized signal energy of the i-1 frame is greater than the tenth limit, determine the correction of the excitation signal of the tenth frame, or if the absolute value of the difference between the preset synthesized signal energy of frame i "and the synthesized signal energy of (i-1)" frame is less than or equal to the tenth limit, determine the non-correction of the excitation signal of the ith table, or to determine whether a relationship o between the energy of the presynthesized signal of frame i and the energy of a synthesized signal of (i-1) 'frame is greater than an eleventh limit, where the eleventh limit is greater than 1, and if the ratio between the preset synthesized signal energy of the ith frame and the synthesized signal energy of the ith frame (i-1) "if greater than the eleventh threshold, determine the correction of the excitation signal of the nth frame , or if the ratio of the pre-synthesized signal energy of frame i "and the synthesized signal energy of (i-1) 'frame is less than or equal to the eleventh threshold, determine the non-correction of the excitation signal. of the Γth frame; or determining whether a relationship between the energy of a pre-synthesized signal from the (i-1) 'frame and the energy of a synthesized signal from the frame i is less than a twelfth threshold, where the twelfth threshold is smaller. than 1, and if the ratio between the pre-synthesized signal energy of (i-1) 's frame and the synthesized signal energy of frame i is less than the twelfth threshold, determine the correction of the excitation signal or if the ratio of the pre-synthesized signal energy of frame (i-1) to the synthesized signal energy of frame i is greater than or equal to the twelfth threshold, determine frame excitation signal correction Γth [0077] The correction module is specifically configured to: determine a second correction factor according to the energy stability between the i'th frame and the i'th frame (i-1), where the second correction factor is less than 1; and multiplying the excitation signal from the ith frame by the second correction factor to obtain a corrected excitation signal from the ith frame. Optionally, the second correction factor is a ratio between the energy of the nth frame (i-1) and the energy of the nth frame, or the second correction factor is a relationship between the energy of the same amount of subframes of (i-1) is the nth frame and the nth frame.

In a possible way of implementing the second aspect, the evaluation module is configured to determine, according to the correlation of a frame (i-1) signal, the need to correct the frame excitation signal. great. When the evaluation module determines that it must correct the excitation signal of the nth frame, the correction module is configured to correct the excitation signal of the nth frame according to the energy stability between the nth and the nth frame. o (i-1) is the nth frame. Frame Signal Correlation (i-1) - th includes a value relationship between a thirteenth threshold and a "th Frame" (i-1) signal correlation value, and a value relationship between a fourteenth and a deviation of a pitch period from the frame signal (i-1) "esimo.

Correspondingly, the evaluation module is specifically configured to: verify that the signal from frame (i-1) 's, mo corresponds to a seventh condition; and if the signal from (i-1) 'seventh frame corresponds to the seventh condition, determines the correction of the excitation signal from frame i, or if the signal from (i-1)' seventh frame does not correspond to the seventh condition, determine the non-correction of the excitation signal of the nth frame. The seventh condition is: frame (i-1) -th is a lost frame, a signal correlation value of frame (i-1) -th is greater than the thirteenth limit, and the pitch period deviation of the frame sign (i-1) is less than fourteenth.

[0080] The correction module is specifically configured to: determine a third correction factor according to the energy stability between frame i "nth and frame (i-1)" where the third correction factor is lower that 1; and multiplying the ith frame excitation signal by the third correction factor to obtain a corrected ith frame excitation signal.

[0081] In a possible way of implementing the second aspect, the evaluation module is configured to determine, according to the correlation between the nth frame signal and a nth frame signal (−1) “the need of correcting the excitation signal from frame i. When the evaluation module determines that it should correct the excitation signal from frame i, the correction module is configured to correct the excitation signal from frame i accordingly. with the energy stability between frame i and frame i-1. the correlation between the nth frame signal and the nth frame signal (i-1) includes a relationship of values between a thirteenth threshold and a nth th frame (i-1) signal correlation value, and a relation of values between a fourteenth and a deviation of a pitch period of the signal from the nth frame.

Correspondingly, the evaluation module is specifically configured to: verify that the signal from frame (i-1) "and the signal from frame i" correspond to an eighth condition; and whether the signal from (i-1) (-1) 'The eighth frame and the nth frame signal correspond to the eighth condition, determine the correction of the eighth frame excitation signal, or if the eighth frame (i-1) signal and the frame signal If the eighth condition does not correspond to the eighth condition, determine the non-correction of the frame's excitation signal The eighth condition includes: frame (i-1) “nth is a lost frame, a frame signal correlation value ( i-1) ~ th is greater than the thirteenth default limit, and the pitch pitch deviation of the i th frame is less than the fourteenth default limit.

[0083] The correction module is specifically configured to: determine a third correction factor according to the energy stability between the resize frame and the "i-1" frame, where the third correction factor is less than 1 and multiplying the excitation signal from the ith frame by the third correction factor to obtain a corrected excitation signal from the ith frame.

In a possible way of implementing the second aspect, the evaluation module is configured to determine, according to the correlation between a frame signal (i-1) 'and a frame signal (i-2)' nth, the need to correct the excitation signal from frame i. When the evaluation module determines that it should correct the excitation signal from frame i, the correction module is configured to correct the excitation signal from frame i. It is according to the energy stability between table i and table (i-1). the correlation between the nth frame signal (i-1) and the nth frame signal (i-2) th includes: a relationship of values between a thirteenth limit and a nth frame correlation value (i-2) ) 'nth, and whether an excitation signal from frame (i-1)' nth is corrected.

Correspondingly, the evaluation module is specifically configured to: verify that the frame signal (i-2) "is and the frame signal (i-1) corresponds to a ninth condition; and whether the signal from frame (i-2) 'th and the signal from frame (i-1)' th corresponds to the ninth condition, determine the correction of the excitation signal from frame i, or if the signal from (i-2) 'th frame and the signal from frame (jl) "if the ninth condition does not match, determine the non-correction of the excitation signal from frame i. The ninth condition includes: frame (i-2)" nth is a lost frame, a correlation signal value from frame (i-2) "es is greater than the thirteenth limit and frame excitation signal from (i-1)" is corrected.

[0086] The correction module is specifically configured to: determine a fourth correction factor according to the energy stability between the frame and the frame (i-1) "nth, where the fourth correction factor is less than 1; and multiplying the excitation signal of the tenth frame by the fourth correction factor to obtain the corrected excitation signal of the ith frame.

In a possible way of implementing the second aspect, the evaluation module is configured to determine, according to the correlation between a nth frame signal (i-1) and a frame signal (i-2) '. nth, the need to correct the excitation signal of the nth frame. When the evaluation module determines that it must correct the excitation signal of the nth frame, the correction module is configured to correct the excitation signal of the nth frame according to the energy stability between the nth and nth frame. frame (i-1) 'nth. The correlation between the signal of frame (i-1)' nth and the signal of frame (i-2) 'nth includes a relation of values between a thirteenth limit and a value of correlation of the (i-2) -th frame signal, and a relation of values between a fifteenth bound and an algebraic codebook contribution of an (i_1) -th frame excitation signal.

Correspondingly, the evaluation module is specifically configured to: verify that the signal from frame (i-2) "es, mo and the signal from frame (i-1) 'corresponds to a tenth condition; and if the signal from frame (i-2) "esimo and the signal from frame (i-1)" esimo corresponds to the tenth condition, determine the correction of the excitement signal from frame i, or if the signal from frame (i -2) 'nth and the frame signal (jl)' nth do not match the tenth condition, determine the non-correction of the excitation signal of the nth frame. The tenth condition includes: the (i-2) 'nth frame is a frame lost, a correlation value of the frame (i-2) signal is greater than the thirteenth limit, and the algebraic codebook contribution of the frame excitation signal (i-1) is smaller that the fifteenth limit.

[0089] The correction module is specifically configured to: determine a fourth correction factor according to the energy stability between frame I and frame (i-1) "where the fourth correction factor is lower which is 1, and multiplying the excitation signal from the nth frame by the fourth correction factor to obtain a corrected excitation signal from the nth frame.

In a possible way of implementing the second aspect, the evaluation module is configured to determine, according to the correlation between a signal from the nth frame (i-1) and the signal from the nth frame, the need to correct the updated frame excitation signal from frame i. When the evaluation module determines that it should correct the updated frame excitation signal from frame i, the correction module is configured to correct the excitation signal with updated status of frame i "is according to the energy stability between frame i and frame (i-1)". The correlation between the nth frame signal (i-1) and the nth frame signal includes: the correlation between the nth frame (i-1) and the nth frame, and whether an excitation signal of the nth frame Table (i-1) 'is correct.

Correspondingly, the evaluation module is specifically configured to: determine whether the frame signal i, m and the frame signal i-1 'correspond to an eleventh condition; and whether the signal of the ith frame and the signal of the frame (i-1) - th correspond to the eleventh condition, determine the correction of the excitation signal with updated status of the frame i, or if the signal of the frame i and the Frame sign (i-1) "If not meet the eleventh condition, determine non-correction of excitation signal with updated status of frame i" The eleventh condition includes: Table i "or frame (i -1) 'nth is a highly correlated frame and the excitation signal of frame (i-1) -this is corrected.

[0092] The correction module is specifically configured to: determine a fifth correction factor according to the energy stability between frame i "nth and frame (jl)" where the fifth correction factor is less than 1 ; and multiplying the updated status excitation signal from the nth frame by the fifth correction factor to obtain an updated status corrected excitation signal from the nth frame.

In accordance with the frame loss compensation processing method and apparatus provided in the embodiments of the present invention, a check is made to verify that a parent frame is a lost frame using a lost frame flag bit. When the nth frame is a lost frame, a frequency spectrum parameter, a pitch period, and a gain of the nth frame are estimated according to at least one of an interlayer relationship between the first N frames of frame i. 'nth or an intra-frame relationship between the first N frames of the nth frame. The inter-frame relationship between the first N-frames includes at least one of the correlation between the first N-frames or the energy stability between the first N-frames, and the inter-frame relationship between the first N frames includes at least one of the correlation between the first N frames. intersubquads between the first N frames or the energy stability of the intersubquads between the first N frames. A parameter of the nth frame is determined using the correlation between the signals of the first N frames, the energy stability between the signals of the first N frames. N-frames, the correlation of Intra-frame signals of each frame and the energy stability of the Intra-frame signal of each frame. The relationship between the signals is considered in order to obtain a more accurate parameter from frame i by estimating and improving the voice signal decoding quality.

BRIEF DESCRIPTION OF DRAWINGS

To describe the technical solutions in the embodiments of the present invention more clearly, the following guide will describe the accompanying drawings necessary to illustrate the embodiments. Of course, the accompanying drawings in the following description show only certain embodiments of the present invention and persons skilled in the art may derive other drawings from these attached drawings without creative effort.

Figure 1 is a flow chart of a frame loss compensation processing method according to embodiment 1 of the present invention;

Figure 2 is a flowchart of a frequency spectrum parameter estimation method according to embodiment 2 of the present invention;

Fig. 3 is a flow chart of a pitch period estimation method according to embodiment 3 of the present invention;

Figure 4 is a flowchart of a gain estimation method according to embodiment 4 of the present invention;

Fig. 5 is a flow chart of a frame loss compensation processing method according to embodiment 5 of the present invention;

[0100] Figure 6 is a comparison diagram between before the correction and after the correction of a spectrogram of a th frame. Figure 7 is a comparison diagram between before correction and after correction of a time domain signal of a nth frame;

Fig. 8 is a flow chart of a frame loss compensation processing method according to embodiment 6 of the present invention;

Figure 9 is a schematic structural diagram of a frame loss compensation processing apparatus according to embodiment 7 of the present invention;

Figure 10 is a schematic structural diagram of a frame loss compensation processing apparatus according to embodiment 8 of the present invention; and Figure 11 is a schematic diagram of a physical structure of a frame loss compensation processing apparatus according to embodiment 9 of the present invention. DESCRIPTION OF MODALITIES

In order to clarify the objectives, technical solutions and advantages of the embodiments of the present invention, the following guide will describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Of course, the embodiments described are some and not all embodiments of the present invention. Accordingly, all other embodiments obtained by persons skilled in the art based on embodiments of the present invention without creative efforts will fall within the scope of protection of the present invention.

Figure 1 is a flow chart of a frame loss compensation processing method according to embodiment 1 of the present invention. As shown in Figure 1, the method in this embodiment may include the following steps.

[0108] Step 101: Determine, using a lost frame flag bit, whether a frame is a lost frame.

[0109] A frame sent by an encoder may be lost during a transmission process. A network side correspondingly records if a current frame is a lost frame. A decoder determines, according to a lost frame flag bit in a received data packet, whether the nth frame is a lost frame. The parent frame here is a current frame being processed. By analogy, a frame (i-1) -th is a frame before the current frame and a (i + 1) -th frame is a frame after the current frame The frame before the current frame refers to a frame that is adjacent to the current frame and that precedes the current frame in a time domain, and the frame after the current frame refers to a frame that is adjacent to the current frame and that accompanies the current frame in a time domain.

[102] Step 102: If the frame is a lost frame, estimate a parameter of the very frame according to at least one of an inter-frame relationship between the first N frames of the frame or an intra-frame relationship between the first frames N of frame i'th.

The interframe relationship between the first N frames includes at least one of the correlation between the first N frames or the energy stability between the first N frames, and the intra-frame relationship between the first N frames includes at least one of the inter-frame correlation between the first N-frames or the energy stability of the inter-frame between the first N-frames. The correlation includes a relationship of values between signal frequency spectrum parameters, a relationship of values between signal correlation values. , a relationship of values between signal spectrum slope parameters, a relationship of values between signal pitch periods, the relationship between excitation signals and the like. The nth frame parameter includes a frequency spectrum parameter, a pitch period, a gain, and an algebraic codebook, and N is a positive integer greater than or equal to 1. The frequency spectrum parameter, the pitch period and gain can be obtained by estimation using at least one of the interframe ratio between the first N frames of the parent frame or the intraframe ratio between the first N frames of the tenth frame.

Correlation of a signal can be represented using a normalized signal autocorrelation value, and the normalized signal autocorrelation value is obtained by performing normalized autocorrelation processing on the signal. Alternatively, the correlation of a signal can be represented using an autocorrelation value, and the autocorrelation value can be obtained by autocorrelation processing, and is determined without normalized processing. The normalized autocorrelation value and autocorrelation value can be mutually converted and the same signal correlation is finally obtained. Specifically, signal correlation can be obtained by performing autocorrelation processing or normalized autocorrelation processing on any or any combination of a decoded signal correlation value from each frame, a relationship of values between periods. pitch, a slope value of the spectrum of each frame, or a zero crossing rate of each frame.

[0113] Signal correlation can include the following several cases: a low correlation, a low correlation rising edge, a low correlation falling edge, a moderate correlation, a high correlation, a high correlation rising edge and a high correlation descending edge. When signal correlation is being determined, a signal correlation value is compared to a correlation limit and the correlation limit may be some critical values selected from the previous cases. For example, if the correlation boundary is the low correlation descending edge, the signal correlation value will be greater than the low correlation descending edge, ie the correlation will be a moderate correlation, high correlation value, on the high correlation rising edge and the high correlation falling edge.

In this embodiment, the inter-frame energy stability of the first N frames refers to a relationship between the energy between the adjacent frames of the first N frames and the adjacent frames refers to two frames that are consecutive in a domain of time during transmission. Energy stability can be represented using a relationship between the energy of one frame and the energy of another frame. The energy of each frame can be obtained by determining an average quadratic value of the average energy of a signal or it can be obtained by determining the average amplitude of a signal. Specifically, the average energy E and the average amplitude M of each frame can be determined using the following two formulas: where N is a frame length or a subframe length, represents the amplitude of a frame already , and a value of j is 1.2, or N.

The frequency spectrum parameter includes an immittance spectral frequency (Immitance Spectral Frequency, ISF to shorten), a line spectral frequency (Line Spectral Frequency, LSF to shorten) and the like. The gain includes an adaptive codebook gain and an algebraic codebook gain. The pitch period is a characteristic of periodicity caused by the vibration of the vocal cords when a person makes a voiced sound, that is, a period of vibration of the vocal cords when a person makes a sound. The pitch period is in a reciprocal relationship with the vibration frequency of the vocal cords.

[0116] In this embodiment, when the i'th frame parameter is being estimated, the i'th frame parameter is determined according to the correlation between historical frames (ie the first N frames), the energy stability between the historical frames, the correlation of each frame and the energy stability of each frame.The relationship between the signals is considered in order to obtain a more accurate parameter of the ith frame by estimation.

[0117] Step 103; get an algebraic codebook from frame Γ [0118] Optionally, the algebraic codebook from frame i can be obtained by random noise estimation, or the algebraic codebook from frame i ' The same can be obtained by weighting the algebraic codebook of the first frames N of the nth frame, or the algebraic codebook of the tenth table can be estimated using an existing method.

[0119] Step 104: Generating an excitation signal from the ith frame according to a pitch period and a gain relative to the ith frame which are obtained by estimation and the algebraic codebook obtained from the i frame. "esimo.

Before this step is performed, the weight of a contribution from the algebraic codebook of the nth frame and an adaptive codebook of the nth frame must also be estimated. The adaptive codebook may be obtained by interpolation according to an excitation signal with updated status of frame (i-1). The weight of the algebraic codebook contribution can be obtained by performing a weighting operation according to any or any combination of a pitch (i-1) pitch shift, a correlation of a signal of the frame (i-1) _th, a frame's tilt spectrum rate (i-1) 'nth or a frame zero crossing rate (i-^ -th [0121] In this embodiment, the frame gain Ith includes an adaptive codebook gain and an algebraic codebook gain.When the excitation signal of the ith frame is first synthesized, the contribution of the algebraic codebook of the i th frame is obtained. by multiplying the algebraic codebook of table one by the gain of the algebraic codebook of table one, and a contribution of the adaptive codebook of table one is obtained by multiplying the adaptive codebook of table one. nth frame by l gain adaptive code book of table i. Then a weighting operation is performed on the algebraic codebook contribution from table i and the adaptive codebook contribution from table i according to the weight of the contribution from algebraic codebook from table i nth and a weight of the contribution of the adaptive codebook of frame i to the excitation signal of the nth frame, and a fixed value of an adaptive codebook weight is 1.

[105] Step 105: Synthesize a signal from the ith frame according to a frequency spectrum parameter referring to the ith frame and which is obtained by estimation and the generated excitation signal from the ith frame.

[0123] A specific way of implementing step 105 can be an existing method or a simple transformation of an existing method, and details of these options will not be described here.

In this embodiment, when a resize frame is a lost frame, a parameter of the i'th frame is estimated according to at least one of an interframe relationship between the first N frames of the resume frame or an intraframe relationship between the first frames N of the parent frame. the interframe relationship between the first N frames includes at least one of the correlation between the first N frames or the energy stability between the first N frames, and the intra-frame relationship between the first N frames includes at least one of the correlation between the first N frames. Intersubquads between the first N frames or the energy stability of the intersubquads between the first N frames. The parameter of frame i is determined using the correlation of signals between the first N frames, the signal energy stability between the first N-frames, the intra-frame signal correlation of each frame and the energy stability of the intra-frame signal of each frame The relationship between the signals is considered in order to obtain a more accurate parameter of the nth frame by means of estimation and improve the quality of voice signal decoding.

Based on embodiment 1, embodiment 2 of the present invention provides a frequency spectrum parameter estimation method. In this embodiment, a frequency spectrum parameter of a single frame is obtained by estimation according to an inter-frame relationship between the first frames N of the single frame. Figure 2 is a flow chart of an estimation method. frequency spectrum parameter according to embodiment 2. As shown in Figure 2, the method provided in this embodiment may include the following steps.

Step 201: Determine the weight of a frequency spectrum parameter of a frame (i-1 and the weight of a predefined frequency spectrum parameter of an iesimode table according to the correlation between the first N frames In this embodiment, the correlation between the first N frames of the parent frame includes a relation of values between a second boundary and a spectrum slope parameter of a signal from the (1-1) frame, a ratio of values between a first boundary and a nth frame normalized autocorrelation value (i-1), and a relation of values between a third boundary and a deviation of a pitch signal period from the frame (i-1) 1) The first limit, the second limit and the third limit are all predefined In one embodiment of the present invention, the first limit may be selected from a numerical range [0.3, 0.8] For example, the first limit may be and specifically 0.3, 0.5, 0.6 or 0.8. In one embodiment of the present invention, the second limit may be selected from a numerical range [-0.5, 0.5]. For example, the second limit may be specifically -0.5, -0.1, 0, 0.1, or 0.5. In one embodiment of the present invention, the third limit may be selected from a numerical range [0.5, 5]. For example, the third limit may be specifically 0.5, 1, or 5. For a signal from each frame, a spectrum slope parameter, a normalized autocorrelation value, and a pitch period for the signal are determined and stored for each signal. a decoder to decode a signal from a current frame according to the correlation between the first N frames of the nth frame. For example, a frequency spectrum parameter of the ith frame may be determined according to signal correlation and a frequency spectrum parameter for a frame prior to the ith frame (i.e. 1) “esimo). Generally, when the signal correlation and frequency spectrum parameter correlation for frame (i-1) are high and when the frequency spectrum parameter of frame ies is determined, the weight of the parameter (i-1) frequency frame is high and the weight of the default frame frequency parameter is low. When the signal correlation and frequency spectrum parameter correlation for frame (i-1) are low, the weight of the frame frequency spectrum parameter (i-1 is low and the weight of parameter default frequency spectrum of the nth frame is high, [0128] In one embodiment, if the nth frame (i-1) signal corresponds to at least one of a first condition, a second condition, or a third condition , the weight of the frequency spectrum parameter of table (i-1) is determined as a first weight and the weight of the default frequency spectrum parameter of table i is determined as a second weight. greater than the second weight. The first condition is: the normalized frame signal autocorrelation value (i-1) 'is higher than the first limit. The second condition is: the frame signal spectrum slope parameter ( i-1) 'esimo is more The third condition is that the deviation of the signal period from the frame (i-1) signal is less than the third limit.

Alternatively, if the frame (i-1) 's signal does not correspond to a first condition, a second condition, or a third condition, the weight of the frame frequency spectrum parameter (i-1) ~ This is determined as a second weight and the weight of the preset frequency spectrum parameter of Table I is determined as a first weight. In this embodiment, the first weight and the second weight may be predefined or may be determined according to inter-frame correlation between the frequency spectrum parameters of the first frames N of the seventh frame. Correspondingly, prior to step 201, the The first weight and the second weight also need to be determined according to the inter-frame correlation between the frequency spectrum parameters of the first N frames of Table I.

[0130] The normalized value of frame signal autocorrelation (i-1 seventh can be determined by performing normalized autocorrelation processing on a decoded frame (i-1) signal, q (jesvj0 ^ 0 pitch period of the signal from frame (i-1) 'nth § the sum of pitch period deviations of all subframes in frame (i-1)' nth relative to an average value of pitch periods of all subframes. If the pitch signal period of the frame (i-1) is being determined, the average value of the pitch periods of all subframes is first obtained by weighting the sum of the pitch periods of all subframes in the frame (i-1). -1) -651171 °, then a deviation of a pitch period from each subframe relative to the mean value of pitch periods is determined, and finally, the pitch period deviation of the signal from frame (i-1) ' This is obtained by calculating the sum of absolute values of the deviations of the pitch periods of t Alternatively, the pitch period deviation of the signal from frame (i-1) is obtained by determining the sum of the absolute values of the differences between the pitch periods of adjacent subframes.

[0131] For example, the first weight is 0.8, the second weight is 0.2, the first limit is 0.8, the second limit is 0.6, and the third limit is 0.2. In this case, when the normalized autocorrelation value of the frame (i-1) 's signal is greater than 0.8, the frame signal skew parameter (i-1)' s is greater than 0.6 and the deviation of the pitch (i-1) 's pitch signal period is less than 0.2, the weight of the pitch (i-1)' s frequency spectrum parameter 's is 0.8 and the weight of the default frequency spectrum parameter of frame 0,2 is 0.2, otherwise the weight of frequency spectrum parameter of frame (i-1)) is 0.2 and the weight of default frequency spectrum parameter of de Table I is the 0.8th.

[0132] Step 202: Carry out a weighting operation on the frame frequency spectrum parameter (i-1) and the default frame frequency spectrum parameter according to the weight of the frequency spectrum parameter of Table (i-1) is the weight of the default frequency spectrum parameter of Table i to obtain a frequency spectrum parameter of Table i.

[0133] In this embodiment, a decoder presets a frequency spectrum parameter for a missed frame, ie a predefined frequency spectrum parameter. When a nth frame is a lost frame, a weighting operation is performed accordingly. having a frequency frame parameter of the nth frame (i-1) and a preset frequency spectrum parameter of the nth frame for obtaining a frequency spectrum parameter of the nth frame. When the correlation of frame (i-1) is high, the correlation between adjacent frames is very likely to be high. Therefore, when a weight of the nth frame frequency spectrum parameter "i-1" is high, a weight of the nth frame default frequency spectrum parameter is correspondingly low. Thus, the frequency spectrum parameter Table I is determined mainly according to the default frequency spectrum parameter of Table I and is more accurate.

Based on embodiment 1, embodiment 3 of the present invention provides a pitch period estimation method. In this embodiment, a pitch period of a frame i 'is obtained by estimating according to the correlation between the first N frames of frame i' and the correlation of intersubquads between the first frames N of the frame i the correlation includes a value relationship between a fifth boundary and a normalized autocorrelation value of a single frame (i — 2) signal, a value relationship between a fourth boundary and a deviation of a frame signal pitch period (i- 2) 2) "nth, and a relation of values between the fourth boundary and a deviation of a pitch period from a signal of a frame (i-1th. In one embodiment of the present invention, the fourth boundary may be selected from from a numerical range [2,50]. For example, the fourth limit may be specifically 2, 5, 10, or 50. In one embodiment of the present invention, the fifth limit may be selected from a range of a rising edge low correlation to a high correlation rising edge. For example, the fifth boundary may be specifically a low correlation rising edge, a low correlation falling edge, or a high correlation rising edge. Low correlation rising edge and high correlation rising edge are classifications of predefined correlation values. For example, correlation values can be sequentially classified into low correlation, a low correlation rising edge, a low correlation falling edge, a high correlation rising edge, high correlation, moderate correlation, a low correlation edge. high correlation and similar, according to the magnitudes of the correlation values.

Figure 3 is a flow chart of a pitch period estimation method according to embodiment 3 of the present invention. As shown in Figure 3, the method provided in this embodiment may include the following steps.

[0136] Step 301: Determine if a pitch shift of a single frame (i-1) signal is less than a fourth threshold.

If the frame signal pitch period deviation (i-1) is even smaller than the fourth threshold, step 302 is performed, or if the frame signal pitch period deviation (i-1) is smaller. 1) If th is greater than or equal to the fourth limit, step 303 is performed.

Each subframe includes multiple subframes and the pitch period deviation of the frame signal (i-1) 'nth is the sum of the pitch period deviations of all subframes of the frame (i-1)' nth relative to an average value of the pitch periods of all subframes. For the deviation of the pitch period of the frame (i-1) signal, see the method of determination in mode 2.

[0139] Step 302: Determine a Frame Signal Pitch Period (i-1) Deviation Value 'nth According to the Frame Signal Pitch Period (i-1)' Nth and Determine a Pitch Period of a signal from a frame i 's according to the deviation value of the frame signal pitch period (i-1)' eso and the frame signal period's pitch period (i-1) 'is th e.

[0140] In this embodiment, the pitch period shift value of the nth frame signal (i-1) is an average value of the differences between the pitch periods of all adjacent sub frames of the nth frame. Assuming that each frame includes four subframes, the pitch period deviation value Pv of the frame (i-1) signal can be determined according to the following formula: [0141], where is a pitch period of a nth subframe of table (i-1) "th, ej = 0, 1,2, 3.

[0142] The pitch period of the i th frame signal can be determined according to the following formula: [0143], where P ® is a pitch period of a third subframe (the last subframe of the frame (i-1 ) 'th) of frame (i-1)' th, Pv is the pitch period shift value of the frame (i-1) signal 'th, and P ™ - ^ is a pitch period of a subframe ith of the ith frame.

[0144] Step 303: If a normalized autocorrelation value of a single-frame (i-2) signal is greater than a fifth bound and a deviation of a single frame's (i-2) signal pitch period is less than the fourth limit, determine a deviation value of the frame signal pitch period (i-2) "nth and the frame signal (i-1)" nth according to the frame signal pitch period (i-2) 'is the pitch signal period of the frame (i-1) -th, and determine the pitch period of the nth frame signal according to the pitch signal period of the frame (i-1) -1) 'th and the deviation value of the frame signal pitch period (i-2) "is th and the frame signal (i-1)" is th.

[0145] Table (i-2) "nth is a table before table (i-1)" nth. The deviation value of pitch period P2 of the nth frame (i-2) signal and the nth frame (i-1) signal can be determined according to the following formula: [0146], where P (m ) is a pitch period of a nth frame subframe (i-2) -th, P is a pitch period of a nth frame subframe (i-1) "th, m = 2, 3 en = 0.1.

Next, the pitch period of the frame signal "i" is determined according to the deviation value of the pitch period P 'of the frame signal (i-2) "es" and the signal of the frame (i -1) 'nth using the following formula: [0148], where P ® is a pitch period of the third frame subframe (i-1) “nth, ^ is the deviation value of the frame signal pitch period (i-2) 'nth is the sign of frame (i-1) "nth, and P ™ ^ is a pitch period of a subframe of the nth frame. n (~ 2) f) n (~ Z) (2) [0149] In the above formula, FJ and 1 y J are the last two subframes of table (i-2) "esimo and P ® and P ^ are the first two subframes of table (i-1) 'esimo . It can be learned that in the above formula, the deviation value of the frame signal pitch period (i-2) 's, mo and the frame signal (i-1)' nth is determined by selecting four consecutive subframes including the last two subframes of frame (i-2) -th and the first two subframes of frame (i-1) "nth. It can be understood that the deviation value of the pitch period of the signal of frame (i-2) "nth and the sign of frame (i-1)" nth can be determined by selecting six consecutive subframes, including the last three subframes of frame (i-2) "nth and the first three subframes of frame (i-1) "es", or the deviation value of the frame signal pitch period (i-2) "eso and the frame signal (i-1)" eso can be determined by selecting all subframes of the frame (i-2) -2) 'nth and all subframes of frame (i-1)' nth, or the deviation value of the pitch period of the signal from frame (i-2) 'th and the signal of frame (i-1)' th may be determined Selecting two consecutive subframes, including the last subframe of frame (i-2) 's and the first subframe of frame (i-1)' is th.

Based on embodiment 1, embodiment 4 of the present invention provides a method of estimating gain. Figure 4 is a flow chart of a gain estimation method according to embodiment 4 of the present invention. A one frame gain includes an adaptive codebook gain and an algebraic codebook gain. In this embodiment, the gain of the nth frame is obtained by estimation according to the correlation between the first N frames of the parent frame and the energy stability between the first N frames of the frame n. As shown in Figure 4, the method provided in this embodiment may include the following steps.

[0151] Step 401: Determining an adaptive codebook gain of a nth frame according to an adaptive codebook gain of a nth frame (i-1) or a predefined fixed value, frame correlation (i-1 seventh is a sequence number of the i th frame in multiple consecutive lost frames.

First, it is determined whether the nth frame is the first lost frame of the multiple consecutive lost frames. If all first m frames of frame i are frames lost, frame i is not a first frame lost of multiple consecutive missed frames, and is a positive integer greater than or equal to 1. If frame i is not a first lost frame of multiple consecutive lost frames, the adaptive codebook gain of the nth frame is determined according to an adaptive codebook gain that corresponds to the first lost frame of multiple consecutive lost frames, a factor attenuation and sequence number of the ith frame in the multiple consecutive missed frames.

[0153] If the first m frames of the ith frame are all lost frames, there will be m + 1 lost frames in total, including the ith frame. When the first frame lost in the m + 1 frame lost is lost, a decoder sets an adaptive codebook gain for the first frame lost, and an adaptive codebook gain gradually fades as the number of frames lost consecutive increases. In one embodiment, when consecutive frames are lost, each time a frame is lost, an adaptive codebook gain from a previous frame is multiplied by an attenuation factor. Considering that the adaptive codebook gain that corresponds to the first lost frame of consecutive lost frames is 1 and that the attenuation factor is 0.8, an adaptive codebook gain of the second lost frame of consecutive lost frames will be 1 * 0.8; an adaptive codebook gain of the third lost frame of consecutive lost frames will be i * co.8) 2 if by analogy, an adaptive codebook gain (rn + 1) 'lost of consecutive lost frames will certainly be , an attenuation factor can be subtracted from an adaptive codebook gain. For example, if the adaptive codebook gain that corresponds to the first lost frame of consecutive lost frames is 1 and the attenuation factor is 0.1, then an adaptive codebook gain of the second lost frame of consecutive lost frames is 1. -0.1, an adaptive codebook gain of the third lost frame of consecutive lost frames will be 1—2 * 0.1 and p0r analogy, an adaptive codebook gain of the lost frame (m + 1) "nth of the lost frames In this mode, the attenuation factor may be a fixed value or may vary according to power stability between frames.For example, the attenuation factor is smaller at a rising edge.

[0154] If the nth frame is the first lost frame that accompanies a normal frame, ie frame (i — 1) "nth frame is a normal frame and the nth frame is a lost frame, it is determined that the adaptive codebook gain of the nth frame is a fixed value, that is, when the first frame that accompanies a normal frame is lost, an adaptive codebook gain is set on the first lost frame, and if there is no no consecutive lost frame following the first lost frame, the adaptive codebook gains of these nonconsecutive lost frames will all be equal to the adaptive codebook gain of the first lost frame.

Step 402: Determine the weight of an algebraic codebook gain of the nth frame (i-1) and the weight of a gain of a voice frame detection frame (Voice Frame Detection, VAD for shorten) according to the energy stability of (i-1) "

[0156] It should be noted that step 402 can be performed prior to step 401, that is, there is no sequence for determining an algebraic codebook gain and for determining an adaptive codebook gain. Voice Frame Detection Frame (VAD) gain can be obtained by determining using an average square energy value, an average amplitude, and the like.

[0157] Â Sum of the algebraic codebook gain weight of frame (i-1) "nth and the weight of the VAD frame gain is a fixed value. More stable energy of frame (i-1)" nth corresponds algebraic codebook gain weight of frame (i-1) "and a lower VAD frame gain weight. Alternatively, as the number of consecutive lost frames increases, the weight of the gain of the VAD frame increases correspondingly and the weight of the algebraic codebook gain decreases correspondingly. If the frame (i-1) "energy is more stable and the number of consecutive lost frames increases as a function of the energy stability and the number of consecutive frames lost, the weight of the algebraic codebook gain of frame (i-1) will not increase or the increment will decrease. In a voice frame, the decoder periodically performs VAD detection to obtain VAD frame power.

[0158] Step 403: Perform a weighting operation on the algebraic codebook gain weight of frame (i-1) "and the VAD frame gain weight according to the algebraic codebook gain of ( i-1) "nth frame and the VAD frame gain, for obtaining an algebraic codebook gain of the jth frame [0159] Considering that the weight of the algebraic codebook gain of (i -1 nth frame is ^ and the weight of the VAD frame gain is β, the algebraic codebook gain of the nth frame will be gc ~ a'Sc + β'8 ^^ 0 where g (°} represents the book gain algebraic codebook from frame (i-1) 'and Scg is the VAD frame gain. When the algebraic codebook gain is less than the VAD frame gain, as the number of frames increases, algebraic codebook gain weight remains unchanged or gradually increases based on previous chart.

Optionally, before step 403 is performed, the method also includes: determining a first correction factor according to a coding and decoding rate and correcting the gain of the frame algebraic codebook (i-1) ) - nth using the first correction factor. For example, the gain of the algebraic codebook of frame (i-1) is corrected by multiplying the gain of the algebraic codebook of (i-1) by the first correction factor.

Modes 1 to 4 specifically describe how to determine a parameter of an ith frame according to at least one of an interframe relationship between the first N frames of the ith frame or an intra frame relationship between the first. frames N of the ith frame when the ith frame is a lost frame Embodiment 5 of the present invention describes how to correct the parameter of the resize frame when the resume frame is a normal frame. Frame loss compensation processing method according to the embodiment of the present invention As shown in Figure 5, the method provided in this embodiment may include the following steps.

[0162] Step 501; Obtain a parameter of the ith frame by decoding according to a received bit stream, where the parameter of the ith frame includes a frequency spectrum parameter, a pitch period, a gain, and a textbook. algebraic codes.

[0163] Step 502: Generate an excitation signal from the nth frame and an excitation signal with updated status from the nth frame according to the pitch period, gain, and algebraic codebook for frame i ' and which are obtained by decoding.

[0164] The excitation signal includes an adaptive codebook contribution and an algebraic codebook contribution. The adaptive codebook contribution is obtained by multiplying an adaptive codebook by an adaptive codebook gain. The algebraic codebook contribution is obtained by multiplying an algebraic codebook by a gain from the algebraic codebook. The adaptive codebook is obtained by interpolation according to a pitch period and an updated status excitation signal for a current frame. The algebraic codebook can be obtained by estimation using an existing method. The excitation signal is used for signal synthesis of frame i and the updated status excitation signal is used to generate an adaptive codebook of a later frame.

Step 503: If a (i-1) ~ frame or a (i-2) 'frame is a missing frame, determine, according to at least one of the inter-frame relationships or intra-frame relationships between the Resume Frame and First Frames N of Frame I and I, the need to correct at least one of the frequency spectrum parameter, the excitation signal or the updated status excitation signal of Frame I and I.

[0166] The interframe relationship includes at least one of the correlation between the nth frame and the first N frames of the nth frame or the energy stability between the nth frame and the first N frames of the nth frame , and the intraframe ratio includes at least one of the intersubramel correlation between the nth frame and the first N frames of the nth frame or the energy stability of the intersubrams between the nth frame and the first N frames of the nth frame That's right. When the need to correct at least one of the frequency spectrum parameter, the excitation signal, or the updated status excitation signal of the nth frame is determined, step 504 and step 506 are performed. When the non-correction of the frequency spectrum parameter, the excitation signal or the updated status excitation signal of the nth frame is determined, step 505 is performed.

[0167] Step 504: Correct at least one of the frequency spectrum parameter, excitation signal, or excitation signal with the updated status of the nth frame according to at least one of the interframe relations or the intraframe relations between the ith frame and the first frames N of the ith frame. Step 506 is performed after step 504.

[0168] Step 505: Synthesize a signal from the optimal frame according to the frequency spectrum parameter, the excitation signal and the updated status excitation signal from the summary frame.

[0169] Step 506: Synthesize a signal from the ith frame according to a correction result of at least one of the frequency spectrum parameter, the excitation signal or the updated status signal from the ith frame .

If only the frequency frame parameter of the i th frame is corrected, the signal of the frame i th will be synthesized according to a frequency spectrum corrected parameter of the frame i, mo, the excitation signal for frame i and which is obtained by decoding and the excitation signal with updated status for frame i and which is obtained by decoding. If only the excitation signal of frame i is corrected, the signal from frame i will be synthesized according to a corrected excitation signal from frame i, the frequency spectrum parameter referring to frame i, which is obtained by decoding and the excitation signal with updated status for the i'th frame and which is obtained by decoding. If only the updated frame excitation signal from the nth frame is corrected, the signal from the frame nth is synthesized according to the updated status excitation signal from the nth frame that has been corrected, the spectrum parameter of frequency referring to the esimoe frame which is obtained by means of decoding and the excitation signal referring to frame i 's and obtained by means of decoding. If the frequency spectrum parameter and the excitation signal from frame i' If they are corrected, the Frame I signal will be synthesized according to the Frame I frequency corrected spectrum parameter, the Frame I excitation corrected signal and the Frame updated status excitation signal. It is obtained by means of decoding. If the frequency spectrum parameter and the updated status excitation signal of the nth frame are corrected, the signal of the nth frame will be synthesized according to a corrected frequency spectrum parameter of the nth frame, a signal corrected excitation signal with updated status of frame i "and the excitation signal referring to the resume frame and which is obtained by decoding. If the excitation signal and excitation signal with updated status of the i'th frame are corrected, the signal from the ith frame will be synthesized according to a corrected excitation signal from the ith frame, a corrected excitation signal with updated status from the ith frame, and the frequency spectrum parameter for the resimoe frame that is obtained by If the frequency spectrum parameter, the excitation signal and the excitation signal with updated status of the nth frame are corrected, the frame signal The nth will be synthesized according to a corrected frequency spectrum parameter of the nth frame, a corrected excitation signal of the nth frame and a corrected excitation signal with the updated status of the nth frame.

[0171] It should be noted that if both frame (i-1) -th and frame (i-2) 'n are normal frames, the signal from frame n can be synthesized directly according to the parameter referring to frame i and which is obtained by decoding, without any need to correct the parameter of frame i. If frame (i-1) 'th or frame (i-2)' is a missing frame, there may be a particular deviation in a parameter from frame (i-1) 'n or frame (i-2) This is obtained by estimation, a relatively high change in interframe energy that is subsequently elicited and a decoded voice signal that is not globally stable. Therefore, in this embodiment, a decoder corrects at least one of the frequency spectrum parameter, the excitation signal, or the updated status excitation signal of the ith frame according to the correlation between the resume frame and the first N frames. Frame I and the energy stability between Frame I and the first N frames of Frame I so that an easy transition of both the overall energy between adjacent frames and the energy in the same frequency band can be achieved. (1) Frequency Spectrum Parameter Correction [0172] The frequency spectrum parameter includes an ISF or an LSF An ISF parameter is used in an example, The ISF parameter is obtained by weighting and converting it. if a frame ISP parameter and a frame ISP parameter (μι seventh, Quancj0 0th frame (i-1) "nth or frame (i-2)" nth is a lost frame, there may be a deviation particular between one determines of the ISF parameter of the ith frame and a normal ISF parameter (an ISF parameter obtained when the ith frame is not lost) of the ith frame. Therefore, certain energy in a low frequency formant location is much higher than the actual energy.

[0173] In one embodiment, the need to correct the frequency spectrum parameter of the ith frame can be determined according to the correlation of the ith frame. When the need for correction of the frequency spectrum parameter of the i th frame is determined, the frequency spectrum parameter of the i th frame is corrected according to the frequency spectrum parameter of the i th frame and a frequency parameter. frame frequency spectrum (i-1), or the frame frequency spectrum parameter is corrected according to the frame frequency spectrum parameter and a preset frequency spectrum parameter of the very nice picture. The nth frame correlation includes a ratio of values between a sixth limit and one of the two frequency spectrum parameters that corresponds to an index of a minimum value of the difference between the adjacent frequency spectrum parameters of the tenth frame, a ratio of values between a seventh limit and the minimum value of the difference between the adjacent frequency spectrum parameters of Table I and a ratio of values between an eighth limit and the index of the minimum value of the difference between the adjacent frequency spectrum parameters from frame i "esimo. In one embodiment of the present invention, the sixth limit may be selected from a numerical range [500, 2000]. For example, the sixth limit may be specifically 500, 1000, or 2000. In one embodiment of the present invention, the seventh limit may be selected from a numerical range [100, 1000]. For example, the seventh limit may be specifically 100, 200, 300, or 1000. In one embodiment of the present invention, the eighth limit may be selected from a numerical range [1,5]. For example, the eighth limit may be specifically 1, 2, or 5.

Correspondingly, determining, according to the correlation of the nth frame, the need to correct the frequency spectrum parameter of the nth frame specifically means: first, determining the difference between the adjacent frequency spectrum parameters. frequency of the nth frame, where each difference corresponds to an index, frequency spectrum parameters are arranged in ascending order and index values are also arranged in ascending order; then verify that the difference between the adjacent frequency spectrum parameters of the nth frame corresponds to at least one of a fourth condition or a fifth condition, where the fourth condition includes: one of the two frequency spectrum parameters corresponding to the minimum value index of the difference between the adjacent frequency spectrum parameters of Table 1 is less than the sixth limit, and the fifth condition includes: an index value of the minimum value of the difference between the adjacent frequency spectrum parameters of the ith frame is less than the eighth preset threshold, and the minimum difference is less than the preset seventh threshold, and if the difference between the adjacent frequency spectrum parameters of the tenth frame corresponds to at least one of the fourth condition or fifth condition, to determine whether the frequency spectrum parameter of table i should be corrected, or whether the between adjacent frequency spectrum parameters of the nth frame does not correspond to the fourth or fifth condition, determine that the jth frame frequency spectrum parameter should not be corrected [0175] In another embodiment, the The need to correct the frequency spectrum parameter of the ith frame is determined according to the correlation between the ith frame and the ith frame (i-1). When the need to correct the frequency spectrum parameter of the nth frame is determined, the frequency spectrum parameter of the nth frame is corrected according to the frequency spectrum parameter of the nth frame and a frequency parameter. frame frequency spectrum (i-1) "nth, or the frequency spectrum parameter of the nth frame is corrected according to the frequency spectrum parameter of the nth frame" and a predefined frequency spectrum parameter of the Ith frame. The correlation between the nth frame and the nth frame (i-1) includes a relation of values between a ninth limit and the sum of the differences between the frequency spectrum parameters that correspond to some or all of the same indices in the frame. (il) 'esimo' and the esimoimo frame. In one embodiment of the present invention, the ninth limit may be selected from a numerical range [100, 2000]. For example, the ninth limit may be specifically 100, 200, 300, or 2000.

Correspondingly, determining, in accordance with the correlation between frame i and frame (il), the need to correct the frequency spectrum parameter of frame i and specifically means: first, to determine the difference between the adjacent frequency spectrum parameters of the resume frame, where each difference corresponds to an index, then check whether the frequency spectrum parameter of the nth frame and the frequency spectrum parameter of the frame (i- 1) "nth corresponds to a sixth condition, where the sixth condition includes: the sum of the differences between frequency spectrum parameters that correspond to some or all of the equal indices of table (i-1)" nth and table i " nth is greater than the ninth limit; and if the frequency spectrum parameter of table i "and the frequency spectrum parameter of table (i-1)" nth meets the sixth condition, it is determined that the frequency spectrum parameter of the nth frame shall be corrected. , or if the frequency spectrum parameter in Table 1 and the frequency spectrum parameter in Frame (i-1) do not meet the sixth condition, determine that the frequency spectrum parameter in the Table should not be corrected. That's right.

[0177] In the previous two ways of implementation, the correction of the frame frequency spectrum parameter "i" according to the frame frequency spectrum parameter "i" and a frame frequency spectrum parameter (i-1 ) 'Specifically means: determining a frequency spectrum corrected parameter of the ith frame according to a weighting operation performed on the frequency spectrum parameter of the frame (i-1)' and the frequency spectrum parameter of the paintings, hand. The correction of the frequency spectrum parameter of Table 1 according to the frequency spectrum parameter of Table 2 and a predefined frequency spectrum parameter of Table 2 specifically means: determining a frequency spectrum corrected parameter of the table i "s according to a weighting operation performed on the frequency spectrum parameter of table i" s and the default frequency spectrum parameter of table i "s.

[0178] An ISF parameter is used in an example. The difference between the adjacent ISF parameters of the i th frame can be represented as isf diff O) and isf _ diff (i) = isf 0 +1) - isf (i), i = n - 2 ^ oncje | \ jq an order of the ISF parameter. If an ISF parameter that corresponds to an index of a minimum value of ISF - DIFF W from table i "th is lower than the sixth limit (for example, 800) and the minimum value of ISF - D1FF W is less than the seventh threshold (for example, 200), or the sum of the differences between the frequency spectrum parameters that correspond to some or all of the equal indices of table (i-1) 'nth and table i'th is greater than the ninth limit , a nth frame ISF parameter and a nth frame (i-1) ISF parameter are weighted to determine and obtain the nth frame corrected ISF parameter, or a ninth frame ISF parameter and a predefined ISF parameter of the nth frame are weighted to obtain the corrected ISF parameter of the nth frame. That the sum of the differences between the frequency spectrum parameters that correspond to some or all of the same indices of the frame (i-1) "esimo and table i" esimo is greater than the ninth limit meaning It indicates that the correlation of ISF parameters between adjacent frames is low.

[0179] Figure 6 is a comparison diagram between before the correction and after the correction of a nth frame spectrogram. As shown in Fig. 6, Fig. 6 (a) is a spectrogram of an original signal, and the original signal is a signal sent by an encoder. Figure 6 (b) is a spectrogram of a signal synthesized in the prior art. Figure 6 (c) is a spectrogram of a synthesized signal in accordance with the present invention. It can be learned by comparing figure 6 (a) to figure 6 (b) that an ellipse part in figure 6 (b) is much lighter than an ellipse part of the original sign in figure 6 (a). That is, the amount of energy recovered from a low frequency formant of the nth frame is much greater than the energy obtained by the correct recovery. Of course, an ISF parameter of the nth frame needs to be corrected correspondingly so that the energy in a formant location of the nth frame is closer to the actual amount of energy to achieve the effect shown in figure 6. (ç). (2) Excitation signal correction [0180] There is a particular deviation between an estimated pitch period of a lost frame and an actual pitch period of the lost frame. Therefore, when an adaptive codebook of the nth frame is obtained by interpolation using a frame excitation signal (M), the adaptive codebook of the nth frame is excessively strong and when emphasis reduction processing is performed on the excitation signal of the ith frame using a linear predictive coding synthesis filter (LPC to shorten) and a synthesized signal from the ith frame, the amount of The energy obtained is much greater than the actual energy of a synthesized signal. Of course, this can affect a normal frame that accompanies a missed frame (sometimes one or two frames that accompany the lost frame are affected, and sometimes more frames may be affected if the frequency of an excitation signal is too strong). In this case, an excitation signal and / or an updated status excitation signal needs to be corrected to some extent so that the energy of a synthesized signal is close to the actual energy.

[0181] In a first embodiment, the need to correct the excitation signal of frame i is determined according to the correlation between frame resume and frame (i-1) and power stability between Frame I and Frame I-1 When determining the need for correction of the frame excitation signal, the frame excitation signal is corrected for energy stability. between table i'th and table (i-1) 'th.

Specifically, a preset synthesized signal from the ith frame is first determined according to the excitation signal from the ith frame and the frequency spectrum parameter from the ith frame. Next, it is determined whether an absolute value of the difference between the preset synthesized signal energy of the ith frame and the synthesized signal energy of the ith frame (i-1) is greater than one tenth limit. If the absolute value of the difference between the pre-synthesized signal energy of the nth frame and the synthesized signal energy of the nth frame is greater than the tenth limit, the correction of the excitation signal of the nth frame is determined. table i, or if the absolute value of the difference between the power of the synthesized signal of table i and the energy of the synthesized signal of table (jl) is less than or equal to the tenth limit is determined not correction of the frame excitation signal Γ es, mo. Specifically, in one embodiment of the present invention, the tenth limit may be a value 0.2 to 1 times lower in the frame pre-synthesized signal energy and the frame synthesized signal energy (i- 1) 'esimo. For example, the tenth limit may be a value of 0.2, 0.5, or 1 time smaller.

Alternatively, it is determined whether a relationship between the power of the synthesized signal from frame i "and the power of a synthesized signal from frame (jl)" is greater than an eleventh limit, where the eleventh If the ratio between the preset synthesized signal energy of the nth frame and the synthesized signal energy of the i-1 frame is greater than the eleventh limit, the correction is determined. of the excitation signal from frame i ', or if the ratio between the pre-synthesized signal energy of the tenth frame and the energy of the synthesized signal from frame (i-1)' is less than or equal to the eleventh limit, is non-correction of the excitation signal of the nth frame is determined.In one embodiment of the present invention, the eleventh limit may be selected from a numerical range [1,1, 5]. limit may be specifically 1,1,1.25, 2, 2.5 or 5.

Alternatively, it is determined whether a relationship between the energy of a pre-synthesized signal from frame (i-1) _th and the energy of a synthesized signal from frame i "is less than a twelfth limit, where the twelfth limit is less than 1. If the ratio of the pre-synthesized signal energy of frame (i-1) 's to the synthesized signal energy of frame i's is less than the twelfth limit, it is determined the correction of the excitation signal of the nth frame, or if the ratio of the presynthesized signal energy of the nth frame (i-1) to the synthesized signal energy of the nth frame is greater than or equal to the tenth second limit, the non-correction of the excitation signal of frame i is determined. In one embodiment of the present invention, the twelfth limit may be selected from a numerical range [0.1, 0.8]. For example, the twelfth limit may be specifically 0.1, 0.3, 0.4, or 0.8.

Correspondingly, the correction of the excitation signal from frame i 's according to the energy stability between frame i and frame (i-1)' es specifically means: first, determining a second factor of correction according to the energy stability between frame i'esimo and frame (i-1) 'esimo, where the second correction factor is less than 1, and then multiply the excitation signal of the frame resimo by the second correction factor for obtaining a corrected excitation signal from the resume frame, [0186] Determine a second correction factor according to the energy stability between the nth frame and the (i-1) frame "neatly means determining that a ratio between the energy of (i-1) 's frame and the energy of the' i 'frame is the second correction factor; or determine that a ratio of the energy of the same number of subframes in frame (i-1) "to the nth frame is the second correction factor. Preferably, the same amounts of subframes in frame (i-1) For example, the last two subframes of the (i-1) -th frame and the first two subframes of the nth frame are selected to determine an energy relationship. selected subframes may not be consecutive.

[0187] In a second embodiment, the need to correct the nth frame excitation signal is determined according to the correlation of a nth frame (i-1) signal. When the need for correction is determined of the excitation signal from frame i, the excitation signal from frame i is corrected according to the energy stability between frame i and frame (i-1). frame (i-1) signal "includes a ratio of values between a thirteenth limit and a correlation value of frame (i-1) - and a ratio of values between a fourteenth and a deviation of a signal period of the frame signal (i-1) 'ith.

Correspondingly, determining, in accordance with the correlation of a signal from frame (i-1) "eso, the need to correct the excitation signal from frame i" and specifically means: verifying that the signal from frame (i-1) "nth corresponds to a seventh condition, where the seventh condition is: frame (i-1)" nth is a lost frame, a signal correlation value of frame (il) is greater than the thirteenth boundary and the deviation of the pitch signal from frame (i-1) "nth" is less than the thirteenth; and if the signal from frame (i-1) "nth is the seventh condition, determine that if shall correct the excitation signal of the nth frame, or if the frame signal (i-1 seventh does not meet the seventh condition, determines that the excitation signal of the nth frame shall not be corrected. excitation of the i-th frame according to the energy stability between the i-th frame and the i-1 frame Specifically: determine a third correction factor according to the energy stability between frame i and frame (i-1), where the third correction factor is less than 1; and multiplying the excitation signal from the nth frame by the third correction factor to obtain a corrected excitation signal from the nth frame. In one embodiment of the present invention, the thirteenth limit may be selected from a low correlation falling edge to a high correlation rising edge For example, the thirteenth limit may be specifically a low correlation falling edge or a high correlation rising edge In one embodiment of the present invention , the fourteenth may be selected from a numerical range [0.5, 20] For example, the fourteenth may be specifically 0.5, 2, 5, 10, or 20.

[0189] In a third way of implementation, the need to correct the nth frame excitation signal is determined according to the correlation between the nth frame signal and a nth frame signal (i-1) When the need for correction of the frame excitation signal is determined, the frame excitation signal is corrected according to the energy stability between frame i, es and frame (i-1). ~ the correlation between the tenth frame signal and the frame signal (jl) includes a value relationship between a thirteenth threshold and a frame signal correlation value (i-1) and a value relationship between a fourteenth and a deviation of a pitch period from the frame signal.

Correspondingly, determining, in accordance with the correlation between the i-frame signal and a n-frame signal (i-1), the need to correct the i-frame excitation signal specifically means : Verify that the signal from frame (i-1) 'e and the signal from frame i and i correspond to an eighth condition, where the eighth condition includes: frame (i-1)' esimo is a lost frame, a value Frame Signal Correlation (i-1) -th is greater than the thirteenth limit and the pitch period deviation of the signal of the i-th frame is less than the fourteenth; and if the signal from frame i-1 and the signal from frame i correspond to the eighth condition, it is determined that the excitation signal from frame i should be corrected, or if the signal from frame (i-1) 1) 'nth and the sign of the nth frame do not match the eighth condition, determine that the excitation signal of the nth frame should not be corrected. The correction of the excitation signal of the frame i-es, according to the energy stability between the frame i-1 and frame (i-1) 'specifically means: determining a third correction factor according to the stability of between the i-th frame and the frame (i-1 yesim ° where the third correction factor is less than 1; and then multiply the excitation signal from the i-es frame by the third correction factor for the obtaining a corrected excitation signal from the nth frame.

Determining a third correction factor according to the energy stability between frame i and frame (i-1) 'may specifically mean: determining that a relationship between the energy of frame (jl) and frame i' and the energy of the nth frame is a third correction factor; or determine that a relationship between the energy of the same number of subframes in frame (i-1) is the third correction factor.

In a fourth embodiment, the need to correct the i-frame excitation signal is determined according to the correlation between a nth frame (i-1) signal and a frame signal (i-1). 2) 'esimo. When the need for correction of the i-th frame excitation signal is determined, the i-th frame excitation signal is corrected according to the energy stability between the i-th frame and frame (i-1). nth the correlation between the frame signal (i-1) "nth and the frame signal (i-2)" nth includes: a relation of values between a thirteenth limit and a correlation value of the frame signal (i— 2) 'nth, and whether an excitation signal from frame (i-1)' is nth corrected.

Correspondingly, determining, in accordance with the correlation between a nth frame (i-1) signal and a nth frame (i-2) signal, the need to correct the frame excitation signal i nth specifically means: verifying that the nth frame sign (i-2) and the nth frame sign (i-1) correspond to a ninth condition, where the ninth condition includes: frame (i-2) th is a lost frame, a frame signal correlation value (i-2) 'th is greater than the predefined thirteenth limit and the frame excitation signal (i-1)' th is corrected; and if the signal from frame (i-2) 's and the signal from frame (i-1)' is the ninth condition, it is determined that the excitation signal from frame i 's should be corrected, or if the signal from (i-2) If the nth frame and the sign of the nth frame do not match the ninth condition, determine that the excitation signal of the nth frame should not be corrected. I-th according to the energy stability between the i-th frame and the (i-1) frame-th specifically means: determining a fourth correction factor according to the energy stability between the i-th frame and the frame (i-1) ', where the fourth correction factor is less than 1, and then multiply the excitation signal from frame i-1 by the fourth correction factor to obtain a corrected excitation signal from frame That's right.

[0194] In a fifth embodiment, the need to correct the ith frame excitation signal is determined according to the correlation between a nth frame (i-1) signal and a frame signal (i— 2) 'esimo. When the need for correction of the ith frame excitation signal is determined, the ith frame excitation signal is corrected according to the energy stability between the ith frame and frame (i-1). esimo. The correlation between the signal of the "i-1" frame and the signal of the frame (i-2) "nth includes a relation of values between a thirteenth limit and a correlation value of the frame signal (i-2) "nth, and a relation of values between a fifteenth limit and an algebraic codebook contribution of a nth frame excitation signal (i-1). In one embodiment of the present invention, the fifteenth limit may from 0.1 to 0.5 X the excitation signal from frame (i-1) "is selected. For example, the fifteenth limit may be specifically 0.1, 0.2 or 0.5 X the frame excitation signal (jl) "nth, correspondingly determining, according to the correlation between a frame (i-1) signal and a frame (i-2) signal ', the need to correct the frame excitation signal specifically means: checking whether the frame (i-2) signal The tenth and the sign of frame (i-1) 'nth correspond to a tenth condition, where the tenth condition includes: frame (i-2)' nth is a lost frame, a correlation value of the signal of frame (i -2) “nth is greater than the thirteenth limit and the algebraic codebook contribution of the frame excitation signal (i-1) 'nth is less than the fifteenth limit, and if the signal of the frame (i- (2) 'nth and the frame signal (i-1)' and nth meet the tenth condition, determine that the excitation signal of the frame is to be corrected, or if the signal of frame (i-2) If the signal from frame (i-1) does not correspond to the tenth condition, it must be determined that the excitation signal from the frame is not to be corrected. The correction of the excitation signal of the i-esim frame according to the energy stability between the i-es frame and the (i-1) frame es specifically means: determining a fourth correction factor according to the energy stability between the i-th frame and the frame (j-iyesinro 0 where the fourth correction factor is less than 1; and then multiply the excitation signal of the i-th frame by the fourth correction factor to obtain an excitation signal corrected from table i'th.

Figure 7 is a comparison diagram between before correction and after correction of a time domain signal of an ith frame. As shown in Fig. 7, Fig. 7 (a) shows an original time domain signal, and the original time domain signal is a time domain signal sent by an encoder. Figure 7 (b) is a time domain signal synthesized in the prior art. Figure 7 (c) is a time domain signal synthesized in accordance with the present invention. It can be learned by comparing figure 7 (a) to figure 7 (b) that the energy in an ellipse part in figure 7 (b) is much greater than the energy in an ellipse part of the original signal in figure 7 (The). Of course, an excitation signal or an excitation signal with updated status of the i-frame needs to be corrected so that the energy of a signal recovered from the i-frame is closer to the energy of the original signal to obtain the effect. shown in figure 7 (c). (3) Updated status excitation signal correction [0197] In this embodiment, the need to correct the updated status excitation signal from the ith frame can be determined according to the correlation between a frame signal (i-1). ) ~ th is the sign of the frame resume. When the need for correction of the updated frame excitation signal is determined, the updated frame excitation signal of the frame is corrected according to the power stability between the frame and the frame. (i-1) 'the correlation' between the signal from the (i-1) frame 'esimo' and the signal from the esimo frame includes: the correlation between the esimo '(i-1)' frame and the resume frame, and if an excitation signal from frame (i-1) is corrected.

Correspondingly, determining, in accordance with the correlation between a frame (i-1) signal and the frame signal, the need to correct the updated frame excitation signal of the frame. specifically means: checking whether the i-th frame signal and the i-th frame signal (i-1) correspond to an eleventh condition, where the eleventh condition includes: the i-th frame or the (i-1) ) 'esimo is a highly correlated frame and the excitation signal of frame (i-1)' esimo is corrected; and if the i-frame signal and the (i-1) 'frame signal correspond to the eleventh condition, determine that the updated status excitation signal of the i-frame should be corrected, or if the frame i-th and the signal of frame (i-1) "if not meet the eleventh condition, determine that the excitation signal with updated status should not be corrected from frame Γ, the correction of the excitation signal with updated status of frame i-es, mode according to the energy stability between frame i-frame and frame (i-1) -this specifically means: determining a fifth correction factor according to energy stability between frame i-es esimoe the table (i-1) "esimo, where the fifth correction factor is less than 1; and multiplying the updated frame excitation signal of the ith frame by the fifth correction factor to obtain an updated frame corrected excitation signal resume [0199] In this embodiment, if an ith frame is a normal frame , a parameter of the ith frame is obtained by decoding according to a received bit stream, and an excitation signal and an excitation signal with updated status of the ith frame are generated according to a pitch period. , a gain and an algebraic codebook referring to the nth frame and which are obtained by decoding. If an 'esimo' (i-1) frame or an esimo (i-2) 'frame is a lost frame, at least one of a frequency spectrum parameter, the excitation signal or the updated frame excitation signal Ith is also corrected according to interframe ratios and intra-frame ratios between the Ith frame and the first N frames of the Ith frame, and a signal from the Ith frame is synthesized according to a corrected parameter. According to the method in this embodiment, at least one of the frequency spectrum parameter, the excitation signal or the updated status excitation signal of the i-esimo frame is corrected, so that an easy transition of the total energy between adjacent frames can be deployed and voice signal decoding quality can be improved.

Figure 8 is a flow chart of a frame loss compensation processing method according to embodiment 6 of the present invention. As shown in Figure 8, based on mode 5, the method in this mode may also include the following steps: Step 0: Process a decoded signal from an ith frame to obtain a signal correlation value Resume Frame Decode [0202] In one embodiment, normalized autocorrelation processing can be performed on the decode Resume Frame signal. The decoded signal of the nth frame is normalized in a particular range by normalized autocorrelation processing and can be processed using an existing normalized autocorrelation function. In another embodiment, autocorrelation processing, rather than normalized processing, is directly performed on the decoded signal of the nth frame. For example, 100 points are sampled from the decoded frame signal and then autocorrelation processing is performed at points 0 to 98 and points 1 to 99 to obtain a signal correlation value. decoded from the ith frame. Of course, 50 points can be selected from each one of a nth frame signal (jl) and a nth frame signal, and there are 100 points in total. Then autocorrelation processing is performed as follows. above for obtaining a correlation value of the decoded signal of the ith frame.

[0203] Step 602: Determine the correlation of an i-frame signal according to any or any combination of a decoded i-frame correlation value, a relationship of values between the pitch periods of all the subframes of the frame are, for example, a slope value of the nth frame spectrum or a zero crossing rate of the nth frame.

For example, when the signal correlation of the ith frame is determined according to a correlation value of the decoded signal of the ith frame, a limit is generally set. If a correlation value of the decoded frame signal is greater than the threshold, it is determined that the correlation of the signal frame already is high, or if a correlation value of the decoded frame signal is less than the limit, it is determined that the correlation of the i-esimo frame signal is low.

[0205] Step 603: Determine the power of the i-th frame according to the decoded signal of the i-th frame and determine the energy stability between the power of the i-th frame and the power of a frame (i-1 th according to the power of the ith frame and the energy of the frame (i-1, and / or determining the energy of each subframe of the ith frame according to the decoded signal of the ith frame, and determining the Energy stability between the subframes of the i-frame is according to the energy of each subframe of the i-frame.

In this embodiment, to estimate a one-frame signal (t + 1), the correlation of signals and the energy stability between an ith frame and a (i-1) frame and / or the stability of intraframe energy of the ith frame is determined. In this embodiment, when a parameter from each frame is estimated, the correlation and energy stability for a previous frame are used.

Figure 9 is a schematic structural diagram of a frame loss compensation processing apparatus according to embodiment 7 of the present invention. As shown in Fig. 9, the frame loss compensation processing apparatus provided in this embodiment includes: a lost frame determination module 11, an estimation module 12, a obtaining module 13, a generation module 14 and a module of signal synthesis 15.

The lost frame determination module 11 is configured to determine, using a lost frame flag bit, whether a frame is already a lost frame.

[0209] Estimation module 12 is configured to: when i-esimofor frame is a lost frame, estimate a parameter of i-es frame according to at least one of an interframe relationship between the first N frames of the frame nth or an intraframe relationship between the first N frames of the nth frame. The interframe relationship between the first N frames includes at least one of the correlation between the first N frames or the energy stability between the first N frames. The intra-frame relationship between the first N frames includes at least one of the correlation between the first N frames. -frames between the first N frames or the energy stability of the inter-subframes between the first N frames. The nth frame parameter includes a frequency spectrum parameter, a pitch period, and a gain, and N is a number. integer greater than or equal to 1.

[0210] Obtaining module 13 is configured to obtain an algebraic codebook from the nth frame.

[0211] Generation module 14 is configured to generate an excitation signal from the ith frame according to the pitch period and the gain relative to the ith frame which are obtained by the estimation module by estimation and of the algebraic codebook referring to the ith table and which is obtained by the obtainment module.

Signal Synthesis Module 15 is configured to synthesize a signal from the ith frame according to the frequency spectrum parameter for the ith frame and which is obtained by the estimation module by means of estimation and the excitation signal for frame i and which is generated by the generation module. (1) Estimation of the frequency spectrum parameter of the i th frame [0213] The frequency spectrum parameter of the i th frame is obtained by the estimation module 12 by estimation according to the interquartile relationship between the first Table N of the Resume Table The estimation module is specifically configured to: determine the weight of a frequency frame parameter of a nth frame (jl) and the weight of a preset frequency spectrum parameter of the nth frame according to with the correlation between the first N frames of Table I, and perform a weighting operation on the frequency spectrum parameter of Table (I-1) "and the default frequency spectrum parameter of Table I according to with the weight of the frequency spectrum parameter of table (i-1) 's and the weight of the default frequency spectrum parameter of table i' to obtain the frequency spectrum parameter of the i'th frame.

Optionally, the correlation includes a ratio of values between a second boundary and a spectrum slope parameter of a "frame" (i-1) signal, a ratio of values between a first boundary and a normalized value. autocorrelation of the nth frame (i-1) signal, and a relation of values between a third threshold and a deviation of a pitch period of the nth frame (i-1) signal.

Correspondingly, estimation module 12 is specifically configured to: if the signal from frame (i-1) -this corresponds to at least one of a first condition, a second condition, and a third condition, determine that the weight of frame frequency spectrum parameter (i-1) is the first weight, and the weight of the default frame frequency spectrum parameter is a second weight when the first weight is greater than the second weight, the first condition is: the normalized autocorrelation value of the (i-1) 's frame signal is greater than the first bound, the second condition is: the signal's spectrum slope parameter of the (i-1) This is greater than the second limit and the third condition is: the deviation of the signal period from frame (ii) is less than the third limit; or if the signal from frame (i-1) "nth does not match a first condition, a second condition, or a third condition, determines that the weight of the frequency spectrum parameter of frame (i-1)" th is a second The weight and weight of the default frequency spectrum parameter of the ith frame is a first weight, where the first weight is greater than the second weight. (2) Estimation of the pitch frame pitch period [0216] pitch of the ith frame is obtained by the estimation module 12 by estimation according to the correlation between the first N frames of the ith frame and the inter-subframe correlation between the first N frames of the ith frame. Correlation includes a ratio of values between a fifth boundary and a normalized autocorrelation value of a nth frame (i-2) signal, a relation of values between a fourth boundary and a deviation of a pitch period from the Table (i-2), and a ratio of values between the fourth limit and a deviation of a pitch period from a single frame (M) signal.

Correspondingly, estimation module 12 is specifically configured to: if the pitch period deviation of the nth frame (i-1) signal is less than the fourth threshold, determine a period deviation value of frame signal pitch (i-1) 'nth according to the frame signal pitch period (i-1)' nth, and determine a frame nth signal pitch period according to the value of deviation from the pitch frame signal period (i-1) 'nth and the pitch signal frame period (i-1)' nth, where the pitch period of the ith frame signal includes a pitch period of each subframe of the nth frame, and the deviation value of the pitch signal period of the frame (i-1) is the average value of the differences between the pitch periods of all adjacent subframes of the frame (i-1). 1) "eso; or if the deviation of the pitch (i-1) 's signal period is greater than or equal to the fourth limit, the normalized autocorrelation value of the (i-2)' s signal is greater than the fifth limit, and the deviation from the frame signal pitch period (i-2) "is less than the fourth limit, determine a deviation from the frame signal pitch period (i-2)" and the frame signal (i-1) "nth according to frame pitch period (i-2)" nth and frame pitch period (i-1) 'nth, and determine a signal pitch period of the ith frame according to the pitch signal period of the frame (i-1) 'th and the deviation value of the pitch signal of the frame signal (i-2)' th and the signal of the frame (i -1) "th.

Optionally, the estimation module 12 determines the pitch period shift value Pv of the frame (i-1) signal according to the following formula: [0219], where P ^ is a period pitch of a nth subframe of frame (i-1) "th, ej = 0, 1,2, 3.

Correspondingly, estimation module 12 determines the pitch period of the nth frame signal according to the following formula: [0221], where P ® is a pitch period of a third frame subframe ( i-1) "th," is the pitch shift value of the signal from frame (i-1) (th and ParU) is a pitch period of a twelfth frame from the th frame.

Optionally, estimation module 12 determines the pitch period shift value Pv of frame signal (i-2) "nth and frame signal (i-1)" nth according to the following [0223] where P is a pitch period of a nth frame subframe (i-2) ~ th, p (l) (n) is a pitch period of a nth frame subframe ( (1) ~ th, m = 2,3 and n = 0,1.

Correspondingly, the estimation module 12 determines the pitch period of the i-th frame signal according to the following formula: [0225] Pair (x) = P (~ '\ 3) + (x + \ ) * pvi x = 0, 1, 2,3j where P ® is a pitch period of a third frame subframe (i-1) 'nth, ^ is the deviation value of the frame signal pitch period (i- 2) 'nth is the sign of frame (i-1)' is nth, and P ™ ^ is a pitch period of a subframe> nth of frame nth. (3) Estimation of the i-th frame gain [0226] The i-th frame gain includes an adaptive codebook gain and an algebraic codebook gain, and the i-th frame gain is obtained. by estimation module 12 by estimating according to the correlation between the first N frames of the ith frame and the energy stability between the first N frames of the ith frame.

Estimation module 12 is specifically configured to: determine the adaptive codebook gain of the ith frame according to an adaptive codebook gain of an "i" frame or a fixed value default, a nth frame correlation (i-1) and a nth frame sequence number in multiple consecutive lost frames, determine the weight of a frame algebraic codebook gain (i-1) 'nth and the weight of a gain of a voice frame detection VAD frame according to the energy stability of the frame (i-1) is e, and performing a weighting operation of the frame algebraic codebook gain (i -1) ~ nth and the VAD frame gain according to the weight of the algebraic codebook gain of frame (i-1) 'nth and the weight of the VAD frame gain to obtain the book gain of algebraic codes from the nth table.

More stable frame energy (i-1) 'nth indicates a higher weight of the algebraic codebook gain of frame (i-1)' nth, or the weight of the VAD frame gain increases correspondingly to as the number of consecutive lost frames increases.

Optionally, before performing a weighting operation on the algebraic codebook gain of frame (i-1) and the VAD frame gain according to the weight of the algebraic codebook gain of the (i-1 seventh and the VAD frame gain weight, to obtain the algebraic codebook gain of frame i, the estimation module 12 is also configured to: determine a first correction factor according to with a coding and decoding rate; and correcting the gain of the algebraic codebook of frame (i-1) "esimo using the first correction factor. (4) Obtaining the algebraic codebook of frame i'th [0230] Obtaining module 12 is specifically configured to: obtain the ith frame algebraic codebook by random noise estimation, or determine the ith frame algebraic codebook according to the codebook algebraic of the first frames N of the ith frame.

[0231] Obtaining module 12 is also configured to: determine the weight of an algebraic codebook contribution from the ith frame according to any of a pitch offset of a frame (i-1) "nth, a correlation of a frame signal (i-1) - nth, a tilt rate value of the frame (i-1)" nth or a zero crossing rate of a frame (i-1) “Nth, or determine the weight of a contribution from the i-th frame algebraic codebook by performing a weighting operation on any combination of a deviation from a (i-1) th frame pitch period, a correlation a frame (i-1) signal "nth, a frame spectrum tilt rate value (i-1)" ninth or a frame zero crossing rate (i-1) nth; and performing an interpolation operation on an excitation signal with updated frame (i-1) status to determine a frame adaptive codebook.

[0232] Generation module 14 is specifically configured to: determine the contribution of the ith-frame algebraic codebook according to a product obtained by multiplying the ith-frame algebraic codebook by the gain of the ith-frame algebraic codes of table i; determining a contribution from the i-frame adaptive codebook according to a product obtained by multiplying the i-frame adaptive codebook by the gain of the i-frame adaptive codebook; and carry out a weighting on the contribution of the i-table algebraic codebook and the contribution of the table-i adaptive codebook according to the weight of the contribution of the i-table algebraic codebook and a weight of the contribution of the adaptive codebook of the ith frame to determine the excitation signal of the ith frame, where a weight of the adaptive codebook is 1.

[0233] The apparatus in this embodiment can be configured to execute methods in modalities 1 through 4. Thus, specific modes of implementation and technical effects in this embodiment are similar to those in modalities 1 through 4, and details thereof will not be repeated repeatedly. described here.

Figure 10 is a schematic structural diagram of a frame loss compensation processing apparatus according to embodiment 8 of the present invention. As shown in Figure 10, based on the apparatus shown in Figure 9, the apparatus in this embodiment also includes: a decoding module 16, an evaluation module 17 and a correction module 18.

[0235] The nth frame is a normal frame in this mode. The decoding module 16 is configured to obtain the ith frame parameter by decoding according to a received bit stream. The nth frame parameter includes the frequency spectrum parameter, pitch period, gain, and algebraic codebook.

[0236] Generation module 14 is also configured to generate the i-th frame excitation signal and an i-th frame updated status excitation signal according to pitch period, gain and codebook algebraic data for table i and which are obtained by decoding module 16 by means of decoding.

[0237] Evaluation module 17 is configured to: when a 'nth frame' or one (i-2) 'nth frame is a lost frame, determine according to at least one of the interframe relationships or intra-frame relationships between the ith frame and the first N frames of the ith frame, the need to correct at least one of the frequency spectrum parameter, the excitation signal or the updated frame excitation signal Fifth, the inter-frame relationship includes at least one of the correlation between the ith frame and the first N frames of the ith frame or the energy stability between the ith frame and the first N frames of the frame. esimo. The intraframe ratio includes at least one of the interframe correlation between the ith frame and the first N frames of the ith frame or the energy stability of the interframes between the ith frame and the first frames. No of the nth frame.

[0238] Correction module 18 is set to: when the evaluation module 17 determines that it must correct at least one of the frequency spectrum parameter, excitation signal or excitation signal with updated status of the i'th frame. , correct at least one of the frequency spectrum parameter, excitation signal or excitation signal with updated status of the ith frame according to at least one of the interframe relations or the interframe relations between the ith frame. nth and the first frames N of the nth frame.

[0239] Signal synthesis module 15 is also configured to: synthesize the signal from the ith frame according to a correction result performed by the correction module on at least one of the frequency spectrum parameter, the frequency signal. excitation or the excitation signal with updated status of the i'th frame; or when the evaluation module 17 determines that it should not correct the frequency spectrum parameter, the excitation signal, or the updated status excitation signal of the ith frame, synthesize the ith frame signal according to the parameter. frequency spectrum, the excitation signal and the excitation signal with updated status of the i'th frame. (1) Correction of the frequency spectrum parameter of Table I

Optionally, the evaluation module 17 is configured to determine, according to the ith frame correlation, the need to correct the frequency spectrum parameter of the ith frame. When the evaluation module 17 determines that it must correct the frequency spectrum parameter of the ith frame, the correction module 18 is configured to: correct the frequency spectrum parameter of the ith frame according to the spectrum parameter. frame frequency parameter and a frame frequency spectrum parameter (i-1) ~ or correct the frame frequency spectrum parameter according to the frame frequency spectrum parameter is a preset frequency spectrum parameter of the table res, mo.

The correlation of the nth frame includes a ratio of values between a sixth limit and one of the two frequency spectrum parameters that corresponds to an index of a minimum value of the difference between the adjacent frequency spectrum parameters of the nth frame. , a value relationship between a seventh limit and the minimum value of the difference between the adjacent frequency spectrum parameters of Table I and a ratio of values between an eighth limit and the minimum value index of the difference between the adjacent parameters frequency spectrum of the nth frame.

The evaluation module 17 is specifically configured to: determine the difference between adjacent frequency spectrum parameters of the nth frame, where each difference corresponds to an index and the frequency spectrum parameter includes an immittance spectral frequency. (ISF) or a line spectral frequency (LSF); check whether the difference between the adjacent frequency spectrum parameters of the nth frame corresponds to at least one of a fourth condition or a fifth condition, where the fourth condition includes: one of the two frequency spectrum parameters that corresponds to the index of the Minimum value of the difference between adjacent frequency spectrum parameters of the i-th frame is less than the sixth limit and the fifth condition includes: an index value of the minimum value of difference between adjacent frequency spectrum parameters of the i-th frame nth is less than the eighth limit and the minimum difference is less than the seventh limit; and if the difference between adjacent frequency spectrum parameters of frame i corresponds to at least one of the fourth or fifth condition, determine the correction of the frequency spectrum parameter of frame i, or if the difference between adjacent frequency spectrum parameters of the ith frame do not correspond to the fourth condition or the fifth condition, determine the non-correction of the frequency spectrum parameter of the ith frame.

[0243] Correction module 18 is specifically configured to: determine a corrected frequency spectrum parameter of the i-th frame according to a weighting operation performed on the frequency spectrum parameter of frame (i-1) _th and the frequency spectrum parameter of the ith frame; or determining a corrected frequency spectrum parameter of the ith frame according to a weighting operation performed on the frequency spectrum parameter of the ith frame and the default frequency spectrum parameter of the ith frame.

Optionally, the evaluation module 17 is configured to determine, according to the correlation between frame I and frame (M), the need to correct the frequency spectrum parameter of frame i. 'esimo. When the evaluation module 17 determines that it must correct the frequency spectrum parameter of the ith frame, the correction module 18 is configured to: correct the frequency spectrum parameter of the ith frame according to the spectrum parameter. frame frequency parameter and a frame frequency spectrum parameter (i-1) 'esim ° or correct the frame frequency spectrum parameter according to the frame frequency spectrum parameter is a predefined frequency spectrum parameter of the ith table. The correlation between the nth frame and the ninth frame (i-1) includes a relation of values between a ninth limit and the sum of the differences between the frequency spectrum parameters that correspond to some or all of the equal indices of the frame. (i-1) ~ th is the table i "th.

Evaluation module 17 is specifically configured to: determine the difference between adjacent frequency spectrum parameters of the nth frame, where each difference corresponds to an index and the frequency spectrum parameter includes an immittance spectral frequency. (1SF) or a line spectral frequency (LSF); determine whether the frame frequency spectrum parameter and frame frequency spectrum parameter (i-1) correspond to a sixth condition, where the sixth condition includes: the sum of the differences between the spectrum parameters of frequency corresponding to some or all of the same indices of table (i-1) -th and the table is already greater than the ninth limit; and if the ith frame frequency spectrum parameter and the ith frame frequency spectrum parameter (i-1) 'meet the sixth condition, determine the correction of the ith frame frequency spectrum parameter, or If the ith frame frequency spectrum parameter and the ith frame frequency spectrum parameter (i-1) do not meet the sixth condition, determine the non-correction of the ith frame frequency spectrum parameter.

[0246] Correction module 18 is specifically configured to: determine a corrected frequency spectrum parameter of the i-th frame according to a weighting operation performed on the frequency spectrum parameter of frame (i-1) “nth and the frequency spectrum parameter of the nth frame; or determine a frequency frame corrected frequency parameter according to a weighting operation performed on the frequency frame frequency parameter and the default frame frequency frequency parameter (2) Signal Correction optionally, the evaluation module 17 is configured to determine, according to the correlation between the ith frame and the ith frame (i-1) and the energy stability between the Frame I and Frame (I-1) is the need to correct the excitation signal from Frame I. When the evaluation module 17 determines that it should correct the excitation signal of the nth frame, the correction module 18 is configured to correct the excitation signal of the nth frame according to the power stability between the nth frame. ese and frame (i-1) "ese.

The evaluation module 17 is specifically configured to: determine a pre-synthesized frame signal according to the frame excitation signal and frame frequency spectrum parameter; determining whether an absolute value of the difference between the power of the i-th frame pre-synthesized signal and the power of a th-frame's synthesized signal (i-1) "is greater than one tenth limit; and whether the absolute value of the difference between the preset synthesized signal energy of the i-th frame and the synthesized signal energy of the (1-1) "tenth frame is greater than the tenth limit, determine the correction of the excitation signal of the i th frame", or if the absolute value of the difference between the preset synthesized signal energy of the ith frame and the synthesized signal energy of the ith frame is less than or equal to the tenth limit, determine the non-correction of the excitation signal of the or determine whether a relationship between the power of the preset synthesized signal of frame I and the power of a synthesized signal of frame (jl) is greater than an eleventh limit, where the eleventh limit is greater than 1, and if the relationship between the preset signal energy of the ith frame and the synthesized signal energy of the frame (i-1 is greater than the eleventh threshold, determine the correction of the ith frame excitation signal, or if the relationship between the pre-signal energy -th frame synthesized and the synthesized signal energy of frame (i-1) -th is less than or equal to the eleventh threshold, determine the non-correction of the frame excitation signal Γ; or determine whether a relationship between the energy of a synthesized signal from frame (i-1) -th and the energy of a synthesized signal from frame-i is less than a twelfth threshold, where the twelfth threshold is smaller that 1; and if the ratio between the power of the synthesized signal from frame (i-1) "esimo and the power of the synthesized signal from frame i" is less than the twelfth threshold, determine the correction of the excitation signal from frame i 'nth, or if the ratio between the pre-synthesized signal energy of frame (i-1)' and the nth signal synthesized signal energy is greater than or equal to the twelfth threshold, determine the non-correction of the signal excth frame excitation [0249] The correction module 18 is specifically configured to: determine a second correction factor according to the energy stability between the i-esimo frame and the "esimo" (i-1) frame, where the second correction factor is less than 1; and multiplying the excitation signal from the ith frame by the second correction factor to obtain a corrected excitation signal from the ith frame. The second correction factor can be a ratio between the frame energy (i-1) "esimo and the frame energy reso" or the second correction factor is a relationship between the energy of the same number of subframe frames (i -1) ~ th and tenth frame [0250] Optionally, the evaluation module 17 is configured to determine, according to the correlation of a signal from frame (i-1) -th, the need to correct the signal When the evaluation module 17 determines that it must correct the excitation signal of the nth frame, the correction module 18 is configured to correct the excitation signal of the nth frame according to energy stability between frame i and frame (i-1). Frame Signal Correlation (i-1) "nth includes a value ratio between a thirteenth threshold and a Frame Signal Correlation value (i-1)" nth, and a value ratio between a fourteenth and a a deviation of a pitch signal from frame (i-1) "th '[0251] Evaluation module 17 is specifically configured to: verify that frame (i-1)" th signal corresponds to a seventh condition, where the seventh condition is: frame (i-1) -th is a lost frame, a signal correlation value of frame (i-1) -th is greater than the thirteenth limit, and the pitch period offset of the sign in table (i-1) 'th is less than the fourteenth; and if the signal from frame (i-1) "nth meets the seventh condition, determines the correctness of the excitation signal from frame nth, or if the signal from frame (i-1)" nth does not correspond to the seventh condition, determine the non-correction of the excitation signal of the nth frame.

[0252] Correction module 18 is specifically configured to: determine a third correction factor according to the energy stability between frame i-esimo and frame (i-1) "esimo, where the third correction factor is smaller which is 1, and multiplying the excitation signal from the ith frame by the third correction factor to obtain a corrected excitation signal from the ith frame.

Optionally, the evaluation module 17 is configured to determine, according to the correlation between the i-frame signal and a "n-frame" signal (i-1), the need to correct the Frame I excitement When the evaluation module 17 determines that it should correct the frame excitation signal, the correction module 18 is configured to correct the frame excitation signal according to the stability of the frame. energy between frame i and frame (i-1). the correlation between the nth frame signal and the nth frame signal (i-1) includes a relationship of values between a thirteenth threshold and a nth th frame (i-1) signal correlation value, and a relation of values between a fourteenth and a deviation of a pitch period of the signal from the nth frame.

[0254] Evaluation module 17 is specifically configured to: verify that the signal from frame (i-1) "esimo and the signal from frame i" correspond to an eighth condition, where the eighth condition includes: frame (i -1) "nth is a lost frame, a frame signal correlation value (i-1) - nth is greater than the default thirteenth threshold, and the pitch period deviation of the nth frame signal is smaller that the default fourteenth limit; and if the signal from frame i-1 and the signal from frame i correspond to the eighth condition, determine the correction of the excitation signal from frame i, or if the signal from frame i-1 If the ith frame signal does not match the eighth condition, determine the non-correction of the ith frame excitation signal.

[0255] Correction module 18 is specifically configured to: determine a third correction factor according to the energy stability between frame i-esimo and frame (i-1) 'esimo, where the third correction factor is lower that 1; and multiplying the ith frame excitation signal by the third correction factor to obtain a corrected ith frame excitation signal.

Optionally, the evaluation module 17 is configured to determine, according to the correlation between a frame (i-1) signal and a frame (i-2) signal, the need for correcting the excitation signal of the tenth frame. When the evaluation module 17 determines that it must correct the excitation signal of the nth frame, the correction module 18 is configured to correct the excitation signal of the nth frame according to The energy stability between the i-frame and the i-1 frame is the correlation between the e-frame's i-1 signal and the i-frame's signal includes: relationship of values between a thirteenth threshold and a correlation value of the nth frame (i-2) signal, and whether a nth frame excitation signal (i-1) is corrected.

The evaluation module 17 is specifically configured to: verify that the nth frame signal (i-2) and the nth frame signal (i-1) correspond to a ninth condition, where the ninth condition includes: (i-2) 'nth frame is a lost frame, a signal correlation value of frame (i-2)' nth is greater than the thirteenth limit and frame excitation signal (i-1) 'nth is corrected, and if the signal from frame (i-2) 'th and the signal from frame (i-1)' th corresponds to the ninth condition, determine the correction of the excitation signal from frame i 'or of frame (i-2) 'nth and the signal of frame (i-1)' nth did not match the ninth condition, determine the non-correction of the excitation signal of frame i 'nth.

[0258] Correction module 18 is specifically configured to: determine a fourth correction factor according to the energy stability between frame i-esimo and frame (i-1) 'esimo, where the fourth correction factor is lower that 1; and multiplying the excitation signal from the nth frame by the fourth correction factor to obtain a corrected excitation signal from the nth frame.

Optionally, the evaluation module 17 is configured to determine, according to the correlation between a frame (i-1) signal and a frame (i-2) signal, the need for Correcting the Resume Frame Excitation Signal When the Evaluation Module 17 determines that it should correct the Resume Frame Excitation Signal, the Correction Module 18 is configured to correct the ith Frame Excitation Signal according to the stability. of energy between frame i and frame (i-1). correlation between frame (i-1) signal 'and frame (i-2)' thirteenth limit and a "(i-2)" signal correlation value of the nth frame, and a relation of values between a fifteenth limit and a contribution of the algebraic codebook of a frame (i-1) signal "esimo.

[0260] Evaluation module 17 is specifically configured to: verify that the signal from frame (i-2) -th and the signal from frame (i-1) -this correspond to a tenth condition, where the tenth condition includes: frame (i-2) is a lost frame, a signal correlation value of frame (i-2) is greater than the thirteenth limit, and the algebraic codebook contribution of the table (i-1) "nth is less than the fifteenth limit; and if the sign of table (i-2) 'nth and the sign of table (i-1)" nth meet the tenth condition, determine the correction of the tenth frame excitation signal, or if the tenth frame (i-2) signal and the tenth frame (i-1) signal do not match the tenth condition, determine that the frame excitation signal is not corrected That's right.

[0261] Correction module 18 is specifically configured to: determine a fourth correction factor according to the energy stability between frame i-esimo and frame (i-1) 'esimo, where the fourth correction factor is lower that 1; and multiplying the ith frame excitation signal by the fourth correction factor to obtain a corrected ith frame excitation signal. (3) Correction of excitation signal with updated frame status Γ th [0262] The evaluation module 17 is configured to determine, according to the correlation between a "th" frame signal (i-1) and the frame signal i "nth, the need to correct the excitation signal with updated status of the nth frame. When the evaluation module 17 determines that it must correct the updated status excitation signal of the i-th frame, the correction module 18 is configured to correct the updated status excitation signal of the i-th frame according to the stability of energy between frame I and frame (I-1) " The correlation between the signal from frame (i-1) "eso and the signal from frame i" includes: the correlation between frame (i-1) "esimo and frame i, and whether an excitation signal from Table (i-1) is corrected.

[0283] Evaluation module 17 is specifically configured to: verify that the i-th frame signal and the (i-1) -th frame signal correspond to an eleventh condition, where the eleventh condition includes: the frame i-esimo or frame (i-1) 'esimo is a highly correlated frame and the excitation signal of frame (i-1)' esimo is corrected; and if the ith frame signal and the ith frame signal (i-1) match the eleventh condition, determine the correction of the excitation signal with updated status of the ith frame, or if the signal from the i frame -th and Frame Signal (i-1) "If not meet the eleventh condition, determine non-correction of excitation signal with updated status of Frame i" ith [0264] Correction module 18 is specifically configured for: determine a fifth correction factor according to the energy stability between table i-esimo and table (i-1) “esimo, where the fifth correction factor is less than 1; and multiplying the i-frame updated status excitation signal by the fifth correction factor to obtain an i-frame updated status corrected excitation signal.

[0265] For specific implementation modes of the functional modules of the frame loss compensation processing apparatus provided in mode 7 and mode 8, refer to the related descriptions of the methods shown in modalities 1 to 6. Details of the same will not be repeated repeatedly. described herein, Figure 11 is a schematic diagram of a physical structure of a frame loss compensation processing apparatus according to embodiment 9 of the present invention. As shown in Fig. 11, a frame loss compensation processing apparatus 200 includes a communication interface 21, a processor 22, a memory 23, and a bus 24. The communication interface 21, processor 22 and memory 23 are interconnected using bus 24. Bus 24 can be a peripheral component interconnect bus (peripheral component interconnect, PCI to shorten), an extended industry standard architecture bus (extended industrial standard architecture, EISA to shorten) or the like. The bus may include an address bus, a data bus, a control bus, and the like. For ease of representation, bus 24 is illustrated using only one thick line in figure 11. However, this does not mean that there is only one bus or only one type of bus. Communication interface 21 is configured to deploy communication between a database access device and another device (such as a client, a read / write database, or a read-only database). Memory 23 may include random access memory (random access memory, shortening RAM) and may also include nonvolatile memory (nonvolatile memory), such as at least one magnetic disk memory.

Memory 22 executes program code stored in memory 23 to implement methods of modalities 1 through 6.

The aforesaid processor 22 may be a generic processor, including a central processing unit (Central Processing Unit, CPU to shorten), a network processor (Network Processor, NP to shorten) and the like; or it may be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable logic gate (FPGA) or other programmable logic device, a discrete port or a transistor logic device, or a discrete hardware.

Those skilled in the art will understand that all or some of the steps of the method embodiments may be implemented by a program that instructs relevant hardware. The program can be stored on a computer readable storage medium. When the program is executed, the steps of the method embodiments are performed. The above storage medium includes: any medium that can store a program code, such as ROM, RAM, magnetic disk or optical disk.

Finally, it should be noted that the above embodiments are merely intended to describe the technical solutions of the present invention and not to limit the present invention. While the present invention has been described in detail with reference to the above embodiments, those skilled in the art will understand that they may make modifications to the technical solutions described in the above embodiments or make equivalent substitutions in some or all of the technical characteristics thereof without escape the scope of the technical solutions of the embodiments of the present invention.

Claims (18)

A frame loss compensation processing method, characterized in that it comprises: determining (101) using a lost frame flag bit whether an ith frame is a lost frame; If the ith frame is a lost frame, estimate (102) a parameter of the ith frame according to at least one of an interframe relationship between the first N frames of the ith frame or an intraframe relationship. among the first N frames of the nth frame, in which the interframe relationship between the first N frames comprises at least one of the correlation between the first N frames or the energy stability between the first N frames, the intra-frames between the first N-frames comprises at least one of the inter-subframe correlation between the first N-frames or the energy stability of the inter-subframes between the first N-frames, the j-frame parameter comprises a parameter of frequency spectrum, a pitch period and a gain, and N is an integer greater than or equal to 1; obtain (103) an algebraic codebook from table i; generating (104) an excitation signal from the ith frame according to the pitch period and gain referring to the ith frame and which are obtained by estimation and the algebraic codebook obtained from the ith frame; and synthesizing (105) a signal from the ith frame according to the frequency spectrum parameter referring to the ith frame and which is obtained by estimation and the generated excitation signal from the frame res. Method according to claim 1, characterized in that if the ith frame is a normal frame, the method also comprises: obtaining (501) the ith frame parameter by decoding according to a received bit stream, wherein the ith frame parameter comprises the frequency spectrum parameter, pitch period, gain, and algebraic codebook; generate (502) the i-th frame excitation signal and an i-th frame updated status excitation signal according to the pitch period, gain and algebraic codebook for the i-th frame and which are obtained by decoding; if a frame (i-1) or frame (i-2) is a lost frame, determine (503) according to at least one of the interframe relationships or intraframe relationships between the i- and the first N frames of the nth frame, the need to correct at least one of the frequency spectrum parameter, the excitation signal or the updated status excitation signal of the nth frame, in which the ratio of The interframe comprises at least one of the correlation between frame I and the first frames N of frame I or the energy stability between frame I and the first frames N of frame I, and the ratio of intraframe comprises at least one of the inter-subframe correlation between the ith frame and the first N frames of the ith frame or the energy stability of the inter-frames between the ith frame and the first N frames of the Γth table, _ 3 “when it is determined that If at least one of the frequency spectrum parameter, the excitation signal or the updated status excitation signal of the ith frame is corrected, correct (504) at least one of the frequency spectrum parameter, the frequency signal excitation or the updated status excitation signal of frame i-es, according to at least one of the interframe relationships or intra-frame relationships between the i-frame and the first N frames of the i-frame; and synthesizing (506) the i-frame frame signal according to a correction result of at least one of the frequency spectrum parameter, the excitation signal or the updated frame-state excitation signal; or when the non-correction of the frequency spectrum parameter, excitation signal or excitation signal with updated status of the ith frame is determined, synthesize (505) the ith frame signal according to the spectrum parameter. frequency, the excitation signal and the excitation signal with updated status of the i'th frame. Method according to claim 1, characterized in that the frequency spectrum parameter of the ith frame is obtained by estimation according to the inter-frame relationship between the first N frames of the ith frame; and the frequency spectrum parameter of the ith frame is obtained by estimation as follows: determining (201) the weight of a frequency spectrum parameter of a frame (i-1 and the weight of a frequency parameter). preset frequency spectrum of the ith frame according to the correlation between the first N frames of the ith frame, and (202) perform a weighting operation on the frequency spectrum parameter of the ith frame (i-1) and the default frequency spectrum parameter of the ith frame according to the weight of the frequency spectrum parameter of the (i-1) ith frame and the weight of the default frequency spectrum parameter of the ith frame to obtain the frequency spectrum parameter of the tenth frame, in which the correlation comprises a ratio of values between a second boundary and a spectrum slope parameter of a signal from the nth frame (i-1), a ratio of values between a firstthreshold and a normalized frame signal autocorrelation value (1-th, and a relationship of values between a third threshold and a pitch shift of the signal from the (i-1) frame; and determining the weight of a frequency frame parameter of a nth frame (i-1) and the weight of a preset frequency spectrum parameter of the nth frame according to the correlation between the first N frames of frame i It comprises: if the signal from frame (i-1) 'corresponds to at least one of a first condition, a second condition, and a third condition, it determines that the weight of the frame frequency spectrum parameter (i-1) 1) -th is a first weight and whereas the weight of the preset frequency spectrum parameter of table i is a second weight, where the first weight is greater than the second weight, the first condition is: the normalized value of autocorrelation of frame (i-1) -th signal is greater than the first limit, the second condition is: the signal slope parameter of frame (i-1) -th is greater than the second and third limit condition is: the deviation of the signal pitch period from (i -1) ~ th frame is less than the third limit; or if the signal from (i-1) -th frame does not correspond to a first condition, a second condition, or a third condition, determines that the weight of the frequency spectrum parameter of frame (i-1) -th is a second weight and that the weight of the preset frequency spectrum parameter of Table I is a first weight, in which the first weight is greater than the second weight. Method according to claim 1, characterized in that the pitch period of the ith frame is obtained by estimation according to the correlation between the first N frames of the ith frame and the correlation of inter subframes between the first N frames of the nth frame, and the correlation comprises a relation of values between a fifth boundary and a normalized autocorrelation value of a signal of a nth frame (i — 2), a relation of values between a fourth boundary and a deviation of a pitch signal period from the nth frame (i-2), and a relation of values between the fourth boundary and a deviation of a pitch period from a signal in a frame (i-1) ) "th and the pitch period of the nth frame is obtained by estimation as follows: if the deviation of the pitch period of the nth frame (i-1) signal is less than the fourth threshold, determine (302) a deviation value of the frame signal pitch period (i-1) -this according to the frame signal pitch period (i-1) "nth, and determine (302) a frame signal pitch period i-th in accordance with the deviation value of the frame signal pitch period (i-1) 'th and the frame signal pitch period (i-1)' es, in which the pitch period of the ith frame signal comprises a pitch period of each subframe of the ith frame, and the deviation value of the pitch signal of the frame (jl) is an average value of the differences between the pitch periods of all adjacent subframes of frame (i-1) "es"; or if the pitch period deviation of the frame (i-1) 's signal is greater than or equal to the fourth threshold, the normalized value of au signal correlation of frame (i-2) "nth is greater than the fifth boundary and deviation of the pitch pitch of the (i-2) signal" nth is less than the fourth boundary, determine (303) a value of deviation of pitch signal from frame (i-2) 'nth and frame signal (i-1)' nth according to pitch period of frame (i-2) ith signal and period of frame signal pitch (i-1) "and determine (303) a frame pitch signal period according to the frame (i-1) signal pitch period and the value deviation of the pitch period of the nth frame signal (i-2) and the nth frame signal (i-1); wherein the deviation value of the pitch (pv) signal period of frame (i-1) is determined according to the following formula: ^ V = (2)) + ^ - ^ (2) (1)) + ^^ (1) (0)) / 35 no! P ^ is a pitch period of a subframe j "th of frame (i-1)" th and j = 0, 1,2,3; and the pitch period of the i-th frame signal is determined according to the following formula: /'[-C/')=p'_,'(3)+0'+lced.;=ftU3i on w | ^ - "(3) is a pitch period of a third subframe of frame (i-1)" nth, ^ is the deviation value of the pitch signal period of frame (i-1) 'nth and ParU) is a pitch period of a nth subframe of the nth frame; or wherein the deviation value of the pitch period of the (i-2) -th frame signal and the (i-1) 'seventh frame signal is determined according to the following formula: Pv = {p ^ ( 3) - pP2 ^ (2)) + (0) - (3)) + (/? (”” ') (1) (0)) / 3 ^ nQ qugj P (m) is a pitch period of one nth frame subframe (i-2) 'th, ^ ^ is a pitch period of a nth frame subframe (i-1) "th, m = 2, 3 and n = 0, 1; and the pitch period of the i-th frame signal is determined according to the following formula: ΡαΛχ) ^ Ρ (~~]) (Ρ) + (* + Ψρν, x = 0, 1, 2,3j nQ w | FKJ and a pitch period of a third subframe of frame (i-1) ~ th, is the value of deviation from the pitch period of signal of frame (i-2) ~ th and signal of frame (i-1) ~ ith, and is a pitch period of a nth> frame of the ith frame. Method according to claim 1, characterized in that the ith frame gain comprises an adaptive codebook gain and an algebraic codebook gain and the ith frame gain is obtained by according to the correlation between the first N-frames of the ith frame and the energy stability between the first N-frames of the ith frame and the gain of the ith frame is obtained by estimation as follows : determine (401) the adaptive codebook gain of the ith frame according to an adaptive codebook gain of a ith frame or a predefined fixed value, a frame correlation (i- 1) 'nth and a sequence number of the nth frame in multiple consecutive lost frames; determine (402) the weight of an algebraic codebook gain of frame (i-1)' nth and the weight of a gain of a voice frame detection, VAD, frame according to stability energy of (i-1) ‘th frame; and performing (403) a weighting operation on the algebraic codebook gain of frame (i-1) and the VAD frame gain according to the weight of the algebraic codebook gain on frame (jl) ' th and the weight of the VAD frame gain to obtain the algebraic codebook gain of the ith frame; where more stable energy of (table i-1) "nth indicates a greater weight of the gain of the algebraic codebook of (i-1)" nth; wherein the gain weight of the VAD frame increases correspondingly as the number of consecutive lost frames increases; in which before performing a weighting operation on frame algebraic codebook gain (i-1) 's and the VAD frame gain according to the weight of frame algebraic codebook gain (i-1) ) The value and the weight of the VAD frame gain, for obtaining the gain of the algebraic codebook of the nth frame, the method also comprising: determining a first correction factor according to a coding and decoding rate; and correct the gain of the frame's algebraic codebook (i-1 y-esimo usanc] 0 0 first correction factor. Method according to claim 1, characterized in that obtaining an algebraic codebook from the ith frame comprises: obtaining the algebraic codebook from the ith frame by estimation according to random noise ; or determine the algebraic codebook of table i in accordance with the algebraic codebook of the first tables N of table i. A method according to claim 1, characterized in that the i-frame gain comprises an adaptive codebook gain and an algebraic codebook gain; prior to the generation of an excitation signal from the ith frame according to the pitch period and gain for the ith frame and which are obtained by estimation and the algebraic codebook obtained from the ith frame, the method also comprising: determining the weight of an algebraic codebook contribution from the i-th frame according to any one of a pitch (i-1) 's pitch deviation, correlation of a signal from the (i-1) -this, a tilt-spectrum value of table (i-1) -this or a zero crossing rate of a (i-1) -this or determine the weight of a contribution of the ith frame algebraic codebook by performing a weighting operation on any combination of a deviation of a frame pitch period (i-1) 'nth, a correlation of a frame signal (i-1) " nth, a tilt spectrum value of frame (i-1) 'nth or a crossover rate zero by frame (i-1) "and perform an interpolation operation on an excitation signal with updated frame (i-1) status to determine an adaptive codebook for frame i"; and generating an excitation signal from the ith frame according to the pitch period and gain referring to the ith frame and which are obtained by estimation and the algebraic codebook obtained from the i-es' frame. It comprises: determining the contribution of Table I algebraic codebook according to a product obtained by multiplying Table I algebraic codebook by the gain of Table I algebraic codebook; contribution of the ith framework adaptive codebook according to a product obtained by multiplying the ith framework adaptive codebook by the gain of the ith framework adaptive codebook, and performing a weighting operation on the contribution from the algebraic codebook of Table I and the contribution of the adaptive codebook of Table I according to the weight of the contribution from the algebraic codebook of Table I and a weight of the contribution of the adaptive codebook of the ith frame to determine the excitation signal of the ith frame, in which a weight of the adaptive codebook is 1. A method according to claim 2, characterized in that the method comprises any of the ways of implementation as follows: a first way of implementation: determining, according to the correlation of the nth frame, the need for correct the frequency spectrum parameter of the ith frame and when the need to correct the frequency spectrum parameter of the ith frame is determined, correct the frequency spectrum parameter of the ith frame according to the frame frequency spectrum and a frame frequency spectrum parameter (i-1), or correct the frame frequency spectrum parameter according to frame frequency spectrum parameter nth and a predefined frequency spectrum parameter of table i, in which the correlation of table i comprises a relationship of values between a sixth limit and one of the two parameters frequency spectrum values which correspond to an index of a minimum value of the difference between adjacent frequency spectrum parameters of the nth frame, a ratio of values between a seventh limit and the minimum value of the difference between adjacent spectrum parameters frequency of the ith frame, and a relationship of values between an eighth limit and the index of the minimum value of the difference between adjacent frequency spectrum parameters of the ith frame; Determining, in accordance with the correlation of the nth frame, the need to correct the frequency spectrum parameter of the nth frame comprises: determining the difference between the adjacent frequency spectrum parameters of the nth frame, in which each difference is an index and the frequency spectrum parameter comprises an immittance spectral frequency, ISF, or a line spectral frequency, LSF; determining whether the difference between the adjacent frequency spectrum parameters of the nth frame corresponds to at least one of a fourth condition or a fifth condition, in which the fourth condition comprises: one of the two frequency spectrum parameters corresponding to the index of the minimum value of the difference between the adjacent frequency spectrum parameters of frame i is less than the sixth limit and the fifth condition comprises: an index value of the minimum value of the difference between the adjacent frequency spectrum parameters of table i -th is less than the eighth limit, and the minimum difference is less than the seventh limit; and if the difference between adjacent frequency spectrum parameters of the i th frame corresponds to at least one of the fourth or fifth condition, it is determined that the frequency spectrum parameter of the i th frame should be corrected, or if if the difference between adjacent frequency spectrum parameters of the ith frame does not correspond to the fourth or fifth condition, it is determined that the frequency spectrum parameter of the ith frame is not to be corrected; and the correction of the spectrum parameter frame frequency parameter according to the frame frequency spectrum parameter and a frame (i-1) frequency spectrum parameter comprises: determining a frame frequency corrected parameter -th in accordance with a weighting operation performed on the frame frequency spectrum parameter (i-1) -this and the frame frequency spectrum parameter Γ or the correction of the frequency frame parameter of the ith frame according to the frequency spectrum parameter of the ith frame and a predefined frequency spectrum parameter of the ith frame comprises: determining a corrected parameter of frequency frame of the ith frame according to a weighting operation performed on the frequency spectrum parameter of the ith frame and the default frequency spectrum parameter of the ith frame; A second way of implementation is to determine, according to the correlation between the ith frame and the (il) frame, the need to correct the frequency spectrum parameter of the ith frame, and when the need is determined. frequency setting parameter of the ith frame, correct the frequency spectrum parameter of the ith frame according to the frequency spectrum parameter of the ith frame and a frequency spectrum parameter of the frame ( i-1) "nth, or correct the frequency spectrum parameter of the nth frame according to the frequency spectrum parameter of the nth frame and a predefined frequency spectrum parameter of the nth frame, in which The correlation between the nth frame and the ninth frame (i-1) comprises a relation of values between a ninth limit and the sum of the differences between the frequency spectrum parameters that correspond to some or all of the same indices. s from table (i-1) -th and table i "is th; determining, in accordance with the correlation between frame i and frame (j-1) "es, the need to correct the frequency spectrum parameter of frame j includes: determining the difference between adjacent spectrum parameters the frequency frame of the nth frame, where each difference corresponds to an index and the frequency spectrum parameter comprises an immittance spectral frequency, ISF, or a line spectral frequency, LSF; determine whether the frequency spectrum parameter of the table i and the frequency spectrum parameter of table (i-1) 'nth correspond to a sixth condition, in which the sixth condition comprises: the sum of the differences between the frequency spectrum parameters which correspond to some or all the equal indices of table (i-1) "nth and the nth table is greater than the ninth limit; and if the frequency spectrum parameter of the nth frame and the frequency spectrum parameter of the nth frame (i-1) meet the sixth condition, determine that the frequency spectrum parameter of the nth frame shall be corrected. , or if the i-th frame frequency spectrum parameter and the i-th frame frequency spectrum parameter (i-1) do not meet the sixth condition, it is determined that the i frequency frame parameter shall not be corrected. and the correction of the frequency spectrum parameter of the ith frame according to the frequency spectrum parameter of the ith frame and a frequency spectrum parameter of the ith frame (i-1) comprises: determine a corrected frequency spectrum parameter of the ith frame according to a weighting operation performed on the frequency frame parameter (i-1) 'th and the frequency spectrum parameter of the frame Γ or the correction of the frequency frame parameter of the ith frame according to the frequency spectrum parameter of the ith frame and a predefined frequency spectrum parameter of the ith frame comprises: determining a corrected parameter of frequency frame of the ith frame according to a weighting operation performed on the frequency spectrum parameter of the ith frame and the default frequency spectrum parameter of the ith frame; a third way of implementation: determining, according to the correlation between the ith frame and the ith frame (i-1) and the energy stability between the ith frame and the (i-1) frame, the need to correct the excitation signal of the ith frame, and when the necessity to correct the excitation signal of the tenth frame is determined, to correct the excitation signal of the ith frame according to the energy stability between the frame j-esimoe the frame (il) "esimo, in which to determine, according to the correlation between the frame i-esimo and the frame (i-1) 'esimo and the energy stability between the frame eima and the frame ( (1) "nth, the need to correct the ith frame excitation signal comprises: determining a pre-synthesized ith frame signal according to the ith frame excitation signal and spectrum parameter frequency table i "; determine whether an absolute value of the difference between the signal energy p re-synthesized of the i-th frame and the energy of a synthesized frame signal (i-1 is greater than one tenth limit; and if the absolute value of the difference between the pre-synthesized signal energy of the i-th frame and the synthesized signal energy of the i-th frame "is greater than the tenth limit, determine that the signal of the excitation of the nth frame, or if the absolute value of the difference between the pre-synthesized signal energy of the nth frame and the synthesized signal energy of the frame (i-1) "nth is less than or equal to the tenth limit, determine that the excitation signal of the table i'es, mo should not be corrected; and the correction of the excitation signal of the ith frame according to the energy stability between the ith frame and the (i-1) frame "comprises: determining a second correction factor according to the energy stability between frame i and frame (jl), where the second correction factor is less than 1, and multiply the excitation signal from frame i by the second correction factor to obtain a corrected excitation of Table I; A fourth way of implementation is to determine, according to the correlation between the i-th frame and the i-1 frame and the energy stability between the i-th frame and the i-1 frame. the need to correct the excitation signal of the ith frame, and when the need to correct the excitation signal of the ith frame is determined, to correct the excitation signal of the ith frame according to the energy stability between Table I-es and Table (I-1) "Esimo, in which you determine, according to the correlation between Table I-es and Table (I-1) 'Esimo and the energy stability between Frame I-Es and frame (i-1) "nth, the need to correct the frame excitation signal comprises: determining a pre-synthesized frame signal according to the frame excitation signal and the frequency spectrum parameter of frame i; determine whether a relationship between the pre-synthesized signal energy of the adth and the energy of a synthesized signal from table (i-1) 'nth is greater than an eleventh limit, where the eleventh limit is greater than 1; and if the ratio of the preset synthesized signal energy of the nth frame to the synthesized signal energy of the nth frame is greater than the eleventh threshold, determine that the excitation signal of the table, or if the ratio of the pre-synthesized signal energy of the i-esimo table to the synthesized signal energy of the i-1's frame is less than or equal to the eleventh threshold, determine that the frame excitation signal Γth. _ 5 fc? the correction of the excitation signal of the ith frame according to the energy stability between the ith frame and the (i-1) frame "comprises: determining a second correction factor according to the energy stability between the nth frame and the (j-1) table "nth, in which the second correction factor is less than 1; and multiplying the excitation signal from the ith frame by the second correction factor to obtain a corrected excitation signal from the ith frame; the fifth way of implementation: to determine, according to the correlation between frame i-es and frame o (i-1) 'and the energy stability between frame i and es (i-1)' es, Thus, the need to correct the frame excitation signal is, and when the need to correct the frame excitation signal is determined, to correct the frame excitation signal according to the stability of the frame. energy between table i-th and table (i-1) 'th, in which to determine, according to the correlation between table i-es, grinds table (ji)' and the energy stability between table i nth and frame (i-1) “nth, the need to correct the excitation signal of the nth frame comprises: determining a pre-synthesized signal from the nth frame according to the excitation signal of the nth frame. nth and the frequency spectrum parameter of table i "nth; determining whether a ratio of the energy of a synthesized signal from frame (i-1) "esimo to the energy of a synthesized signal from frame i" 1 is less than a twelfth threshold, where the twelfth threshold is smaller that 1; and if the ratio of the pre-synthesized signal energy of frame (i-1) "esimo to the synthesized signal energy of frame ios is less than the twelfth threshold, determines that the signal should be corrected frame excitation signal, or if the ratio of the pre-synthesized signal energy of frame (i-1) to the synthesized signal energy of frame i is greater than or equal to the twelfth threshold, determine that the excitation signal of frame i should not be corrected; and the correction of the excitation signal of the ith frame according to the energy stability between the ith frame and the (i-1) frame "comprises: determining a second correction factor according to the energy stability between frame i and frame (i-1) "es, where the second correction factor is less than 1; and multiplying the excitation signal from the ith frame by the second correction factor to obtain a corrected excitation signal from the ith frame; A sixth way of implementation is to determine, according to the correlation of a frame (i-1) signal, the need to correct the frame excitation signal and when the need to correct the frame signal is determined. excitation of the ith frame, correct the excitation signal of the ith frame according to the energy stability between the ith frame and the (i-1) frame, in which the correlation of the frame signal ( i-1) 'sth comprises a relation of values between a thirteenth limit and the signal correlation of the frame (jl) ~ mo, and a relation of values between a fourteenth and a deviation of a signal pitch period of frame (i-1) "th. 3 determine, according to the correlation of a signal from frame (i-1)" th, the need to correct the excitation signal of frame i-th comprises: determining whether the signal of table (i-1) 'seventh corresponds to a seventh condition, in which the seventh condition is: table (i-1)' and simo is a lost frame, a signal correlation value of frame (i-1) "eso is greater than the thirteenth limit and deviation of the pitch pitch of the (i-1)" signal is less than fourteenth; and if the signal from frame (i-1) 'nth meets the seventh condition, it is determined that the excitation signal from frame i' shall be corrected, or if the signal from frame (i-1) 'nth does not correspond to the seventh condition, determine that the excitation signal of the nth frame should not be corrected; and correcting the excitation signal of the ith frame according to the energy stability between the ith frame and the (i-1) frame comprises: determining a third correction factor according to the energy stability between the ith frame and the ith frame (il) ", where the third correction factor is less than 1; and multiply the excitation signal from the ith frame by the third correction factor to obtain a corrected excitation of the nth frame: a seventh way of implementation: determining, according to the correlation between the nth frame signal and a nth frame (i-1) signal, the need to correct the excitation signal of the ith frame, and when the need to correct the excitation signal of the ith frame is determined, correct the excitation signal of the ith frame according to the energy stability between the ith frame and the frame. (i-1) ~, in which the correlation between the quad signal ro-th and the frame signal (i-1 seventh comprises a ratio of values between a thirteenth limit and a correlation value of the frame (i-1) signal and a ratio of values between a fourteenth and a deviation of a pitch period from the frame signal i; To determine, in accordance with the correlation between the ith frame signal and a "th frame" signal (i-1), the need to correct the ith frame excitation signal comprises: determining whether the frame signal ( i-1) "nth and the signal of the nth frame correspond to an eighth condition, in which the eighth condition comprises: the (i-1) 'nth frame is a lost frame, a correlation value of the frame signal ( i-1) -th is greater than the preset thirteenth limit and the deviation of the signal pitch period of the i-esimo frame is smaller than the preset fourteenth limit; and if the frame signal (i-1) 's and the frame signal' i 'correspond to the eighth condition, determine that the excitation signal of frame i' is to be corrected, or if the frame signal (i-1) (1) If the ith frame signal does not correspond to the eighth condition, determine that the ith frame excitation signal is not to be corrected. with the energy stability between frame j and frame (i-1) 'comprises: determining a third correction factor according to energy stability between frame i and frame (i-1) This is where the third correction factor is less than 1, and multiplying the excitation signal from the ith frame by the third correction factor to obtain a corrected excitation signal from the ith frame; implementation: determine, according to the correlation between a frame signal (i-1 seventh and a sin In addition to Frame (I-2), the need to correct the excitation signal of Frame I and when determining the need to correct the excitation signal of Frame I, correct the excitation signal of Frame I nth according to the energy stability between the ith frame and the (i-1) frame, in which the correlation between the (i-1) −th frame signal and the frame (i-1) signal 2) The thirtieth comprises: a relation of values between a thirteenth limit and a correlation value of the (i-2) frame signal and whether an excitation signal of the (i-1) frame is corrected; determining, in accordance with the correlation between a 'th frame' (i-1) signal and a th frame '(i-2)' signal, the need to correct the i-th frame excitation signal comprises: determining whether the frame sign (i-2) 'nth and frame sign (i-1)' nth corresponds to a ninth condition, in which the ninth condition comprises: frame (i-2) 'nth is a lost frame, a correlation value of frame (i-2) signal is higher than the thirteenth limit and frame excitation signal (i-1) is corrected; and if the signal from frame (i-2) ~ and the signal from frame (i-1) ~ meets the ninth condition, it is determined that the excitation signal from frame i should be corrected, or if the signal from frame (i-2) ~ and the signal from frame (i-1) “if it does not meet the ninth condition, determine that the excitation signal from frame Γ should not be corrected. The correction of the excitation signal of the ith frame according to the energy stability between the ith frame and the (i-1) frame comprises: determining a fourth correction factor according to the energy stability between the ith frame and the ith frame (il), where the fourth correction factor is less than 1; and multiply the excitation signal of the ith frame by the fourth correction factor to obtain a corrected excitation of the nth frame; a ninth way of implementation: determining, according to the correlation between a nth frame signal (i-1) and a nth frame signal (i — 2), the need to correct the excitation signal of the ith frame, and when the need to correct the excitation signal of the ith frame is determined, correct the excitation signal of the ith frame according to the energy stability between the ith frame. and the (i-1) 'nth frame, in which the correlation between the frame signal (i-1) 'nth and the signal from frame (i-2)' es, m and m, comprises a relationship of values between a thirteenth limit and a correlation value of the signal from (i-2) 'nth, and a relationship of values between a fifteenth bound and an algebraic codebook contribution of an excitation signal from (i-1) ~ frame; determining, in accordance with the correlation between a nth frame (i-1) signal and a nth frame (i-2) signal, the need to correct the nth frame excitation signal comprises: determining whether the frame sign (i-2) 'nth and frame sign (i-1)' nth correspond to a tenth condition, in which the tenth condition comprises: frame (i-2) 'nth is a lost frame, a Frame Signal Correlation Value (i-2) -th is greater than the thirteenth limit, and the algebraic codebook contribution of frame excitation signal (i-1) -this less than the fifteenth limit and if the signal from frame (i-2) 's and the signal from frame (il) -this meet the tenth condition, it is determined that the excitation signal from frame i should be corrected, or if the signal from frame (i-2) "nth and sign of table (i-1)" nth does not meet the tenth condition, determine that the excitation signal of table Γ is not to be corrected. _ 3 fcf the correction of the excitation signal of the ith frame according to the energy stability between the ith frame and the (i-1) frame comprises: determining a fourth correction factor according to the energy stability between frame i and frame (j — 1) -th frame3 where the fourth correction factor is less than 1; and multiplying the excitation signal from the ith frame by the fourth correction factor to obtain a corrected excitation signal from the ith frame; a tenth way of implementation: determining, according to the correlation between a nth frame signal (jl) and the nth frame signal, the need to correct the updated status excitation signal of the nth frame, and when determining the need for correction of the i-frame updated status excitation signal, correct the i-frame's updated status excitation signal according to the power stability between the i-frame and the frame (í -1) "thth, in which the correlation between the thth frame (il) signal and the ththth frame signal comprises: the correlation between the thth (i-1)" frame and the thth frame, and if a frame excitation signal (i-1) is corrected; determine, according to the correlation between a frame (i-1) signal and the frame signal, the need to correct the Excitation signal with updated status of frame j comprises: determining if the frame signal is the signal of the frame (i-1 and j correspond to an eleventh condition, in which the eleventh condition comprises: the frame and the frame (i-1) 'is a highly correlated frame and the frame's excitation signal (i-1) 'nth is corrected; and if the ith frame signal and the ith frame signal (i-1) match the eleventh condition, it is determined that the updated status excitation signal of the ith frame shall be corrected, or if the signal from the ith frame and the signal from frame (i-1) 'e do not correspond to the eleventh condition, determine that the excitation signal with updated status of frame Γ shall not be corrected. α the correction of the excitation signal with updated status of the ith frame according to the energy stability between the ith frame and frame (i-1) 'comprises: determining a fifth correction factor according to the energy stability between the ith frame and the ith frame (i-1), where the fifth correction factor is less than 1; and multiply the updated status excitation signal of the ith frame by the fifth correction factor. correction to obtain a corrected excitation signal with updated status of the nth frame. A method according to claim 8, wherein, when the method comprises any of the third mode of implementation, the fourth mode of implementation or the fifth mode of implementation, determining a second correction factor according to the energy stability between frame i and frame i-1 comprising: determining that a relationship between frame energy (i-1 and frame i energy is the second correction factor, or determine that a relationship between the energy of the same number of subframes in frame (i-1) and the frame is the second correction factor. 10. Frame loss compensation processing apparatus, comprising: a frame loss determination module (11), configured to determine, using a frame loss bit, whether a frame is esimo is a lost picture; an estimation module (12), configured to: when the i-esimofor frame is a lost frame, estimate a parameter of the i-frame according to at least one of an inter-frame relationship between the first N frames of the i-frame or an intraframe relationship between the first N frames of the nth frame, wherein the interframe relationship between the first N frames comprises at least one of the correlation between the first N frames or the energy stability between the first N frames, the intraframe relationship between the first N frames comprises at least one of the inter-subframe correlation between the first N frames or the energy stability of the inter-frames between the first N frames, the ith frame parameter comprises a frequency spectrum parameter, a pitch period, and a gain, and N is an integer greater than or equal to 1; a retrieval module (13), configured to obtain an algebraic codebook from the nth frame; a generation module (14), configured to generate an excitation signal from the ith frame according to the pitch period and gain referring to the ith frame and which are obtained by the estimation module (12) by means of estimation and the algebraic codebook for the ith table and obtained by the obtainment module (13); and a signal synthesis module (15), configured to synthesize a signal from the ith frame according to the frequency spectrum parameter for the ith frame and which is obtained by the estimation module (12) by means of estimation and the excitation signal for the i-th frame and which is generated by the generation module (14). Apparatus according to claim 10, characterized in that if the i-esimofor frame is a normal frame, the apparatus also comprises: a decoding module (16) configured to obtain the i-esim frame parameter by decoding means according to a received bit stream in which the i_th frame parameter comprises the frequency spectrum parameter, pitch period, gain and algebraic codebook in which the generation module (14) ) is also configured to generate the i-th frame excitation signal and an i-th frame updated status excitation signal according to the pitch period, gain and algebraic codebook for the i-th frame and which are obtained by the decoding module (16) by means of decoding; an evaluation module (17), configured for: when a nth frame (i-1) or a nth frame (i-2) is a lost frame, determine according to at least one of the inter-frame relationships or intra-frame relationships between the ith frame and the first N frames of the ith frame, the need to correct at least one of the frequency spectrum parameter, the excitation signal or the updated frame excitation signal Second, in which the interframe relationship comprises at least one of the correlation between the ith frame and the first N frames of the ith frame or the energy stability between the ith frame and the first N frames of the frame. The intra-frame ratio comprises at least one of the inter-subframe correlation between the nth frame and the first N frames of the nth frame or the energy stability of the inter-frames between the nth frame and the first frames N of frame i and a correction module (18), configured for: when the evaluation module (17) determines that it must correct at least one of the frequency spectrum parameter, the excitation signal or the updated status excitation signal of frame i ", correct at least one of the frequency spectrum parameter, excitation signal or excitation signal with updated status of frame i according to at least one of the interframe relations or the intraframe relations between the ith frame and the first N frames of the ith frame, in which the signal synthesis module (15) is also configured to: synthesize the signal from the ith frame according to a correction result performed by the correction module (18) in at least one of the frequency spectrum parameter, the excitation signal or the updated status excitation signal of the nth frame; or when the evaluation module (17) determines that it should not correct the frequency spectrum parameter, the excitation signal or the updated status excitation signal of the i th frame, synthesizes the i th frame signal according to the frequency spectrum parameter, the excitation signal and the updated status excitation signal of Table I. Apparatus according to claim 10, characterized in that the frequency spectrum parameter of the ith frame is obtained by the estimation module (12) by estimation according to the interframe relationship between the first frames. No of the parent table; and the estimation module (12) is specifically configured to: determine the weight of a frequency frame parameter of an i-1 frame and the weight of a predefined frequency spectrum parameter of an i-frame according to the correlation between the first frames N of frame i; and perform a weighting operation on the frame frequency spectrum parameter (i-1) 'and the default frame frequency spectrum parameter according to the weight of the frame frequency spectrum parameter (i-1)'. 1) 's is the weight of the default frequency spectrum parameter of the ith frame to obtain the frequency spectrum parameter of the ith frame; wherein the correlation comprises a relationship of values between a second boundary and a spectrum slope parameter of a frame signal (i-1 yesim ° a relationship of values between a first boundary and a normalized signal frame autocorrelation value ( jl) "es, mo, and a relation of values between a third threshold and a deviation of a signal period from the frame signal (jl)" and the estimation module (12) is specifically configured for: if the signal of frame (i-1) 'th is at least one of a first condition, a second condition, and a third condition, determining that the weight of the frequency spectrum parameter of frame (i-1) -th is a first weight and that the weight of the default frequency spectrum parameter of the ith frame is a second weight, where the first weight is greater than the second weight, the first condition is: the normalized autocorrelation value of the frame signal (i- 1) ~ is higher than the first boundary, the second condition is: the frame signal spectrum slope parameter (i-1) "is higher than the second boundary, and the third condition is: the frame signal pitch period deviation ( (1) th is less than the third limit; or if the signal from frame (jl) does not correspond to a first condition, a second condition, or a third condition, determines that the weight of the frequency spectrum parameter of frame (i-1) 'nth is a second weight. and the weight of the preset frequency spectrum parameter of the ith table is a first weight, in which the first weight is greater than the second weight, Apparatus according to claim 10, characterized in that the pitch period of the ith frame is obtained by the estimation module (12) by estimation according to the correlation between the first N frames of frame i -th and the inter-subframe correlation between the first N frames of the nth frame; The correlation comprises a relation of values between a fifth limit and a normalized autocorrelation value of a nth frame (i-2) signal, a relation of values between a fourth limit and a deviation of a pitch period from the frame (i-2) 's, and a relationship of values between the fourth boundary and a pitch shift of a signal from a frame (i-1)' th; and the estimation module (12) is specifically configured to: if the pitch signal period shift deviation from frame (i-1) is less than the fourth threshold, determine a signal pitch pitch shift deviation value from (jl) "nth according to the pitch period of the nth frame signal (i-1), and determine a pitch period of the nth frame signal according to the pitch period deviation value of the nth frame frame (i-1) signal "is the pitch period of the frame (i-1) signal", in which the pitch period of the frame signal i-1 comprises a pitch period of each frame subframe resume and the pitch shift value of the frame (i-1) signal is a mean value of the differences between the pitch periods of all adjacent subframes of the frame (i-1); or if the pitch (i-1) signal pitch deviation of the nth frame is greater than or equal to the fourth threshold, the normalized autocorrelation value of the ninth frame signal is greater than the fifth threshold and If the pitch signal deviation of the frame (i-2) 's pitch is less than the fourth threshold, determine a pitch signal deviation value of the frame (i-2)' s pitch and the frame signal ( i-1) "nth in accordance with the frame signal pitch period (i-2) - nth and the frame signal pitch period (i-1)" nth, and determine a pitch signal period of the frame (i-2) ith frame according to the frame signal pitch period (i-1 and the frame signal pitch period offset value (i-2) "ith and the frame signal (i-1) 'where the estimation module (12) determines the pitch value shift deviation of the signal from frame (i-1) - nth according to the following formula: / n = p-1) (3) -p <-1) (2)) + p-1) (2) _ /, H) 0)) + (/, H) 0) _ /, H) (o)) / 3inaqua | P V) 0 is a pitch period of the twelfth frame of (i-1) -Eth frame, ej = 0, 1, 2, 3; and the estimation module (12) determines the pitch period of the i-esimode frame signal according to the following formula: R „Ü) = ^ H '(3) +0 + l) Vy = 0, 1, 2, 3i on wed | / "(3) is a pitch period of a third subframe of the (i-1) -th frame, ^ is the value of the pitch signal period shift deviation of the frame (jl)", and Parii) q is a pitch period of a subframe already "nth of table i" nth; or wherein the estimation module (12) determines the pitch period shift value of frame (i-2) -th signal and frame (i-1) -th signal according to the following formula: ρν = (ρ (-2) (3) - /, (- 2) (2)) + (ρ (- || (0) -? (- 2) (3)) + (ρ (- || ( 1) -ρ (-,) (0)) / 3ι na qua | P ^ is a pitch period of a nnth frame subframe (i-2) 'nth, P ^ is a pitch period of a n "nth frame subframe (M) -th, m = 2, 3 and n = 0, 1; and the estimation module (12) determines the pitch period of the i-esimode frame signal according to the following formula: ραιι, (χ) = ρΗ} (3) + (χ + ψρν, x = 0, 1, 2,3j ng qug | P ® is a pitch period of a third subframe of the frame (j-1) "nth, Pv is the pitch period deviation value of frame (i-2) is the nth frame signal (M) is θ, and P ^ x ^ is a pitch period of a nth frame subframe . Apparatus according to claim 10, characterized in that the ith frame gain comprises an adaptive codebook gain and an algebraic codebook gain and the ith frame gain is obtained by the module. estimation (12) by estimation according to the correlation between the first N-tables of Table I and the energy stability between the first N-tables of Table I and the estimation module (12) is specifically is set to: determine the i-frame adaptive codebook gain according to a n-frame adaptive codebook gain (i-1) "or a predefined fixed value, frame correlation (i-1) nth and one frame number of the nth frame in multiple consecutive lost frames; determining the weight of a gain from frame algebraic codebook (i-1) and the weight of a gain from a voice frame detection, VAD, frame according to the energy stability of frame (i-1) ) "s, mo; and perform a weighting operation on the algebraic codebook gain in frame (i-1) 'and the VAD frame gain according to the weight of the algebraic codebook gain in frame ( i-1) 'nth and the weight of the VAD frame gain to obtain the algebraic codebook gain of the nth frame; where more stable energy of the (i-1) frame' nth indicates a higher weight of the algebraic codebook gain from table (i-1) 'th. Wherein the weight of the VAD frame gain correspondingly increases as the number of consecutive lost frames increases; where before performing a weighting operation on the algebraic codebook gain of frame (i-1) -th and the VAD frame gain according to the weight of the algebraic codebook gain (i-1) ) 's and the VAD frame gain weight, to obtain the algebraic codebook gain of the nth frame, the estimation module (12) is also configured to: determine a first correction factor according to a encoding and decoding rate; and correct the gain of the algebraic codebook of table (i-1) -ths usancj0 0 the first correction factor. Apparatus according to claim 10, characterized in that the obtainment module (13) is specifically configured to: obtain the algebraic codebook of the nth frame by estimation according to random noise; or determine the algebraic codebook of table i according to the algebraic codebook of the first tables N of table i. Apparatus according to claim 10, characterized in that the i-frame gain comprises an adaptive codebook gain and an algebraic codebook gain; the obtaining module (13) is also configured to: determine the weight of an algebraic codebook contribution of the ith frame according to any one of a pitch period deviation of a frame (i-1) ' nth, correlation of a frame signal (M) "nth, a tilt rate value of the frame (i-1) - nth or a zero crossing rate of a frame (i-1)" nth or determine the weight of a contribution from the ith frame algebraic codebook by performing a weighting operation on any combination of a deviation from a frame pitch period (−1) "nth, a correlation of a frame signal ( i-1) “nth, a tilt spectrum value of the (i-1) 'nth frame or a zero crossing rate of the table (i-1)' nth; and performing an interpellation operation on an excitation signal with updated frame (j-1) status to determine an adaptive codebook of frame 1; and the generation module (14) is specifically configured to: determine the contribution of the algebraic codebook of the jth frame according to a product obtained by multiplying the algebraic codebook of the jth frame by the gain of the textbook algebraic codes in table i; determining a contribution from the i-frame adaptive codebook according to a product obtained by multiplying the i-frame adaptive codebook by the gain of the i-frame adaptive codebook; and carry out a weighting on the contribution of the i-table algebraic codebook and the contribution of the table-i adaptive codebook according to the weight of the contribution of the i-table algebraic codebook and a the contribution of the adaptive codebook of the ith frame to determine the excitation signal of the ith frame, where a weight of the adaptive codebook is 1, Apparatus according to claim 16, characterized in that the apparatus comprises any of the combination ways as follows: a first combination mode: the evaluation module (17) is configured to determine, according to the correlation of the ith frame, the need to correct the frequency spectrum parameter of the ith frame, and when the evaluation module (17) determines that it must correct the frequency spectrum parameter of the ith frame , the correction module (18) is configured to: correct the i-esim frequency frame spectrum parameter according to the i-es frame frequency spectrum parameter and a frame frequency spectrum parameter (jl) " or to correct the frequency spectrum parameter of the ith frame according to the frequency spectrum parameter of the ith frame and a predefined frequency spectrum parameter of the ith frame. which the correlation of the nth frame comprises a relation of values between a sixth limit and one of the two frequency spectrum parameters corresponding to an index of a minimum value of the difference between the adjacent frequency spectrum parameters of the nth frame , a ratio of values between a seventh limit and the minimum value of the difference between the adjacent frequency spectrum parameters of Table I and a ratio of values between an eighth limit and the index of the minimum value of the difference between the adjacent parameters of frequency spectrum of the parent frame; The evaluation module (17) is specifically configured to: determine the difference between adjacent frequency spectrum parameters of the nth frame, where each difference corresponds to an index and the frequency spectrum parameter comprises an immittance spectral frequency. ISF or a line spectral frequency, LSF; determining whether the difference between the adjacent frequency spectrum parameters of the nth frame corresponds to at least one of a fourth condition or a fifth condition, in which the fourth condition comprises: one of the two frequency spectrum parameters corresponding to the index of the minimum value of the difference between the adjacent frequency spectrum parameters of the i th frame is less than the sixth limit, and the fifth condition comprises: an index value of the minimum value of the difference between the adjacent frequency spectrum parameters of the i th frame i-th is less than the eighth limit and the minimum difference is less than the seventh limit; and if the difference between adjacent frequency spectrum parameters of the ith frame corresponds to at least one of the fourth or fifth condition, determine the correction of the frequency spectrum parameter of the ith frame, or if the difference between adjacent frequency spectrum parameters of frame j does not match the fourth or fifth condition, determine non-correction of the frequency spectrum parameter of frame i; and the correction module (18) is specifically configured to: determine a corrected frequency spectrum parameter of the ith frame according to a weighting operation performed on the frequency spectrum parameter of the frame (i-1) ith and the frequency spectrum parameter of the ith frame; or determine a frequency spectrum corrected parameter of the ith frame according to a weighting operation performed on the specimen parameter frequency table of frame i and the default frequency spectrum parameter of table i; a second way of combining: wherein the evaluation module (17) is configured to determine, according to the correlation between frame i and frame (i-1), the need to correct the spectrum parameter frequency module, and when the evaluation module (17) determines that it must correct the frequency spectrum parameter of the nth frame, the correction module (18) is configured to: correct the frequency spectrum parameter. frame frequency according to the frame frequency spectrum parameter and a frame (i-1) frequency spectrum parameter, or correct the frame frequency frequency parameter according to the frequency spectrum parameter of the nth frame and a predefined frequency spectrum parameter of the nth frame, in which the correlation between the nth frame and the (i-1) th frame comprises a relation of values between a ninth limit eas o oma of the differences between frequency spectrum parameters corresponding to some or all of the same indices of table (i-1) and the table; The evaluation module (17) is specifically configured to: determine the difference between adjacent frequency spectrum parameters of the resume frame, where each difference corresponds to an index and the frequency spectrum parameter comprises an immittance spectral frequency, ISF , or an LSF line spectral frequency; determining whether the frequency frame parameter of the ith frame and the frequency spectrum parameter of the frame (i-1) 'nth correspond to a sixth condition, in which the sixth condition comprises: the sum of the differences between the parameters of frequency spectrum corresponding to some or all of the same indices of table (i-1) "nth and frame nth is greater than the ninth limit; and if the frequency spectrum parameter of table nth is the parameter Frequency Spectrum (i-1) - If the sixth condition corresponds to the sixth condition, determine the correctness of the Frequency Spectrum parameter of the nth frame, or if the Frequency Spectrum parameter of the nth frame and the frequency spectrum of frame (i-1) does not match the sixth condition, determine non-correction of the frequency spectrum parameter of frame i, and correction module (18) is specifically configured to: determine a corrected frequency spectrum parameter of the ith frame according to a weighting operation performed on the frequency spectrum parameter of the frame (i-1) ‘and the frequency spectrum parameter of the ith frame; or determining a corrected frequency spectrum parameter of the ith frame according to a weighting operation performed on the frequency spectrum parameter of the ith frame and the default frequency spectrum parameter of the ith frame; a third way of combining: wherein the evaluation module (17) is configured to determine, according to the correlation between frame i and frame i and the power stability between frame i -th and the frame (i-1 seventh, the need to correct the excitation signal of the i'es frame, mo, and when the evaluation module (17) determines that it must correct the excitation signal of the ith frame, correction module (18) is configured to correct the excitation signal of the ith frame according to the energy stability between the ith frame and the (i-1) "th frame; the evaluation module (17) ) is specifically configured to: determine a preset synthesized signal from the ith frame according to the excitation signal from the ith frame and the frequency spectrum parameter of the ith frame; determine whether an absolute value of the difference between the power of the i-th frame pre-synthesized signal and the power of a synthesized signal Table (i-1) 'is greater than one tenth limit; and if the absolute value of the difference between the preset synthesized signal energy of the nth frame and the synthesized signal energy of the nth frame is greater than the tenth limit, determine the correction of the excitation signal of the table i, or if the absolute value of the difference between the power of the synthesized signal of frame i and the power of the synthesized signal of table (i-1) is less than or equal to the tenth limit, determine the not correcting the excitation signal of the nth frame; and the correction module (18) is specifically configured to: determine a second correction factor according to the energy stability between frame I and frame I, where the second correction factor is less than 1; and multiplying the excitation signal from the ith frame by the second correction factor to obtain a corrected excitation signal from the ith frame; a fourth way of combining: wherein the evaluation module (17) is configured to determine, according to the correlation between frame i and frame i and the power stability between frame i -th and frame (i-1) "is the need to correct the excitation signal from frame i, and when the evaluation module (17) determines that it must correct the excitation signal from frame i," correction module (18) is configured to correct the excitation signal of the ith frame according to the energy stability between the ith frame and the ith frame (i-1); the evaluation module (17) is specifically configured to: determine a preset synthesized frame signal according to the frame excitation signal and frame frequency parameter frequency; a relationship between the power of the preset synthesized signal of frame I and the power of a synthesized signal of frame I-1 is greater than an eleventh limit, where the eleventh limit is greater than 1; and if the ratio of the pre-synthesized signal energy of frame j to the synthesized signal energy of frame (i-1) is greater than the eleventh threshold, determine the correction of the excitation signal of frame i "th, or if the ratio of the pre-synthesized signal energy of the ith frame to the synthesized signal energy of the frame (jl) ~ m, when less than or equal to the eleventh threshold, determines the non-correction of the signal excitation of the nth frame; and the correction module (18) is specifically configured to: determine a second correction factor according to the energy stability between frame I and frame (M) where the second correction factor is smaller that 1; and multiplying the excitation signal from the ith frame by the second correction factor to obtain a corrected excitation signal from the ith frame; The fifth way of combining: wherein the evaluation module (17) is configured to determine, according to the correlation between the i-th frame and the i-1 th frame and the energy stability between the i th frame and frame (i-1) "is the need to correct the excitation signal of the nth frame, and when the evaluation module (17) determines that it must correct the excitation signal of the nth frame, the module The correction signal (18) is configured to correct the excitation signal of the ith frame according to the energy stability between the ith frame and the ith frame (i-1); the evaluation module (17) is specifically configured to: determine a pre-synthesized frame signal according to the frame excitation signal and frame frequency spectrum parameter, to verify whether a relationship between the energy of a preset signal -synthesized frame (jl) "es and the energy of a synthesized quadratic signal j-esimo is less than a twelfth limit, where the twelfth limit is less than 1; and if the ratio of the pre-synthesized signal energy of frame (i-1) -th to the synthesized signal energy of frame i-1 is less than the twelfth threshold, determine the correction of the excitation signal of frame i "nth, or if the ratio of the pre-synthesized signal energy of the jth frame to the synthesized signal energy of the nth frame is greater than or equal to the twelfth threshold, determine the non-correction of the excitation signal and the correction module (18) is specifically configured to determine a second correction factor according to the energy stability between the nth frame and the nth frame (i-1) in the which the second correction factor is less than 1, and multiplying the excitation signal from the nth frame by the second correction factor to obtain a corrected excitation signal from the nth frame, a sixth way of combining: the evaluation module (17) is configured to determine, according to the correlation of a frame (i-1) signal, the need to correct the frame excitation signal, and when the evaluation module (17) determines that it should correct the signal of the nth frame excitation, the correction module (18) is configured to correct the excitation signal of the nth frame according to the energy stability between the nth frame and the nth frame (jl), wherein the nth frame (i-1) signal correlation comprises a ratio of values between a thirteenth threshold and a nth frame (i-1) signal correlation value, and a ratio of values between one tenth fourth and a deviation of a signal period from frame signal (i-1) "ith; the evaluation module (17) is specifically configured to: determine whether the signal from frame (i-1) "seventh corresponds to a seventh condition, in which the seventh condition is: frame (jl)" is a lost frame, a frame signal correlation value (i-1) "nth is greater than the thirteenth limit and the deviation from the pitch signal period of frame (i-1)" nth is less than the fourteenth; and if the frame (i-1) signal 'seventh meets the seventh condition, determine the correctness of the excitation signal of frame i'th, or if frame (i-1)' seventh does not meet the seventh condition, determine correction of the excitation signal of the ith frame, and the correction module (18) is specifically configured to: determine a third correction factor according to the energy stability between the ith frame and the frame (i-1) ) "eso, where the third correction factor is less than 1; and multiplying the excitation signal from the ith frame by the third correction factor to obtain a corrected excitation signal from the Psimo frame; a seventh way of combining: wherein the evaluation module (17) is configured to determine, according to the correlation between the i-frame signal and a n-frame frame (i-1) signal, the need to correct the frame excitation signal, and when the evaluation module (17) determines that it should correct the frame excitation signal, the correction module (18) is configured to correct the frame excitation signal I-th according to the energy stability between the i-th frame and the i-th frame (jl), in which the correlation between the i-th frame signal and the (i-1) frame signal “comprises a relationship of values between a thirteenth threshold and a signal correlation value of the nth frame (i-1), and a relationship of values between a thirteenth and a deviation of a pitch period of the nth frame signal; the evaluation module (17) is specifically configured to: determine if the signal of q (i-1) -th frame and the i-th frame signal correspond to an eighth condition, in which the eighth condition comprises: the (i-1) 'th frame is a lost frame, a correlation value of the frame (i-1 seventh is greater than the default thirteenth limit, and the pitch pitch deviation of the i-esimo frame signal is less than the default fourteenth limit; and if the frame signal (i-1 seventh and the frame signal i-1 correspond to the eighth condition, determine the correction of the frame excitation signal i "or the frame signal (jl)" is the signal of the ith frame does not match the eighth condition, determine non-correction of the excitation signal of the ith frame, and the correction module (18) is specifically configured to: determine a third correction factor according to stability of energy between frame i and frame (i-1) "es, where the third correction factor is less than 1; and multiply the excitation signal from frame i by the third correction factor to obtaining a corrected excitation signal from frame i "is eighth; an eighth combination way: wherein the evaluation module (17) is configured to determine, according to the correlation between a signal from frame (i-1)" esimo and a sign from table (i-2) "esimo, the need to correct the frame excitation signal i, and when the evaluation module (17) determines that it should correct the frame excitation signal i, the correction module (18) is configured to correct the frame excitation signal i-th in accordance with the energy stability between frame i-es and frame (i-1) "eso, in which the correlation between the signal of frame (i-1)" eso and the signal of frame (i-1) 2) The thirtieth comprises: a relationship of values between a thirteenth limit and a correlation value of the "i-2" frame signal and whether an excitation signal of the (i-1) frame is corrected; the evaluation module (17) is specifically configured to: determine whether the signal from (i-2) 'nth frame and the signal from frame (i-1)' ninth condition correspond to a ninth condition, wherein the ninth condition comprises: frame (i-2) is a lost frame, a signal correlation value of frame (i-2) is higher than the thirteenth limit and frame excitation signal "il" is corrected ; and if the (i-2) -th frame signal and the (i-1) -th frame signal correspond to the ninth condition, determine the correction of the nth frame excitation signal, or if the frame signal ( i-2) 'nth and the frame signal (i-1)' nth does not match the ninth condition, determine the non-correction of the nth frame excitation signal, and the correction module (18) is specifically configured to : determine a fourth correction factor according to the energy stability between frame I and frame (i-1), where the fourth correction factor is less than 1, and multiply the excitation signal of the ith frame by the fourth correction factor to obtain a corrected excitation signal from frame ith a ninth combination: wherein the evaluation module (17) is configured to determine, according to the correlation between a sign from frame (i-1) "esimo and a sign from frame (i-2)" esimo, the need to correct rotate the excitation signal from frame i, and when the evaluation module (17) determines that it should correct the excitation signal from frame i, the correction module (18) is configured to correct the excitation signal from i-th frame according to the energy stability between the i-th frame and the i-th frame (i-1), in which the correlation between the i-1's frame signal and the i-th frame's correlation (i-1) -2) "nth comprises a relation of values between a thirteenth limit and a signal correlation value of table (i-2)" and a value relation between a fifteenth limit and a contribution from the algebraic codebook of an excitation signal from frame (i-1) "ith; the evaluation module (17) is specifically configured to: determine whether the signal of the (i-2) -th frame and the signal of frame (i-1) 'the tenth condition, in which the tenth condition comprises: Table o (i-2) 'nth is a lost frame, a signal correlation value of frame (i-2)' nth is greater than the thirteenth limit, and the algebraic codebook contribution of the table (i-1) "nth is less than the fifteenth limit; and if the sign of table (i-2) 'nth and the sign of table (i-1)" nth meet the tenth condition, determine the correction of the tenth frame excitation signal, or if the tenth frame (i-2) signal and the tenth frame (i-1) signal do not match the tenth condition, determine that the frame excitation signal is not corrected i'th; and the correction module (18) is specifically configured to: determine a fourth correction factor according to the energy stability between frame i and the frame (i-1), in which the fourth correction factor is less than 1, and multiply the excitation signal from the ith frame by the fourth correction factor to obtain a corrected excitation signal from the ith frame: a tenth combination: in which the evaluation module (17) ) is configured to determine, according to the correlation between a signal from frame (i-1) and the signal from frame i, the need to correct the excitation signal with updated status from frame i, and when the evaluation module (17) determines that it must correct the updated status excitation signal of the i-th frame, the correction module (18) is configured to correct the updated status excitation signal of the i-th frame according to with energy stability ia between frame i-th and frame (i-1) "th, in which the correlation between the signal of frame (i-1) _th and the signal of frame i-th comprises: the correlation between frame (jl ) "Es is the nth frame, and if a frame excitation signal from (jl)" is corrected; the evaluation module (17) is specifically configured to: determine whether the i-th frame signal and the i-1 th frame signal correspond to an eleventh condition, wherein the eleventh condition comprises: table i -th or frame (i-1) 'is a highly correlated frame and the excitation signal of frame (i-1)' is corrected; and if the ith frame signal and the ith frame signal (i-1) corresponder meet the eleventh condition, determine the correction of the updated status excitation signal of the ith frame, or if the i-frame signal -th and the signal from frame (jl) "nth does not match the eleventh condition, determine the non-correction of the excitation signal with updated status of frame i"; and the correction module (18) is specifically configured to: determine a fifth correction factor according to the energy stability between frame i and the frame (i-1), in which the fifth correction factor is less than 1, and multiply the updated status excitation signal from the i-th frame by the fifth correction factor to obtain a corrected excited signal from the i-th frame. Apparatus according to claim 17, characterized in that when the apparatus comprises any one of the third combination mode, the fourth combination mode or the fifth combination mode, the correction module (18) is specifically set to: determine that a ratio between the power of the (i-1 frame) and the energy of the i th frame is the second correction factor; or determine that a relationship between the energy of the same number of subframes (i -1) "nth and the nth frame is the second correction factor.