CN101185127B - Methods and apparatus for coding and decoding highband part of voice signal - Google Patents

Abstract

A wideband speech encoder according to one embodiment includes a lowband encoder and a highband encoder. The lowband encoder is configured to encode a lowband portion of a wideband speech signal as a set of filter parameters and an encoded excitation signal. The highband encoder is configured to calculate values for coding parameters that specify a spectral envelope and a temporal envelope of a highband portion of the wideband speech signal. The temporal envelope is based on a highband excitation signal that is derived from the encoded excitation signal. In one such example, the temporal envelope is based on a difference in levels between the highband portion and a synthesized highband signal, wherein the synthesized highband signal is generated according to the highband excitation signal and a set of highband filter parameters.

Description

The method and apparatus that is used for the highband part of Code And Decode voice signal

The application's case is advocated based on the 60/667th of being entitled as of application on April 1st, 2005 " CODING THE HIGH-FREQUENCYBAND OF WIDEBAND SPEECH ", the right of priority of No. 901 U.S. Provisional Patent Application cases.The application's case is also advocated based on the 60/673rd of being entitled as of on April 22nd, 2005 application " PARAMETER CODING IN AHIGH-BAND SPEECH CODER ", the right of priority of No. 965 U.S. Provisional Patent Application cases.

Technical field

The present invention relates to signal processes.

Background technology

The bandwidth Conventional cap of the voice communication on PSTN (PSTN) is in the frequency range of 300-3400 kHz.For example, for voice communication (cellular phone and ip voice (Internet Protocol, VoIP)) new network may not have identical limit bandwidth, and may on such network, transmit and receive the voice communication that comprises broadband frequency range.For instance, may need to support to extend downwardly into 50Hz and/or until 7 or the audio frequency range of 8kHz.Also may need to support to have other application, for example high quality audio or the audio/video conference of the audio speech content in the scope beyond traditional PSTN restriction.

The scope that speech coder is supported can be improved sharpness to the extension of upper frequency.For instance, distinguishing for example " s " will be in high-frequency with the fricative information spinner of " f ".High frequency band extends other quality that also can improve voice, for example, have rate.For instance, even turbid vowel also can have the spectrum energy limiting higher than PSTN far away.

A kind of method of wideband speech coding relates to convergent-divergent narrow-band speech coding technology (for example, be configured to encode 0-4kHz the technology of scope) with covering wide band spectrum.For instance, voice signal can higher rate sample to comprise in high-frequency component, and narrow-band coding techniques is reconfigurable to represent this broadband signal with more filter factors.Yet for example the narrow-band coding techniques calculated amount of CELP (code book Excited Linear Prediction) is larger, and broadband celp coder may consume too much cycle for the treatment of, to such an extent as to unrealistic for many movements and other Embedded Application.Using this technology is that required quality also may cause bandwidth greatly to increase making us accepting by the whole spectrum coding of broadband signal.In addition, even, in the narrow-band part of this encoded signal can be transferred to the system of only supporting narrow-band coding and/or before by described system decodes, need to carry out code conversion to described encoded signal.

The another kind of method of wideband speech coding relates to from encoded narrow-band spectrum envelope extrapolation high frequency band spectrum envelope.Although this method can be implemented in the situation that not increasing bandwidth and not needing code conversion, rough spectrum envelope or the resonance peak structure of the highband part of the voice signal that conventionally cannot calculate to a nicety out from the spectrum envelope of narrow-band part.

May need to implement wideband speech coding, make the narrow-band part of at least encoded signal can pass through narrow-band channel (for example, PSTN channel) transmission, and not carry out code conversion or other is significantly revised.Also may need broadband code extension efficient, for example, so that () avoids the number that may accept the user of service in the application of the broadcast on for example wireless cellular telephony and wired and wireless channel significantly to reduce.

Summary of the invention

In one embodiment, a kind of method of highband part that coding has a voice signal of low-frequency band part and highband part comprises: a plurality of filter parameters of the spectrum envelope of computational representation highband part; By extending the frequency spectrum of the signal of partly deriving from low-frequency band, calculate frequency spectrum extension signal; High band excitation signal according to (A) based on frequency spectrum extension signal and (B) a plurality of filter parameters produce synthetic high-frequency band signals; And the relation based between highband part and the signal based on low-frequency band part is carried out calculated gains envelope.

In one embodiment, a kind of method of speech processing comprises: based on low band excitation signal, produce high band excitation signal; Based on high frequency band voice signal and high band excitation signal, produce synthetic high-frequency band signals; And the relation based between high frequency band voice signal and the signal based on low band excitation signal is calculated a plurality of gain factors.

In another embodiment, a kind of method of highband part that decoding has a voice signal of low-frequency band part and highband part comprises: a plurality of gain factors that receive a plurality of filter parameters of the spectrum envelope that characterizes highband part and the temporal envelope of sign highband part; By the signal spectrum extending based on low band excitation signal, calculate frequency spectrum extension signal; According to (A) a plurality of filter parameters and the high band excitation signal of (B) extending signal based on frequency spectrum, produce synthetic high-frequency band signals; And the gain envelope of modulating synthetic high-frequency band signals according to a plurality of gain factors.

In another embodiment, be a kind ofly configured to the equipment of highband part that coding has the voice signal of low-frequency band part and highband part and comprise: analysis module, it is configured to one group of filter parameter of the spectrum envelope of computational representation highband part; Frequency spectrum stretcher, it is configured to calculate frequency spectrum extension signal by extending the frequency spectrum of the signal of partly deriving from low-frequency band; Composite filter, its be configured to according to (A) based on frequency spectrum extend the high band excitation signal of signal and (B) described group of filter parameter produce synthetic high-frequency band signals; And gain factor counter, its time-varying relationship being configured to based between highband part and the signal based on low-frequency band part carrys out calculated gains envelope.

In another embodiment, a kind of high frequency band Voice decoder is configured to receive a plurality of filter parameters and (B) the encoded low band excitation signal of the part of the low-frequency band based on voice signal of spectrum envelope that (A) characterizes the highband part of voice signal.Described demoder comprises: frequency spectrum stretcher, and its signal spectrum being configured to by extending based on encoded low band excitation signal calculates frequency spectrum extension signal; Composite filter, it is configured to characterize a plurality of filter parameters of spectrum envelope of highband part and the high band excitation signal of (B) extending signal based on frequency spectrum produces synthetic high-frequency band signals according to (A); And gain control element, it is configured to modulate according to a plurality of gain factors that characterize the temporal envelope of highband part the gain envelope of synthetic high-frequency band signals.。

Accompanying drawing explanation

Fig. 1 a shows according to the block scheme of the wideband speech coding device A100 of an embodiment.

Fig. 1 b shows the block scheme of the embodiment A102 of wideband speech coding device A100.

Fig. 2 a shows according to the block scheme of the broadband Voice decoder B100 of an embodiment.

Fig. 2 b shows the block scheme of the embodiment B102 of broadband Voice decoder B100.

Fig. 3 a shows the block scheme of the embodiment A112 of bank of filters A110.

Fig. 3 b shows the block scheme of the embodiment B122 of bank of filters B120.

The bandwidth of the low and high frequency band of an example of Fig. 4 a displaying bank of filters A110 covers.

The bandwidth of the low and high frequency band of another example of Fig. 4 b displaying bank of filters A110 covers.

Fig. 4 c shows the block scheme of the embodiment A114 of bank of filters A112.

Fig. 4 d shows the block scheme of the embodiment B124 of bank of filters B122.

Fig. 5 a shows the example of the Logarithmic magnitude of voice signal and the curve of frequency.

Fig. 5 b shows the block scheme of basic linear predictive coding system.

Fig. 6 shows the block scheme of the embodiment A122 of narrowband encoder A120.

Fig. 7 shows the block scheme of the embodiment B112 of narrow-band demoder B110.

Fig. 8 a shows the Logarithmic magnitude of residual signal of turbid voice and the example of the curve of frequency.

Fig. 8 b shows the Logarithmic magnitude of residual signal of turbid voice and the example of the curve of time.

Fig. 9 shows the block scheme of the basic linear predictive coding system of also carrying out long-term forecasting.

Figure 10 shows the block scheme of the embodiment A202 of high band encoder A200.

Figure 11 shows the block scheme of the embodiment A302 of high band excitation generator A300.

Figure 12 shows the block scheme of the embodiment A402 of frequency spectrum stretcher A400.

Figure 12 a shows that frequency spectrum extends the curve of the signal spectrum at each point place in the example operating.

Figure 12 b shows that frequency spectrum extends the curve of the signal spectrum at each point place in another example operating.

Figure 13 shows the block scheme of the embodiment A304 of high band excitation generator A302.

Figure 14 shows the block scheme of the embodiment A306 of high band excitation generator A302.

Figure 15 shows the process flow diagram of envelope calculation task T100.

Figure 16 shows the block scheme of the embodiment 492 of combiner 490.

Figure 17 illustrates the method for the cyclic indicator that calculates high-frequency band signals S30.

Figure 18 shows the block scheme of the embodiment A312 of high band excitation generator A302.

Figure 19 shows the block scheme of the embodiment A314 of high band excitation generator A302.

Figure 20 shows the block scheme of the embodiment A316 of high band excitation generator A302.

Figure 21 shows the process flow diagram of gain calculation task T200.

Figure 22 shows the process flow diagram of the embodiment T210 of gain calculation task T200.

The figure of Figure 23 a impression window function.

Figure 23 b shows the subframe that the window function as shown in Figure 23 a is applied to voice signal.

Figure 24 shows the block scheme of the embodiment B202 of high band decoder B200.

Figure 25 shows the block scheme of the embodiment AD10 of wideband speech coding device A100.

Figure 26 a shows the schematic diagram of the embodiment D122 of lag line D120.

Figure 26 b shows the schematic diagram of the embodiment D124 of lag line D120.

Figure 27 shows the schematic diagram of the embodiment D130 of lag line D120.

Figure 28 shows the block scheme of the embodiment AD12 of wideband speech coding device AD10.

Figure 29 shows according to the process flow diagram of the signal processing method MD100 of an embodiment.

Figure 30 shows according to the process flow diagram of the method M100 of an embodiment.

Figure 31 a shows according to the process flow diagram of the method M200 of an embodiment.

The process flow diagram of the embodiment M210 of Figure 31 b methods of exhibiting M200.

Figure 32 shows according to the process flow diagram of the method M300 of an embodiment.

In each figure and the description of enclosing, same reference numerals represents same or similar element or signal.

Embodiment

Embodiment described herein comprises and can be configured to provide and extend that bandwidth only increases by approximately 800 to 1000bps (bps) system, method and apparatus to support transmission and/or storage broadband voice signal to narrow-band speech coder.The potential advantage of this type of embodiment comprise support with narrow band system compatible embedded encoded, between narrow-band and high frequency band encoding channel, relatively easily distribute and reallocate, avoid the broadband synthetic operation that calculated amount is larger, and maintain the low sampling rate for the treatment of by the signal of the larger waveform coding routine processing of calculated amount.

For example, unless be subject to especially context limited, otherwise used term " calculating " to represent any one of its its ordinary meaning herein, calculate, produce and select from the list of value.When this is described and uses term " to comprise " in content and claims, do not get rid of other element or operation.Use term " A is based on B " represents any one of its its ordinary meaning, comprises following situation: (i) " A equals B " and (ii) " A is at least based on B ".Term " Internet Protocol " comprises edition 4 and the subsequent version (for example, version 6) of describing as in IETF (the Internet engineering work group) RFC (Request for Comment) 791.

Fig. 1 a shows according to the block scheme of the wideband speech coding device A100 of an embodiment.Bank of filters A110 is configured to broadband voice signal S10 to carry out filtering to produce narrow-band signal S20 and high-frequency band signals S30.Narrowband encoder A120 is configured to encode narrow-band signal S20 to produce narrow-band (NB) filter parameter S40 and narrow-band residual signal S50.As described in further detail herein, narrowband encoder A120 is conventionally configured to as codebook index or takes another quantized versions and produce narrow band filter parameter S 40 and encoded narrow-band pumping signal S50.High band encoder A200 is configured to according to the information coding high-frequency band signals S30 in encoded narrow-band pumping signal S50 to produce high frequency band coding parameter S60.As described in further detail herein, high band encoder A200 is conventionally configured to as codebook index or takes another quantized versions and produce high frequency band coding parameter S60.A particular instance of wideband speech coding device A100 is configured to the broadband voice signal S10 that encodes under the speed of about 8.55kbps (kbps), wherein about 7.55kbps is used for narrow band filter parameter S 40 and encoded narrow-band pumping signal S50, and about 1kbps is for high frequency band coding parameter S60.

May need encoded narrow-band and high-frequency band signals to be combined as single bit stream.For instance, may need to using as encoded broadband voice signal for example, for transmission (, on wired, optics or wireless transmission channel) or for storage encoded signal is multiplexed together.Fig. 1 b shows the block scheme of the embodiment A102 of wideband speech coding device A100, wideband speech coding device A100 comprises multiplexer A130, and it is configured to narrow band filter parameter S 40, encoded narrow-band pumping signal S50 and high band filter parameter S 60 to be combined as multiplex signal S70.

A kind of equipment that comprises scrambler A102 also can comprise and is configured to multiplex signal S70 to be transferred to the circuit in the transmission channel of wired, optics for example or wireless channel.This equipment also can be configured to signal to carry out one or more chnnel coding operations, for example error correction code (for example, rate-compatible convolutional encoding) and/or error detection code (for example, cyclic redundancy code), and/or one or more layers procotol coding (for example, Ethernet, TCP/IP, cdma2000).

May need multiplexer A130 to be configured to embed encoded narrow-band signal (comprising narrow band filter parameter S 40 and encoded narrow-band pumping signal S50) as separable multiplex signal S70 tributary, the another part (for example, high frequency band and/or low band signal) that makes encoded narrow-band signal can be independent of multiplex signal S70 is resumed and decodes.For instance, multiplex signal S70 can be configured encoded narrow-band signal can be recovered by divesting high band filter parameter S 60.A potential advantage of this feature is to avoid needing described encoded broadband signal to carry out code conversion before the system that encoded broadband signal is delivered to the decoding of supporting the decoding of narrow-band signal but not supporting highband part.

Fig. 2 a shows according to the block scheme of the broadband Voice decoder B100 of an embodiment.Narrow-band demoder B110 is configured to decode narrow band filter parameter S 40 and encoded narrow-band pumping signal S50 to produce narrow-band signal S90.High band decoder B200 is configured to the high frequency band coding parameter S60 that decodes according to narrow-band pumping signal S80 based on encoded narrow-band pumping signal S50, to produce high-frequency band signals S100.In this example, narrow-band demoder B110 is configured to narrow-band pumping signal S80 to be provided to high band decoder B200.Bank of filters B120 is configured to narrow-band signal S90 and high-frequency band signals S100 to combine to produce broadband voice signal S110.

Fig. 2 b is the block scheme of the embodiment B102 of broadband Voice decoder B100, and broadband Voice decoder B100 comprises demultiplexer B130, and it is configured to produce encoded signal S40, S50 and S60 from multiplex signal S70.A kind of equipment that comprises demoder B102 can comprise the circuit that is configured to receive from the transmission channel of for example wired, optics or wireless channel multiplex signal S70.This equipment also can be configured to signal to carry out one or more channel-decoding operations, for example error correction decoding (for example, rate-compatible convolution decoder) and/or error-detecting decoding (for example, cyclic redundancy decoding), and/or one or more layers procotol decoding (for example, Ethernet, TCP/IP, cdma2000).

Bank of filters A110 is configured to, according to division frequency band scheme, input signal is carried out to filtering to produce low frequency sub-band and high-frequency sub-band.Depending on the design standards of application-specific, output sub-band can have the bandwidth that equates or do not wait and possible overlapping or not overlapping.The bank of filters A110 configuration that produces two above sub-bands is also possible.For instance, this bank of filters can be configured to produce one or more low band signal, and described low band signal comprises for example, component in the frequency range (scope of 50-300Hz) below the frequency range in narrow-band signal S20.This bank of filters also may be configured to produce one or more extra high-frequency band signals, and described extra high-frequency band signals comprises for example, component in frequency range (scope of 14-20,16-20 or 16-32kHz) more than the frequency range in high-frequency band signals S30.In the case, can implement wideband speech coding device A100 with this signal (one or more) of encoding separately, and multiplexer A130 can be configured to extra encoded signal (one or more) to be included in (for example,, as removable part) in multiplex signal S70.

Fig. 3 a shows the block scheme of the embodiment A112 of bank of filters A110, and bank of filters A110 is configured to produce the two sub-frequency bands signals with the sampling rate reducing.Bank of filters A110 is configured to receive the broadband voice signal S10 with high-frequency (or high frequency band) part and low frequency (or low-frequency band) part.Bank of filters A112 comprises the low-frequency band processing path that is configured to receive broadband voice signal S10 and produces narrow-band voice signal S20, and the high frequency band processing path that is configured to receive broadband voice signal S10 and produces high frequency band voice signal S30.110 couples of broadband voice signal S10 of low-pass filter carry out filtering with the low frequency sub-band by selected, and 130 couples of broadband voice signal S10 of Hi-pass filter carry out filtering with the high-frequency sub-band by selected.Because the bandwidth ratio broadband voice signal S10 of two sub-frequency bands signals is narrow, so its sampling rate can reduce to a certain extent without information loss.Downsampled device 120 (is for example selected factor according to required, by removing the sample of signal and/or replacing sample with mean value) reduce the sampling rate of low-pass signal, and downsampled device 140 similarly required is selected the sampling rate that factor reduces high communication number according to another.

Fig. 3 b shows the block scheme of the corresponding embodiment B122 of bank of filters B120.Upwards sampler 150 (for example, fill in and/or pass through reproduction copies by zero) increases the sampling rate of narrow-band signal S90, and 160 pairs of low-pass filters upwards sampled signal carry out filtering with only by low-frequency band part (for example,, to prevent aliasing).Equally, upwards sampler 170 increases the sampling rate of high-frequency band signals S100, and 180 pairs of Hi-pass filters upwards sampled signal carry out filtering only to pass through highband part.Then, two passband signals are sued for peace to form broadband voice signal S110.In some embodiments of demoder B100, bank of filters B120 is configured to produce according to one or more flexible strategy that received and/or calculated by high band decoder B200 the weighted sum of two passband signals.Also expect the bank of filters B120 configuration of two above passband signal combination.

Each be embodied as finite impulse response (FIR) (FIR) wave filter of wave filter 110,130,160,180 or be embodied as infinite impulse response (IIR) wave filter.Encoder filters 110 and 130 frequency response can have symmetry between stopband and passband or the limited proportionality of dissimilar shape.Equally, demoder wave filter 160 and 180 frequency response can have symmetry between stopband and passband or the limited proportionality of dissimilar shape.May need (but not being to be strictly necessary) low-pass filter 110 and low-pass filter 160 to there is same response, and Hi-pass filter 130 have same response with Hi-pass filter 180.In an example, two wave filters are quadrature mirror filter (QMF) groups to 110,130 and 160,1 80, its median filter to 110,130 and wave filter to 160,180, there is same factor.

In representative instance, low-pass filter 110 has the passband (for example, 0 to 4kHz frequency band) of the limited PSTN scope that comprises 300-3400Hz.The relative bandwidth of broadband voice signal S10, narrow-band signal S20 and high-frequency band signals S30 in Fig. 4 a and two different embodiments of 4b displaying.In these two particular instances, broadband voice signal S10 has the sampling rate (representing the frequency component within the scope of 0 to 8kHz) of 16kHz, and narrow-band signal S20 has the sampling rate (representing the frequency component within the scope of 0 to 4kHz) of 8kHz.

In the example of Fig. 4 a, between two sub-frequency bands, do not exist obviously overlapping.The Hi-pass filter 130 that high-frequency band signals S30 shown in this example can have a passband of 4-8kHz by use obtains.In the case, may need 2 as factor is downsampled, through filtering signal, sampling rate to be reduced to 8kHz by take.Can expect that this operation can significantly reduce the computational complexity to the further processing operation of signal, this operation will move down into passband energy within the scope of 0 to 4kHz and without information loss.

In the alternate example of Fig. 4 b, upper and lower sub-band has obviously overlapping, makes two sub-frequency bands signals all describe 3.5 to 4kHz region.The Hi-pass filter 130 that high-frequency band signals S30 in this example can have a passband of 3.5-7kHz by use obtains.In the case, may need 16/7 as factor is downsampled, through filtering signal, sampling rate to be reduced to 7kHz by take.Can expect that this operation can significantly reduce the computational complexity to the further processing operation of signal, this operation will move down into passband energy within the scope of 0 to 3.5kHz and without information loss.

In call type code communication handset, the obvious response in the frequency range of one or more transducers (that is, microphone and earphone or loudspeaker) shortage 7-8kHz.In the example of Fig. 4 b, broadband voice signal S10 in 7 and 8kHz between part be not included in encoded signal.Other particular instance of Hi-pass filter 130 has the passband of 3.5-7.5kHz and 3.5-8kHz.

In some embodiments, as provided overlapping permission to use low pass and/or the Hi-pass filter on overlay region with level and smooth decay between sub-band in the example at Fig. 4 b.This type of wave filter is compared with the wave filter having more sharply or " brickwall (brick-wall) " responds, and more easily design is conventionally not too complicated in calculating, and/or the delay causing is less.The wave filter with drastic shift district often has higher secondary lobe (this may cause aliasing) than the wave filter with the similar grade of level and smooth decay.The wave filter with drastic shift district also may have long impulse response, and this may cause ring illusion.For the bank of filters embodiment with one or more iir filters, allow the level and smooth decay on overlay region can make it possible to use limit away from the wave filter (one or more) of unit circle, this is for guaranteeing that stable fixed point embodiment may be more important.

The overlapping permission low-frequency band of sub-band is mixed with the level and smooth of high frequency band, and this can cause the less illusion of hearing, reduces aliasing, and/or makes a frequency band not too obvious to the transformation of another frequency band.In addition, the code efficiency of narrowband encoder A120 (for example, wave coder) can constantly increase and decline along with frequency.For instance, may, under low bitrate, especially in the situation that there is ground unrest, reduce the coding quality of narrowband encoder.In such cases, provide sub-band the overlapping quality that improves the frequency component copying in overlay region.

In addition, the overlapping permission low-frequency band of sub-band is mixed with the level and smooth of high frequency band, and it may cause the less illusion of hearing, reduces aliasing, and/or makes a frequency band not too obvious to the transformation of another frequency band.This feature may especially cater to the need according to the embodiment of different coding method operation for narrowband encoder A120 and high band encoder A200.For instance, different coding technology can produce the signal that sounds very different.Scrambler with codebook index form coding spectrum envelope can produce the signal with the sound different from the scrambler that changes coding amplitude spectrum into.Time domain coding device (for example, pulse-code modulated or PCM encoder) can produce the signal with the sound different from Frequency Domain Coding device.With the scrambler of the representation coded signal of spectrum envelope and corresponding residual signal can produce have from only with the signal of the different sound of the scrambler of spectrum envelope representation coded signal.The scrambler that is the representation of its waveform by Signal coding can produce the output with the sound different from output from sinusoidal coder.In such cases, with the wave filter with drastic shift district, define that non-overlapped sub-band may cause in synthetic broadband signal the transformation between sub-band more suddenly and sensuously more obvious.

Although conventionally use the QMF bank of filters with complementary overlapping frequency response in sub-band technology, this type of wave filter is unsuitable at least some embodiments in broadband coding embodiment described herein.The QMF bank of filters at scrambler place is configured to produce aliasing largely, described in be aliasing in the corresponding QMF bank of filters at demoder place and be eliminated.This configuration may be unsuitable for the application that signal causes a large amount of distortions between bank of filters, because distortion can reduce the effect of aliasing cancellation property.For instance, application described herein comprises be configured to the coding embodiment that operates under low-down bit rate.Because bit rate is very low, so probably compare and be rendered as obvious distortion with original signal through decoded signal, make the use of QMF bank of filters can cause the aliasing of not cancellation.Use the application of QMF bank of filters conventionally to there is high bit speed (for example, surpass 12kbps for AMR, and for G.722 surpassing 64kbps).

In addition, scrambler can be configured to be similar to original signal on sensigenous but in fact significantly be different from the composite signal of original signal.For instance, from narrow-band is residual, the scrambler of derivation high band excitation can produce this signal as described herein, because may not exist actual high frequency band residual completely in decoded signal.The distortion largely of using QMF bank of filters can cause the aliasing by not cancellation to cause in this type of application.

If affected children frequency band is narrower, can reduce the amount distortion that QMF aliasing causes so, because the impact of aliasing is limited to the bandwidth equating with sub-band width.Yet for half example of the pact that each sub-band comprises broadband bandwidth described herein wherein, the distortion being caused by the aliasing of not cancellation may affect the major part of signal.The quality of signal also may be subject to occur above the position influence of frequency band of the aliasing of not cancellation.Near the distortion that near the distortion that for instance, broadband voice signal center (for example, 3 and 4kHz between) produces may for example, occur than (, more than 6kHz) signal edge is much harmful.

Although the response of the wave filter of QMF bank of filters is strictly relevant each other, the low-frequency band of bank of filters A110 and B120 and high frequency band path can be configured to have complete incoherent frequency spectrum except two sub-frequency bands overlapping.The distance of the point of the drop to-20dB of frequency response of the some tremendously low frequency band filter of drop to-20dB of the frequency response that we are high band filter by the overlay defining of two sub-frequency bands.In the various examples of bank of filters A110 and/or B120, this overlapping range is that about 200Hz is to about 1kHz.Approximately 400 to the scope of about 600Hz can presentation code efficiency and perception smoothness between desired trade-off.In a particular instance mentioned above, overlap near 500Hz.

May need to implement the operation that bank of filters A112 and/or B122 illustrate in some stage execution graph 4a and 4b.For instance, Fig. 4 c shows the block scheme of the embodiment A114 of bank of filters A112, a series of interpolations for embodiment A114, samples, selects the high-pass filtering and the downsampled operation that operate to carry out function equivalent with other.This type of embodiment can more easily design and/or can allow to re-use the functional block of logic and/or code.For instance, can with identical function piece carry out as shown in Fig. 4 c to 14kHz select and to the operation of selecting of 7kHz.Can pass through signal function e ^{jn π}or sequence (1) ⁿmultiply each other to implement reversing spectrum operation, described sequence (1) ⁿvalue between+1 and-1 alternately.Spectrum shaping operation can be embodied as and be configured so that thereby signal shaping obtains the low-pass filter of required overall filter response.

Notice, due to the cause of reversing spectrum operation, the reversing spectrum of high-frequency band signals S30.The correspondingly subsequent operation in configuration codes device and respective decoder.For instance, high band excitation generator A300 described herein can be configured to produce the high band excitation signal S120 equally with reversing spectrum form.

Fig. 4 d shows the block scheme of the embodiment B124 of bank of filters B122, and a series of interpolations bank of filters B122 for, sampling and other operate to carry out function equivalent again upwards sampling and high-pass filtering operate.Bank of filters B124 comprises the reversing spectrum operation in high frequency band, and it makes the middle similar operation reversion of carrying out of bank of filters (for example, bank of filters A114) with for example scrambler.In this particular instance, bank of filters B124 also comprises the notch filter in low-frequency band and high frequency band, and it weakens the component of signal at 7100Hz place, but this type of wave filter is optional and does not need to comprise this type of wave filter.The patent application case " SYSTEMS; METHODS; AND APPARATUS FOR SPEECH SIGNAL FILTERING " that together US publication of application is 2007/0088558 therewith comprises about the additional description of the response of the element of the particular of bank of filters A110 and B120 and graphic, and this material is incorporated to by reference at this.

Narrowband encoder A120 implements according to source-filter model, and it is encoded to input speech signal the pumping signal that (A) a group describes the parameter of wave filter and (B) drive the synthetic duplicate of described wave filter generation input speech signal.Fig. 5 a shows the example of the spectrum envelope of voice signal.The peak value that shows the feature of this spectrum envelope represents the resonance of voice range and is called resonance peak.Most of speech coders to this rough spectrum structured coding of major general is one group of parameter (for example, filter coefficient).

Fig. 5 b shows as is applied to the example of elementary sources-filter configuration of coding of the spectrum envelope of narrow-band signal S20.Analysis module calculates one group of parameter describing corresponding to the wave filter of the speech sound in a period of time (common 20 milliseconds).According to the prewhitening filter of those filter parameter configurations (also referred to as analysis or prediction error filter), remove spectrum envelope so that signal is carried out to frequency spectrum leveling.Therefore the whitened signal of gained (also referred to as residual) is compared with primary speech signal, has less energy and changes less and compared with easy-encode.Because the error that residual signal coding is caused also may be dispersed on frequency spectrum more equably.Filter parameter and residual common through quantizing for effectively transmission on channel.At demoder place, according to the composite filter of filter parameter configuration, by the signal based on residual, encouraged to produce the synthetic version of raw tone sound.Composite filter is configured to have transfer function conventionally, and described transfer function is the inverted versions that turns function of prewhitening filter.

Fig. 6 shows the block scheme of the basic embodiment A122 of narrowband encoder A120.In this example, linear predictive coding (LPC) analysis module 210 is encoded to one group of linear prediction (LP) coefficient (for example, the coefficient 1/A (z) of full utmost point wave filter) by the spectrum envelope of narrow-band signal S20.Analysis module is treated to input signal a series of non-overlapped frames conventionally, wherein for each frame calculates one group of new coefficient.Frame period normally can expect the cycle that signal is static in this locality; A common example is 20 milliseconds (being equivalent to lower 160 samples of sampling rate of 8kHz).In an example, lpc analysis module 210 is configured to calculate one group of 10 LP filter coefficient to describe the resonance peak structure of each 20 milliseconds of frame.Also may implement analysis module input signal is treated to a series of overlapping frame.

Analysis module can be configured to the sample of each frame of Direct Analysis, or can first for example, according to window function (, Hamming window), to sample, be weighted.Also the upper execution analysis of the window of frame (for example, 30 milliseconds of windows) can be greater than.This window can be symmetrical (5-20-5 for example, make its comprise 20 milliseconds of frames and then before and afterwards 5 milliseconds) or asymmetric (for example 10-20, last 10 milliseconds of making it comprise previous frame).Lpc analysis module is configured to use Levinson-Durbin recursion or Leroux-Gueguen algorithm to calculate LP filter coefficient conventionally.In another embodiment, analysis module can be configured to calculate one group of cepstrum coefficient rather than one group of LP filter coefficient of each frame.

The output speed of scrambler A120 can significantly reduce by quantizing filter coefficient, and relatively little on the impact of reproduction quality.Coefficient of linear prediction wave filter is difficult to effective quantification, and is conventionally mapped as another representation, and for example line spectrum pair (LSP) or line spectral frequencies (LSF), encode for quantification and/or entropy.In the example of Fig. 6, LP filter coefficient-LSF conversion 220 is transformed to one group of corresponding LSF by described group of LP filter coefficient.Other of LP filter coefficient one to one representation comprises partial autocorrelation coefficient, log area ratio value, adpedance and composes (ISP) and immittance spectral frequencies (ISF), and it is for GSM (global system for mobile communications) AMR-WB (the how fast broadband of self-adaptation) coder.Conventionally, the conversion between one group of LP filter coefficient and one group of corresponding LSF is reversible, but embodiment also comprises the scrambler A120 embodiment that conversion can not be reversible in error free situation.

Quantizer 230 is configured to quantize described group of narrow-band LSF (or other coefficient representation), and the result that narrowband encoder A122 is configured to export this quantification is as narrow band filter parameter S 40.This quantizer comprises vector quantizer conventionally, and it is encoded to the index for the corresponding vectorial entry in table or code book by input vector.

As shown in Figure 6, narrowband encoder A122 is also by making narrow-band signal S20 by (also referred to as analyzing or prediction error filter) produces residual signal according to the prewhitening filter 260 of described group of filter coefficient configuration.In this particular instance, prewhitening filter 260 is embodied as FIR wave filter, but also can use IIR embodiment.This residual signal will contain in narrow band filter parameter S 40 the sensuously more important speech frame information not representing, for example long-term structure relevant with tone conventionally.The quantization means form that quantizer 270 is configured to calculate this residual signal is usingd as encoded narrow-band pumping signal S50 output.This quantizer comprises vector quantizer conventionally, and it is encoded to the index for the corresponding vectorial entry in table or code book by input vector.Or this quantizer can be configured to send one or more parameters, can from described parameter, dynamically produce at demoder place vector, rather than as in sparse code book method, from memory storage, retrieve vector.The method is for the encoding scheme of for example algebraically CELP (code book Excited Linear Prediction) and the coder of for example 3GPP2 (third generation partnership relation 2) EVRC (strengthening variable bit rate coder).

Need narrowband encoder A120 according to the same filter parameter value that can be used for corresponding narrow-band demoder is produced to encoded narrow-band pumping signal.In this way, the encoded narrow-band pumping signal of gained may be considered the undesirable property of those parameter values, for example quantization error to a certain extent.Therefore, need to by demoder place can with same tie numerical value configure prewhitening filter.In the basic example of scrambler A122 as shown in Figure 6,240 pairs of narrow-band coding parameter S40 de-quantizations of quantizer, LSF-LP filter coefficient conversion 250 is shone upon back one group of corresponding LP filter coefficient by income value, and this group coefficient is for configuring prewhitening filter 260 to produce the residual signal being quantized by quantizer 270.

Some embodiments of narrowband encoder A120 are configured to calculate encoded narrow-band pumping signal S50 by identifying with a vector of residual signal optimum matching from one group of code book vector.Yet, noticing, narrowband encoder A120 also can be through implementing in fact not produce residual signal to calculate the quantization means form of residual signal.For instance, narrowband encoder A120 can be configured to (for example produce corresponding composite signal with many code book vectors, according to one group of current filter parameter), and in selection and perceptual weighting territory and the code book that joins of institute's signal correction that produces of original narrow-band signal S20 optimum matching vectorial.

Fig. 7 shows the block scheme of the embodiment B112 of narrow-band demoder B110.310 pairs of narrow band filter parameter S 40 de-quantizations of quantizer (in the case, de-quantization is one group of LSF), and the conversion 320 of LSF-LP filter coefficient is transformed to one group of filter coefficient (for example,, as the quantizer 240 above with reference to narrowband encoder A122 and conversion 250 are described) by LSF.340 pairs of encoded narrow-band pumping signal S50 de-quantizations of quantizer are to produce narrow-band pumping signal S80.Based on filter coefficient and narrow-band pumping signal S80, narrow-band composite filter 330 synthesis of narrow band signal S90.In other words, narrow-band composite filter 330 is configured to, according to de-quantization filter coefficient, narrow-band pumping signal S80 is carried out to spectrum shaping, to produce narrow-band signal S90.Narrow-band demoder B112 is also provided to high band encoder A200 by narrow-band pumping signal S80, and high band encoder A200 derives high band excitation signal S120 with narrow-band pumping signal S80, as described herein.In embodiments more described below, narrow-band demoder B110 can be configured to the extraneous information relevant with narrow-band signal (for example, spectral tilt, pitch gain and hysteresis, and speech pattern) to be provided to high band decoder B200.

The system of narrowband encoder A122 and narrow-band demoder B112 is the basic example of synthesis analysis Codec.Code book Excited Linear Prediction (CELP) coding is the general series of of synthesis analysis coding, and the embodiment of this type of scrambler can be carried out residual waveform coding, wherein comprise for example selector bar object operation from fixing and self-adaptation code book, error minimize operation and/or perceptual weighting operation.Other embodiment of synthesis analysis coding comprises MELP (Mixed Excitation Linear Prediction) (MELP), algebraically CELP (ACELP), lax CELP (RCELP), Regular-Pulse Excitation (RPE), multiple-pulse CELP (MPE) and vector sum Excited Linear Prediction (VSELP) coding.Correlative coding method comprises multi-band excitation (MBE) and prototype waveform interpolation (PWI) coding.ETSI (ETSI) the GSM full rate coder (GSM06.10) that the example of standard synthesis analysis Codec comprises use residual excited linear predictive (RELP), GSM EFR coder (ETSI-GSM 06.60), ITU (International Telecommunications Union (ITU)) standard 11.8kb/s be the IS of Annex E scrambler, IS-136 (interim standard) 641 coders (time division multiple access (TDMA) scheme), GSM adaptive multi-rate (GSM-AMR) coder G.729, and 4GV ^tM(the 4th generation Vocoder ^tM) coder (Qualcomm of Diego California, California (QUALCOMM Incorporated, San Diego, CA)).Narrowband encoder A120 and corresponding demoder B110 can implement according to any one or any other speech coding technology in these technology (known or leaved for development), and described speech coding technology is expressed as voice signal that (A) a group describes the parameter of wave filter and (B) for driving the pumping signal of described wave filter reproduction speech signal.

Even after prewhitening filter is removed rough spectrum envelope from narrow-band signal S20, also may retain quite a large amount of meticulous harmonic structure (especially for turbid voice).The spectrum curve of an example of the residual signal (as produced by prewhitening filter) of Fig. 8 a displaying voiced sound signal (for example, vowel).In this example, visible periodic structure is relevant with tone, and the different voiced sounds that same speaker sends may have different resonance peak structure but have similar tone structure.Fig. 8 b shows the time-domain curve of the example of this residual signal, and it shows the time series of tone pulses.

Can be by increasing code efficiency and/or voice quality by the characteristic of one or more parameter value coding tone structures.The frequency (also referred to as fundamental frequency) that a key property of tone structure is first harmonic, they are conventionally within the scope of 60 to 400Hz.This characteristic is encoded to the inverted versions of fundamental frequency conventionally, also referred to as pitch lag (pitch lag).The number of sample in a pitch period of pitch lag indication, and may be encoded as one or more codebook index.Voice signal from male speaker often recently has larger pitch lag from women speaker's voice signal.

Another characteristics of signals relevant with tone structure is periodically, the intensity of its indication harmonic structure, or in other words, signal is harmonic wave or non-harmonic degree.Periodic two typical designators are zero crossing and normalized autocorrelation function (NACF).Periodically also can be indicated by pitch gain, described pitch gain is encoded to code book gain (for example, quantizing the gain of self-adaptation code book) conventionally.

Narrowband encoder A120 can comprise one or more modules of the long-term harmonic structure of the narrow-band signal S20 that is configured to encode.As shown in Figure 9, a spendable typical CELP example comprises open loop lpc analysis module, and its coding short-term characteristic or rough spectrum envelope, be the closed type loop Long-run Forecasting Analysis stage afterwards, described stage coding fine pitch or harmonic structure.Short-term characteristic is encoded to filter coefficient, and long-time quality is encoded to for example value of the parameter of pitch lag and pitch gain.For instance, narrowband encoder A120 can be configured so that for example, to comprise the encoded narrow-band pumping signal of the formal output S50 of one or more codebook index (, this index of fixed password and self-adaptation codebook index) and corresponding yield value.The calculating of this quantization means form of narrow-band residual signal (for example,, by quantizer 270) can comprise to be selected these index and calculates these values.The coding of tone structure also can comprise interpolation pitch prototype waveform, and described operation can comprise calculates poor between continuous tone pulse.Can be for the modeling of forbidding long-term structure corresponding to the frame of clear voice (it is similar to noise and not systematization conventionally).

According to the embodiment of the narrow-band demoder B110 of the example shown in Fig. 9, can be configured to, after long-term structure (tone or harmonic structure) has been recovered, narrow-band pumping signal S80 is outputed to high band decoder B200.For instance, this demoder can be configured to export narrow-band pumping signal S80 as the de-quantization version of encoded narrow-band pumping signal S50.Certainly, also may implement narrow-band demoder B110, make de-quantization that high band decoder B200 carries out encoded narrow-band pumping signal S50 to obtain narrow-band pumping signal S80.

According in the embodiment of the wideband speech coding device A100 of the example shown in Fig. 9, high band encoder A200 can be configured to receive the narrow-band pumping signal being produced by short run analysis or prewhitening filter.In other words, narrowband encoder A120 can be configured to, before the long-term structure of coding, narrow-band pumping signal is outputed to high band encoder A200.Yet high band encoder A200 need to receive the same-code information being received by high band decoder B200 from narrow-band channel, the coding parameter that high band encoder A200 is produced may be considered the undesirable property of described information to a certain extent.Therefore, may be preferably, high band encoder A200 is from treating the identical parameters of being exported by wideband speech coding device A100 and/or quantizing to rebuild narrow-band pumping signal S80 in encoded narrow-band pumping signal S50.A potential advantage of the method is to calculate more exactly high frequency band gain factor S60b described below.

Except describing the short-term of narrow-band signal S20 and/or the parameter of long-term structure, narrowband encoder A120 also can produce the parameter value relevant with other characteristic of narrow-band signal S20.These values (its may through suitable quantification to be exported by wideband speech coding device A100) can be included in narrow band filter parameter S 40 or output separately.High band encoder A200 also can be configured to for example, calculate high frequency band coding parameter S60 according to one or more (, after de-quantization) in these additional parameter.At broadband Voice decoder B100 place, high band decoder B200 can be configured to for example, receive parameter value via narrow-band demoder B110 (, after de-quantization).Or high band decoder B200 can be configured to direct reception (and may for de-quantization) parameter value.

In an example of additional narrow frequencyband coding parameter, narrowband encoder A120 produces the spectral tilt of each frame and the value of speech pattern parameter.Spectral tilt is relevant with the shape of spectrum envelope in passband, and conventionally by quantizing the first reflection coefficient, represents.For most of voiced sounds, spectrum energy reduces along with the continuous increase of frequency, makes the first reflection coefficient for bearing and can approaching-1.Most of voicelesss sound have smooth frequency spectrum, thereby make the first reflection coefficient approach zero, or under high-frequency, have more energy, thereby make the first reflection coefficient for just and can approach+1.

Speech pattern (also referred to as sounding pattern) indication present frame represents turbid voice or clear voice.This parameter can have binary value, its for example, voice activity (for example, the relation between this index and threshold value) based on periodic one or more indexs (, zero crossing, NACF, pitch gain) and/or frame.In other embodiments, speech pattern parameter has one or more other states with the indication pattern of the transformation between noiseless or ground unrest or noiseless and turbid voice for example.

High band encoder A200 is configured to according to the source-filter model high-frequency band signals S30 that encodes, and wherein the excitation of this wave filter is based on encoded narrow-band pumping signal.Figure 10 shows the block scheme of the embodiment A202 of high band encoder A200, and high band encoder A200 is configured to produce the high frequency band coding parameter S60 stream that comprises high band filter parameter S 60a and high frequency band gain factor S60b.High band excitation generator A300 derives high band excitation signal S120 from encoded narrow-band pumping signal S50.Analysis module A210 produces one group of parameter value of the spectrum envelope of describing high-frequency band signals S30.In this particular instance, analysis module A210 is configured to carry out lpc analysis to produce one group of LP filter coefficient for each frame of high-frequency band signals S30.Coefficient of linear prediction wave filter-LSF conversion 410 is transformed to one group of corresponding LSF by described group of LP filter coefficient.As described in above with reference to analysis module 210 and conversion 220, analysis module A210 and/or conversion 410 can be configured to use other coefficient sets (for example, cepstrum coefficient) and/or coefficient representation is (for example, ISP).

Quantizer 420 is configured to quantize described group of high frequency band LSF (or other coefficient representation, for example ISP), and the result that high band encoder A202 is configured to export this quantification is as high band filter parameter S 60a.This quantizer comprises vector quantizer conventionally, and it is encoded to the index for the corresponding vectorial entry in table or code book by input vector.

High band encoder A202 also comprises composite filter A220, and its high band excitation signal S120 and encoded spectrum envelope (for example, described group of LP filter coefficient) that is configured to produce according to analysis module A210 produces synthetic high-frequency band signals S130.Composite filter A220 is embodied as iir filter conventionally, but also can use FIR embodiment.In particular instance, composite filter A220 is embodied as sextic autoregressive filter.

It is poor that high frequency band gain factor counter A230 calculates one or more between original high-frequency band signals S30 and the level of synthetic high-frequency band signals S130, with the gain envelope of designated frame.Quantizer 430 can be embodied as input vector is encoded to the vector quantizer for the index of the corresponding vectorial entry in table or code book, it quantizes to specify the value (one or more) of gain envelope, and the result that high band encoder A202 is configured to export this quantification is as high frequency band gain factor S60b.

In embodiment as shown in figure 10, composite filter A220 is configured to from analysis module A210 receiving filter coefficient.The alternate embodiment of high band encoder A202 comprises quantizer and inverse transform, it is configured to decoding filter coefficient from high band filter parameter S 60a, and composite filter A220 is configured to change into reception through decoding filter coefficient in the case.This alternative arrangements can be supported the calculated gains envelope more exactly by high frequency band gain calculator A230.

In a particular instance, the respectively every frame output hexad LSF of analysis module A210 and high frequency band gain calculator A230 and one group of five yield value, make only by 11 bonus values of every frame, to realize the broadband extension of narrow-band signal S20.Ear is often more insensitive for the frequency error under high-frequency, thereby the high frequency band of low LPC level coding can produce the signal with the perceived quality that can compare with the narrow-band coding of higher LPC level.The typical embodiments of high band encoder A200 can be configured to 8 to 12 high-quality for spectrum envelope of every frame output and rebuild, and every frame is exported other 8 to 12 high-quality for temporal envelope reconstruction.In another particular instance, the every frame of analysis module A210 is exported one group of eight LSF.

Some embodiments of high band encoder A200 are configured to produce in the following manner high band excitation signal S120: produce and have the random noise signal of high-band frequency component, and according to the temporal envelope of narrow-band signal S20, narrow-band pumping signal S80 or high-frequency band signals S30, noise signal is carried out to which amplitude modulation.Although this method based on noise can produce suitable result for voiceless sound, yet it may be undesirable for voiced sound, and the residual of voiced sound is generally harmonic wave and therefore has certain periodic structure.

High band excitation generator A300 is configured in high-band frequency range, produce high band excitation signal S120 by the frequency spectrum of narrow-band pumping signal S80 is extended to.Figure 11 shows the block scheme of the embodiment A302 of high band excitation generator A300.Quantizer 450 is configured to the encoded narrow-band pumping signal of de-quantization S50 to produce narrow-band pumping signal S80.Frequency spectrum stretcher A400 is configured to produce harmonic wave based on narrow-band pumping signal S80 and extends signal S160.Combiner 470 is configured to the temporal envelope of the random noise signal of noise generator 480 generations and 460 calculating of envelope counter to combine to produce through zoop signal S170.Combiner 490 is configured to harmonic wave extend signal S160 and mix to produce high band excitation signal S120 through zoop signal S170.

In an example, frequency spectrum stretcher A400 is configured to narrow-band pumping signal S80 to carry out spectrum folding operation (also referred to as mirror) to produce harmonic wave extension signal S160.Spectrum folding can be filled in pumping signal S80 execution and then apply Hi-pass filter by zero and retain false signal.In another example, frequency spectrum stretcher A400 is configured to (for example,, via upwards sampling, multiply each other afterwards with constant frequency cosine signal) and produce harmonic wave extension signal S160 by narrow-band pumping signal S80 frequency spectrum being translated in high frequency band.

Spectrum folding and translation method can produce harmonic structure and extend signal with the discontinuous frequency spectrum of original harmonic structure of narrow-band pumping signal S80 in phase place and/or frequency.For instance, these class methods can produce the signal of the peak value with the multiple place that is not conventionally positioned at fundamental frequency, and this may cause the illusion of microphonia in the voice signal of rebuilding.These methods produce and have the unnatural high-frequency harmonic wave of adjusting characteristic compared with forte toward contact.Yet, because PSTN signal can sample under 8kHz, bandwidth is restricted to and is not more than 3400Hz, so the top frequency spectrum of narrow-band pumping signal S80 may contain seldom or not contain energy, make to translate according to spectrum folding or frequency spectrum the extension signal that operation produces and can there is spectral hole more than 3400Hz.

Other method that produces harmonic wave extension signal S160 comprises one or more fundamental frequencies of identifying narrow-band pumping signal S80, and according to described information, produces homophonic.For instance, the harmonic structure of pumping signal can be described together with phase information with amplitude by fundamental frequency.Another embodiment of high band excitation generator A300 produces harmonic wave extension signal S160 based on fundamental frequency and amplitude (for example,, as indicated by pitch lag and pitch gain).Yet, remove anharmonic wave extension signal and narrow-band pumping signal S80 relevant in phase place, otherwise the quality through decoded speech of gained may be unacceptable.

Useful nonlinear function produce with narrow-band excitation phase on phase dry doubling keep harmonic structure and there is no the high band excitation signal of phase discontinuity.Nonlinear function also can provide the noise level of the increase between high-frequency harmonic wave, and its tone high-frequency harmonic wave often producing than the method for translating by for example spectrum folding and frequency spectrum sounds more natural.The memoryless nonlinear function of typical case that can be applied by the various embodiments of frequency spectrum stretcher A400 comprises ABS function (also referred to as full-wave rectification), half-wave rectification, square, cube and slicing.Other embodiment of frequency spectrum stretcher A400 can be configured to the nonlinear function that application has memory.

Figure 12 is the block scheme of the embodiment A402 of frequency spectrum stretcher A400, and frequency spectrum stretcher A400 is configured to apply the frequency spectrum that nonlinear function extends narrow-band pumping signal S80.Upwards sampler 510 is configured to the sampling that makes progress to narrow-band pumping signal S80.May need signal to carry out fully upwards sampling so that aliasing during application nonlinear function minimizes.In a particular instance, upwards sampler 510 take 8 as factor is to the signal sampling that makes progress.Upwards sampler 510 can be configured to fill in and result carried out to low-pass filtering and carry out upwards sampling operation by input signal being carried out to zero.Nonlinear function counter 520 is configured to nonlinear function to be applied to through the sampled signal that makes progress.For frequency spectrum, extend (for example, square), ABS function is not need energy scale with respect to a potential advantage of other nonlinear function.In some embodiments, can effectively apply ABS function by peeling off or remove the sign bit of each sample.Nonlinear function counter 520 also can be configured to carry out through upwards sampling or frequency spectrum extend the amplitude deviation of signal.

Downsampled device 530 is configured to that the frequency spectrum of application nonlinear function is extended to result and carries out downsampled.Downsampled device 530 may need to carry out bandpass filtering operation to select before frequency spectrum to extend the required frequency band of signal reducing sampling rate (for example,, so that the aliasing or the error that reduce or avoid causing due to unnecessary image).Downsampled device 530 may also need to reduce sampling rate in more than one stage.

Figure 12 a shows that frequency spectrum extends the figure of the signal spectrum at each point place in the example operating, and wherein frequency scaling is identical on each curve.Curve (a) is shown the frequency spectrum of an example of narrow-band pumping signal S80.Curve (b) shows that signal S80 take 8 for factor sampling frequency spectrum afterwards upwards.Curve (c) is shown the example of the extension frequency spectrum after application nonlinear function.Curve (d) is shown the frequency spectrum after low-pass filtering.In this example, passband extends to the upper frequency limit (for example, 7kHz or 8kHz) of high-frequency band signals S30.

Curve (e) is shown the frequency spectrum after the downsampled first stage, wherein makes sampling rate take and 4 reduces to obtain broadband signal as factor.Curve (f) shows and to carry out high-pass filtering operation to select to extend the frequency spectrum after the highband part of signal, and curve (g) shows the frequency spectrum after downsampled subordinate phase, wherein makes sampling rate take and 2 reduces as factor.In a particular instance, downsampled device 530 is carried out high-pass filtering and downsampled subordinate phase by the Hi-pass filter 130 and the downsampled device 140 that make broadband signal pass through bank of filters A112 other structure or the routine of same response (or have), to produce, has the frequency range of high-frequency band signals S30 and the frequency spectrum of sampling rate extends signal.

As visible in curve (g), downsampled its reversing spectrum that impels of the high communication number shown in curve (f).In this example, downsampled device 530 is also configured to signal to carry out spectrum inversion operation.Curve (h) is shown the result of application spectrum inversion operation, and described spectrum inversion operation can be passed through signal function e ^{jn π}or sequence (1) ⁿmultiply each other to carry out described sequence (1) ⁿvalue between+1 and-1 alternately.This operation is equivalent in frequency domain the digital spectrum translocation distance π of signal.Notice, also can operate to obtain identical result by and spectrum inversion downsampled with different order application.Upwards sampling and/or downsampled operation also can be configured to comprise sampling again and obtain and have the sampling rate of high-frequency band signals S30 (for example, frequency spectrum 7kHz) extends signal.

As above noticed, bank of filters A110 and B120 can be through implementing to make the one or both in narrow-band and high-frequency band signals S20, S30 have reversing spectrum form in the output of bank of filters A110, with reversing spectrum form, be encoded and decode, and in broadband voice signal S110 before output at bank of filters B120 place reversing spectrum again.Certainly, in the case, by the spectrum inversion operation not needing as shown in Figure 12 a, because high band excitation signal S120 will also need to have reversing spectrum form.

The upwards sampling of the frequency spectrum extension operation that frequency spectrum stretcher A402 carries out and each downsampled task can many different modes configurations and arrangements.For instance, Figure 12 b shows that frequency spectrum extends the figure of the signal spectrum at each point place in another example operating, and wherein frequency scaling is identical on each curve.Curve (a) is shown the frequency spectrum of an example of narrow-band pumping signal S80.Curve (b) shows that signal S80 take 2 for factor sampling frequency spectrum afterwards upwards.Curve (c) is shown the example of the extension frequency spectrum after application nonlinear function.In the case, accept contingent aliasing in upper frequency.

Curve (d) is shown the frequency spectrum after reversing spectrum operation.Curve (e) is shown the frequency spectrum after the downsampled single stage, wherein makes sampling rate take and 2 reduces to obtain required frequency spectrum extension signal for factor.In this example, described signal is taked reversing spectrum form, and can be used for processing in the embodiment of high band encoder A200 of the high-frequency band signals S30 take this form.

The frequency spectrum that nonlinear function counter 520 produces extends signal, and probably along with frequency increases, amplitude obviously reduces.Frequency spectrum stretcher A402 comprises frequency spectrum tenderizer 540, and it is configured to operating through the albefaction of sampled signal execution downwards.Frequency spectrum tenderizer 540 can be configured to carry out fixedly albefaction operation or carry out adaptive whitening operation.In the particular instance of adaptive whitening, frequency spectrum tenderizer 540 comprises: lpc analysis module, and it is configured to according to calculating one group of four filter coefficient through downward sampled signal; And four analysis filters, it is configured to, according to those coefficients, signal is carried out to albefaction.Other embodiment of frequency spectrum stretcher A400 comprises frequency spectrum tenderizer 540 and frequency spectrum is extended the configuration of signal operation before downsampled device 530.

Can implement high band excitation generator A300 and using output harmonic wave extension signal S160 as high band excitation signal S120.Yet, in some cases, only use harmonic wave to extend signal and as high band excitation, may cause the illusion that can hear.The harmonic structure of voice is conventionally not so good as in low-frequency band obvious in high frequency band, and in high band excitation signal, uses too much harmonic structure may cause buzz.This illusion may be especially obvious in the voice signal from women speaker.

Embodiment comprises the embodiment that is configured to harmonic wave to extend the high band excitation generator A300 that signal S160 mixes with noise signal.As shown in figure 11, high band excitation generator A302 comprises noise generator 480, and it is configured to produce random noise signal.In an example, noise generator 480 is configured to produce unit variance white pseudo-random noise signal, but in other embodiments, noise signal do not need for white and can there is the power density along with frequency change.Noise generator 480 may need to be configured to output noise signal as determinacy function to can copy at demoder place its state.For instance, noise generator 480 can be configured to output noise signal as previously for example, at the information (, narrow band filter parameter S 40 and/or encoded narrow-band pumping signal S50) of identical intraframe coding qualitative function really.

Before mixing with harmonic wave extension signal S160, the random noise signal that noise generator 480 produces can extend through which amplitude modulation the temporal envelope of the energy distribution in time of signal S160 to have approximate narrow-band signal S20, high-frequency band signals S30, narrow-band pumping signal S80 or harmonic wave.As shown in figure 11, high band excitation generator A302 comprises combiner 470, and the noise signal that its temporal envelope that is configured to calculate according to envelope counter 460 produces noise generator 480 is carried out which amplitude modulation.For instance, combiner 470 can be embodied as multiplier, and its temporal envelope that is configured to calculate according to envelope counter 460 comes the output of convergent-divergent noise generator 480 to produce through zoop signal S170.

As shown in the block scheme of Figure 13, in the embodiment A304 of high band excitation generator A302, envelope counter 460 is configured to calculate the envelope that harmonic wave extends signal S160.As shown in the block scheme of Figure 14, in the embodiment A306 of high band excitation generator A302, envelope counter 460 is configured to calculate the envelope of narrow-band pumping signal S80.The other embodiments of high band excitation generator A302 can otherwise be configured to harmonic wave, to extend signal S160 in time according to the position of narrow-band tone pulses and add noise.

Envelope counter 460 can be configured to carry out as the task of comprising a series of subtasks envelope and calculate.Figure 15 shows the process flow diagram of the example T100 of this task.Subtask T110 calculate envelope treat the signal of modeling (for example, narrow-band pumping signal S80 or harmonic wave extend signal S160) frame each sample square to produce square value sequence.Subtask T120 carries out smooth operation to square value sequence.In an example, subtask T120 applies an IIR low-pass filter according to following formula to sequence:

y(n)＝ax(n)+(1-a)y(n-1)， (1)

Wherein x is wave filter input, and y is wave filter output, Domain Index when n is, and a is the smoothing factor with the value between 0.5 and 1.The value of smoothing factor a can be fixing, or in alternate embodiment, can be according to the indication of noise in input signal and self-adaptation makes a in muting situation approach 1, and approach 0.5 in the situation that there is noise.Subtask T130 is applied to through each sample of level and smooth sequence square root function to produce temporal envelope.

This embodiment of envelope counter 460 can be configured to according to execute the task each subtask of T100 of serial and/or parallel mode.In the other embodiments of task T100, can be the logical operation of band before the T110 of subtask, the required frequency part of the signal that it is configured to select envelope to treat modeling, for example 3-4kHz scope.

Combiner 490 is configured to harmonic wave extend signal S160 and mix to produce high band excitation signal S120 through zoop signal S170.The embodiment of combiner 490 can be configured (for example) with high band excitation signal S120 is calculated as harmonic wave extend signal S160 with through zoop signal S170's and.This embodiment of combiner 490 can be configured to by extended signal S160 and/or to through zoop signal S170 application weighting factor to harmonic wave before summation, and high band excitation signal S120 is calculated as to weighted sum.Can calculate each this type of weighting factor according to one or more standards, and described weighting factor can be fixed value, or frame by frame or the adaptation value calculating on by sub-frame basis.

Figure 16 shows the block scheme of the embodiment 492 of combiner 490, and described embodiment 492 is configured to that high band excitation signal S120 is calculated as to harmonic wave and extends signal S160 and weighted sum through zoop signal S170.Combiner 492 is configured to extend signal S160 according to harmonic wave weighting factor S180 weighting harmonic wave, according to noise weighting factor S190 weighting, through zoop signal S170, and exports high band excitation signal S120 as the summation that is weighted signal.In this example, combiner 492 comprises weighting factor counter 550, and it is configured to calculate harmonic wave weighting factor S180 and noise weighting factor S190.

Weighting factor counter 550 can be configured to calculate weighting factor S180 and S190 according to the required ratio of harmonic content and noise content in high band excitation signal S120.For instance, combiner 492 may need to produce high band excitation signal S120 with have harmonic energy to high-frequency band signals S30 to noise energy than similar harmonic energy and noise energy ratio.In some embodiments of weighting factor counter 550, one or more parameters relevant according to the periodicity with narrow-band signal S20 or narrow-band residual signal (for example, pitch gain and/or speech pattern) are calculated weighting factor S180, S190.This embodiment of weighting factor counter 550 can be configured to the value of assigning (for example) to be directly proportional to pitch gain to harmonic wave weighting factor S180, and/or is compared to turbid voice signal to the higher value of noise weighting factor S190 appointment for clear voice signal.

In other embodiments, weighting factor counter 550 is configured to calculate according to the cyclic indicator of high-frequency band signals S30 the value of harmonic wave weighting factor S180 and/or noise weighting factor S190.In this type of example, weighting factor counter 550 is calculated as harmonic wave weighting factor S180 the maximal value of the present frame of high-frequency band signals S30 or the coefficient of autocorrelation of subframe, wherein postpones and do not comprise in the hunting zone of zero sample delay to carry out auto-correlation comprising pitch lag.Figure 17 show take a pitch lag postpone centered by and there is the example that the length of the width that is not more than a pitch lag is this hunting zone of n sample.

Figure 17 also shows that weighting factor counter 550 is at the example of the other method of the cyclic indicator of some stages calculating high-frequency band signals S30.In the first stage, present frame is divided into many subframes, and identifies separately the delay of coefficient of autocorrelation when maximum for each subframe.As mentioned above, postpone and do not comprise in the hunting zone of zero sample delay to carry out auto-correlation comprising pitch lag.

In subordinate phase, by the delay of corresponding identification being applied to the subframe of each subframe, connection gained, with the related coefficient of setting up optimal delay frame and harmonic wave weighting factor S180 is calculated as between primitive frame and optimal delay frame, set up deferred frame.In another alternate embodiment, weighting factor counter 550 is calculated as harmonic wave weighting factor S180 the mean value of the maximum coefficient of autocorrelation obtaining for each subframe in the first stage.The embodiment of weighting factor counter 550 also can be configured to convergent-divergent related coefficient, and/or itself and another value is combined, to calculate the value of harmonic wave weighting factor S180.

Weighting factor counter 550 may need only otherwise indicating frame to have the cyclic indicator that calculates high-frequency band signals S30 in periodic situation.For instance, weighting factor counter 550 can be configured to for example, calculate according to the relation between periodic another designator (, pitch gain) of present frame and threshold value the cyclic indicator of high-frequency band signals S30.In an example, the value that weighting factor counter 550 is only configured to the pitch gain (for example, the self-adaptation code book gain that narrow-band is residual) when frame is greater than 0.5 when (or, at least 0.5) and just high-frequency band signals S30 is carried out to auto-correlation computation.In another example, weighting factor counter 550 is configured to only for the frame (for example,, only for voiced sound signal) with special sound mode state, high-frequency band signals S30 be carried out to auto-correlation computation.In such cases, weighting factor counter 550 can be configured to assign default weighting factor for the frame with other speech pattern state and/or less pitch gain value.

Embodiment comprises and is configured to according to being different from periodically or characteristic except periodicity is calculated the other embodiments of the weighting factor counter 550 of weighting factor.For instance, this embodiment can be configured to for the voice signal with large pitch lag than assigning larger value for the voice signal with little pitch lag to noise gain factor S190.This type of embodiment of another of weighting factor counter 550 is configured to the index with respect to the signal energy in other frequency component place according to the signal energy at the multiple place in fundamental frequency, determines the index of the humorous degree of broadband voice signal S10 or high-frequency band signals S30.

Some embodiments of wideband speech coding device A100 are configured to another index based on pitch gain described herein and/or periodicity or humorous degree, export the indication (for example, indication frame is harmonic wave or non-harmonic 1 flag) of periodicity or humorous degree.In an example, corresponding broadband Voice decoder B100 indicates to configure with this operation that for example weighting factor is calculated.In another example, this indication carrys out the value of computing voice mode parameter for scrambler and/or demoder place.

May need high band excitation generator A302 to produce high band excitation signal S120, make the energy of pumping signal roughly not be subject to the impact of the particular value of weighting factor S180 and S190.In the case, weighting factor counter 550 can be configured to calculate the value (or receiving this value from another element of memory storage or high band encoder A200) of harmonic wave weighting factor S180 or noise weighting factor S190, and according to for example following formula, derives the value of another weighting factor:

(W _{harmonic wave}) ²+ (W _noise) ²=1, (2)

W wherein _{harmonic wave}represent harmonic wave weighting factor S180, and W _noiserepresent noise weighting factor S190.Or, weighting factor counter 550 can be configured to select from multipair weighting factor S180, S190 according to the value of the periodic measurement of present frame or subframe corresponding one, wherein said to calculate to meet for example constant energy ratio of expression formula (2) through in advance.For the embodiment of following the weighting factor counter 550 of expression formula (2), the representative value of harmonic wave weighting factor S180 is in approximately 0.7 to approximately 1.0 scopes, and the representative value of noise weighting factor S190 is in approximately 0.1 to approximately 0.7 scope.Other embodiment of weighting factor counter 550 can be configured to operate according to expression formula (2) pattern that adds weight update according to harmonic wave extension signal S160 and the required baseline between zoop signal S170.

When using sparse code book (entry major part is the code book of null value) to calculate residual quantization means form, in synthetic speech signal, may there is illusion.Especially when with low bitrate coding narrow-band signal, code book can be there is sparse.The sparse illusion causing of code book is quasi periodic conventionally in time, and mainly more than 3kHz, occurs.Because thering is good time resolution for people's ear under upper frequency, so these illusions may be more obvious in high frequency band.

The embodiment that embodiment comprises the high band excitation generator A300 that is configured to carry out anti-sparseness filtering.Figure 18 shows the block scheme of the embodiment A312 of high band excitation generator A302, and described embodiment A312 comprises anti-sparseness filtering device 600, and what it was configured to quantizer 450 to produce carries out filtering through de-quantization narrow-band pumping signal.Figure 19 shows the block scheme of the embodiment A314 of high band excitation generator A302, and described embodiment A314 comprises anti-sparseness filtering device 600, and its frequency spectrum that is configured to that frequency spectrum stretcher A400 is produced extends signal and carries out filtering.Figure 20 shows the block scheme of the embodiment A316 of high band excitation generator A302, and described embodiment A316 comprises anti-sparseness filtering device 600, and it is configured to the output of combiner 490 to carry out filtering to produce high band excitation signal S120.Certainly, expection and this disclose clearly by embodiment A304 and A306 any one feature and any one the embodiment of high band excitation generator A300 of Feature Combination of embodiment A312, A314 and A316.Anti-sparseness filtering device 600 is also configurable in frequency spectrum stretcher A400: for example the element in frequency spectrum stretcher A402 510,520,530 and 540 any one after.Pay particular attention to, anti-sparseness filtering device 600 also can be used for the embodiment that execution spectrum folding, frequency spectrum are translated or harmonic wave extends of frequency spectrum stretcher A400.

Anti-sparseness filtering device 600 can be configured to change the phase place of its input signal.For instance, anti-sparseness filtering device 600 may need to be configured and to arrange, and makes the phase place of high band excitation signal S120 along with Timing randomization or alternate manner distribute more equably.The response that may also need anti-sparseness filtering device 600 is frequency spectrum leveling, makes not have sizable change through the amplitude frequency spectrum of filtering signal.In an example, anti-sparseness filtering device 600 is embodied as the all-pass filter with transfer function according to following formula:

$H\left(z\right)=\frac{-0.7+z^{-4}}{1-{0.7z}^{-4}}\·\frac{0.6+z^{-6}}{1+{0.6z}^{-6}}.---\left(3\right)$

An effect of this wave filter can be that the energy dissipation of input signal is opened it is no longer only concentrated in several samples.

The sparse illusion causing of code book is conventionally more obvious for the signal of the residual similar noise that comprises less tone information wherein, and also more obvious for the voice in ground unrest.Sparsely in the situation that excitation has long-term structure, conventionally cause less illusion, and in fact phase modification can cause the noise in voiced sound signal.Therefore, may need to configure anti-sparseness filtering device 600 voiceless sound signal carried out to filtering and to make at least some voiced sound signals pass through in the situation that not making change.Voiceless sound signal (is for example characterised in that low pitch gain, quantize the gain of narrow-band self-adaptation code book) and approach zero or be positive spectral tilt (for example, quantize the first reflection coefficient), thus indication leveling or the acclivitous spectrum envelope along with the continuous increase of frequency.The typical embodiments of anti-sparseness filtering device 600 to voiceless sound (is for example configured to, as indicated in the value of spectral tilt) carry out filtering, when pitch gain is carried out filtering to voiced sound during lower than threshold value (or, be not more than threshold value), and otherwise in the situation that not making change, signal is passed through.

The other embodiments of anti-sparseness filtering device 600 comprise two or more wave filters, and it is configured to have different maximum phases and revises angle (for example,, up to 180 degree).In the case, anti-sparseness filtering device 600 according to pitch gain (for example can be configured to, quantize self-adaptation code book or LTP gain) value at these, form in wave filters and select, to larger maximum phase is revised to angle for having compared with the frame of low pitch yield value.The embodiment of anti-sparseness filtering device 600 also can comprise different composition wave filters, it is configured to revise phase place in the part more or less of frequency spectrum, so as by the wave filter that is configured to revise phase place in the wider frequency range of input signal for having compared with the frame of low pitch yield value.

In order to copy exactly encoded voice signal, may need to make the ratio between the high frequency band of synthetic broadband voice signal S100 and the level of narrow-band part to be similar to the described ratio in original broadband voice signal S10.Except the spectrum envelope that high frequency band coding parameter S60a represents, high band encoder A200 also can be configured to characterize high-frequency band signals S30 by fixed time or gain envelope.As shown in figure 10, high band encoder A202 comprises high frequency band gain factor counter A230, it is configured and arranges for example, to calculate one or more gain factors according to the relation between high-frequency band signals S30 and synthetic high-frequency band signals S130 (, difference or the ratio between the energy of described two signals in frame or its certain part).In other embodiment of high band encoder A202, high frequency band gain calculator A230 can similarly configure but change into through arranging to carry out calculated gains envelope according to this time-varying relationship between high-frequency band signals S30 and narrow-band pumping signal S80 or high band excitation signal S120.

The temporal envelope of narrow-band pumping signal S80 and high-frequency band signals S30 is probably similar.Therefore, the gain envelope of for example, relation based between high-frequency band signals S30 and narrow-band pumping signal S80 (or the signal of therefrom deriving, high band excitation signal S120 or synthetic high-frequency band signals S130) of coding is conventionally by than coding, only the gain envelope based on high-frequency band signals S30 is effective.In typical embodiments, high band encoder A202 is configured to be output as the quantization index of 8 to 12 that each frame is specified 5 gain factors.

High frequency band gain factor counter A230 can be configured to carry out as the task of comprising one or more serial subtasks gain factor and calculate.Figure 21 shows the process flow diagram of example T200 of task that calculates the yield value of corresponding subframe according to the relative energy of high-frequency band signals S30 and synthetic high-frequency band signals S130.Task 220a and 220b calculate the energy of the corresponding subframe of each signal.For instance, task 220a and 220b can be configured to by energy be calculated as each subframe sample square and.Task T230 is calculated as the gain factor of subframe the square root of the ratio of those energy.In this example, task T230 is calculated as gain factor the square root of the ratio of the energy of high-frequency band signals S30 in subframe and the energy of synthetic high-frequency band signals S130.

High frequency band gain factor counter A230 may need to be configured to calculate subframe energy according to window function.Figure 22 shows the process flow diagram of this embodiment T210 of gain factor calculation task T200.Task T215a is applied to high-frequency band signals S30 by window function, and task T215b by uniform window function application in synthetic high-frequency band signals S130.The embodiment 222a of task 220a and 220b and 222b calculate the energy of window separately, and task T230 is calculated as the gain factor of subframe the square root of the ratio of energy.

May need the overlapping window function of application and adjacent sub-frames.For instance, can be overlapping-window function of the generation gain factor of phase add mode application can help to reduce or avoid the uncontinuity between subframe.In an example, high frequency band gain factor counter A230 is configured to the trapezoidal window function of application as shown in Figure 23 a, wherein each overlapping one millisecond of window and two adjacent sub-frames.Figure 23 b shows each of five subframes that this window function is applied to 20 milliseconds of frames.Other embodiment of high frequency band gain factor counter A230 can be configured to the window function that application has negative lap cycle not and/or different windows shape (for example, rectangle, Hamming) (it can be symmetrical or asymmetric).The frame that the embodiment of high frequency band gain factor counter A230 also may be configured to by different windows function application different subframes in frame and/or comprise the subframe with different length.

The example of following value (there is no restriction) as particular is provided.For the frame of one 20 milliseconds of these situation supposition, but can use any other duration.For the high-frequency band signals with 7kHz sampling, each frame has 140 samples.If this frame is divided into five subframes with equal length, each subframe will have 28 samples so, and the window as shown in Figure 23 a will be wide for 42 samples.For the high-frequency band signals with 8kHz sampling, each frame has 160 samples.If this frame is divided into five subframes with equal length, each subframe will have 32 samples so, and the window as shown in Figure 23 a will be wide for 48 samples.In other embodiments, can use the subframe with any width, and even may make the embodiment of high frequency band gain calculator A230 be configured to for the different gain factor of each Sample producing of frame.

Figure 24 shows the block scheme of the embodiment B202 of high band decoder B200.High band decoder B202 comprises high band excitation generator B300, and it is configured to produce high band excitation signal S120 based on narrow-band pumping signal S80.Depending on particular system design, select, can implement high band excitation generator B300 according to any one of the embodiment of high band excitation generator A300 described herein.Conventionally, high band excitation generator B300 need to be embodied as and there is the response identical with the high band excitation generator of the high band encoder of specific coding system.Yet, because narrow-band demoder B110 will carry out the de-quantization of encoded narrow-band pumping signal S50 conventionally, so in most of the cases, high band excitation generator B300 can be through implementing to receive narrow-band pumping signal S80 from narrow-band demoder B110, and do not need to comprise the quantizer being configured to encoded narrow-band pumping signal S50 de-quantization.Narrow-band demoder B110 also may be through implementing to comprise the example of anti-sparseness filtering device 600, and it is configured to, before the narrow-band pumping signal through de-quantization being input to the narrow-band composite filter of wave filter 330 for example, described signal is carried out to filtering.

Quantizer 560 is configured to the de-quantization to high band filter parameter S 60a (being one group of LSF in this example), and LSF-LP filter coefficient conversion 570 is configured to LSF to be transformed to one group of filter coefficient (for example,, as the quantizer 240 above with reference to narrowband encoder A122 and conversion 250 are described).In other embodiments, as mentioned above, can use different coefficient sets (for example, cepstrum coefficient) and/or coefficient representation (for example, ISP).High frequency band composite filter B204 is configured to produce synthetic high-frequency band signals according to high band excitation signal S120 and described group of filter coefficient.For system that wherein high band encoder comprises composite filter (for example, in the example of above-mentioned scrambler A202), may need high frequency band composite filter B204 to be embodied as and to there is the response identical with described composite filter (for example, identical transfer function).

High band decoder B202 also comprises the quantizer 580 being configured to high frequency band gain factor S60b de-quantization, with be configured and arrange for example the gain factor through de-quantization to be applied to synthetic high-frequency band signals, to produce the gain control element 590 (, multiplier or amplifier) of high-frequency band signals S100.Situation for the gain envelope of frame wherein by an above gain factor appointment, gain control element 590 can comprise and is configured to gain factor to be applied to according to window function to the logic of each subframe, described window function can be for example, with the window function that the gain calculator (, high frequency band gain calculator A230) by corresponding high band encoder is applied identical or different.In other embodiment of high band decoder B202, gain control element 590 is through configuring similarly but the gain factor through de-quantization being applied to narrow-band pumping signal S80 or being applied to high band excitation signal S120 to change into through arrangement.

As mentioned above, may in high band encoder and high band decoder, obtain equal state (for example,, by using through de-quantization value during encoding).Therefore, may be according to guaranteeing in the coded system of this embodiment that the corresponding noise generator in high band excitation generator A300 and B300 has equal state.For instance, the high band excitation generator A300 of this embodiment and B300 can be configured and make the state of noise generator be the information of having encoded in same number of frames (for example, narrow band filter parameter S 40 or its part, and/or encoded narrow-band pumping signal S50 or its part) qualitative function really.

One or more (for example, quantizer 230,420 or 430) in the quantizer of element described herein can be configured to carry out class vector and quantize.For instance, the information that this quantizer can be configured to encode in the same number of frames based in narrow-band channel and/or in high frequency band channel is selected a code book from one group of code book.This technology conventionally be take and is stored extra code book and provide the code efficiency of increase as cost.

As above discussed referring to for example Fig. 8 and 9, remove rough spectrum envelope from narrow-band voice signal S20 after, quite a large amount of periodic structures may be retained in residual signal.For instance, residual signal can contain rough recurrent pulses or spiking sequence in time.This structure (conventionally relevant with tone) especially likely occurs in voiced speech signal.The calculating of the quantization means form of narrow-band residual signal can comprise according to the model of the long term periodicities that for example, represented by () one or more code books this tone structure of encoding.

The tone structure of actual residual signals may not mated with periodic model completely.For instance, residual signal may comprise the reduced jitter of the location rule of tone pulses, makes in frame the distance between continuous tone pulse not exclusively equate and described structure suitable rule not.These scramblings tend to reduce code efficiency.

Some embodiments of narrowband encoder A120 be configured to by before quantizing or during auto-adaptive time deviation is applied to residual, or by otherwise comprise auto-adaptive time deviation in encoded pumping signal, carry out the regularization of tone structure.For instance, this scrambler can be configured to select or otherwise computing time deviation degree (for example, according to one or more perceptual weightings and/or error minimize standard), the pumping signal of gained and model the best of long term periodicities are fitted.The regularization of tone structure is carried out by the celp coder subgroup that is called lax code exciting lnear predict (RCELP) scrambler.

RCELP scrambler is configured to execution time deviation conventionally as self-adaptation time shift.This time shift can be negative several milliseconds to the just delay of several milliseconds of scopes, and it changes to avoid the uncontinuity that can hear conventionally smoothly.In some embodiments, this scrambler is configured to segmented mode application rule, wherein each frame or subframe deviation fixedly time shift accordingly.In other embodiments, scrambler is configured to application rule and is turned to continuous offset difference function, makes frame or subframe according to tone contour (also referred to as tone track) and deviation.In some cases (for example, described in No. 2004/0098255 U.S. Patent Application Publication case), scrambler be configured to by by offset applications in comprising time deviation for calculating the perceptual weighting input signal of encoded pumping signal in encoded pumping signal.

The encoded pumping signal of scrambler computation rule and quantification, and demoder to encoded pumping signal de-quantization to obtain for the synthesis of the pumping signal through decodeing speech signal.Therefore through decoded output signal, show the delay of the variation identical be included in delay in encoded pumping signal by regularization.Conventionally, not by the communication of any specified rule amount to demoder.

Regularization often makes residual signal more easily encode, and this has improved from the coding gain of long-term predictor and has therefore advanced overall code efficiency, and conventionally can not produce illusion.May need only to unvoiced frame executing rule.For instance, narrowband encoder A124 can be configured to only be offset frame or the subframe that those have long-term structure (for example, voiced sound signal).Even may need only to the subframe executing rule that comprises tone pulses energy.The 5th, the various embodiments of RCELP coding have been described in 704, No. 003 United States Patent (USP)s people such as () Kleijn and the 6th, 879, No. 955 United States Patent (USP)s (Rao) and No. 2004/0098255 U.S. Patent Application Publication case people such as () Kovesi.The existing embodiment of RCELP scrambler comprises as the enhancing variable bit rate coder (EVRC) described in the IS-127 of telecommunications industry association (TIA), and third generation partnership relation plan 2 (3GPP2) alternative mode vocoder (SMV).

Unfortunately, regularization is for wherein for example, causing some problems from the wideband speech coding device (system that, comprises wideband speech coding device A100 and broadband Voice decoder B100) of encoded narrow-band pumping signal derivation high band excitation.Because high band excitation signal derives from time deviation signal, so high band excitation signal will have the time response different from the time response of original high frequency band voice signal conventionally.In other words, high band excitation signal will no longer be synchronizeed with original high frequency band voice signal.

Temporal misalignment between deviation high band excitation signal and original high frequency band voice signal may cause some problems.For instance, deviation high band excitation signal may be no longer for providing suitable source forcing according to the composite filter of the filter parameter configuration of extracting from original high frequency band voice signal.Therefore, synthetic high-frequency band signals can contain the illusion of hearing reducing through the perceived quality of decoding broadband voice signal.

Temporal misalignment also may cause the poor efficiency of gain envelope coding.As mentioned above, between the temporal envelope of narrow-band pumping signal S80 and high-frequency band signals S30, probably exist relevant.By according to the gain envelope of the relation coding high-frequency band signals between these two temporal envelope, compare the raising that can realize code efficiency with direct coding gain envelope.Yet, when encoded narrow-band pumping signal regularization, this relevant may weakening.Temporal misalignment between narrow-band pumping signal S80 and high-frequency band signals S30 can cause occurring fluctuating in high frequency band gain factor S60b, and code efficiency may reduce.

Embodiment comprises the wideband speech coding method to high frequency band voice signal execution time deviation according to the time deviation comprising in corresponding encoded narrow-band pumping signal.The potential advantage of these class methods comprises to be improved through the quality of decoding broadband voice signal and/or the efficiency of improvement coding high frequency band gain envelope.

Figure 25 shows the block scheme of the embodiment AD10 of wideband speech coding device A100.The embodiment A124 that scrambler AD10 comprises narrowband encoder A120, described embodiment A124 is configured to executing rule during calculating encoded narrow-band pumping signal S50.For instance, narrowband encoder A124 can be according to the one or more configurations in RCELP embodiment discussed above.

Narrowband encoder A124 is also configured to the regularization data-signal SD10 that the degree of applied time deviation is specified in output.For narrowband encoder A124, be configured to fixing time shift to be applied to the various situations of each frame or subframe, regularization data-signal SD10 can comprise a series of values, and it be take sample, millisecond or a certain increment At All Other Times and each time shift amount is designated as to integer or non integer value as unit.The time scale that is configured to otherwise to revise frame or other sample sequence for narrowband encoder A124 (for example, by compressing a part and expanding another part) situation, regularization information signal SD10 can comprise the corresponding description to revising, for example one group of function parameter.In a particular instance, narrowband encoder A124 is configured to the fixedly time shift that frame is divided into three subframes and calculates each subframe, makes regularization data-signal SD10 indicate three time shift amounts of each regularization frame of encoded narrow-band signal.

Wideband speech coding device AD10 comprises lag line D120, and it is configured to the some parts that advance or block high frequency band voice signal S30 according to the retardation of being indicated by input signal, thus generation time deviation high frequency band voice signal S30a.In the example shown in Figure 25, lag line D120 is configured to according to carrying out the execution time deviation to high frequency band voice signal S30 by the deviation of regularization data-signal SD10 indication.In this way, the same time departure comprising in encoded narrow-band pumping signal S50 was also applied to the appropriate section of high frequency band voice signal S30 before analyzing.Although this example is shown as lag line D120 the individual component of high band encoder A200, in other embodiments, lag line D120 is configured to a part for high band encoder.

The spectrum analysis that the other embodiments of high band encoder A200 can be configured to carry out deviation high frequency band voice signal S30 not (for example, and before calculating high frequency band gain parameter S60b, carry out the time deviation of high frequency band voice signal S30 lpc analysis).This scrambler can be including (for example) the embodiment that is configured to the lag line D120 of execution time deviation.Yet, in such cases, the spectrum envelope based on describing the high band filter parameter S 60a of the not analysis of deviation signal S30 with high band excitation signal S120 misalignment in time.

Lag line D120 can be according to being suitable for that required time biased operation is applied to the logic element of high frequency band voice signal S30 and any combination of memory element configures.For instance, lag line D120 can be configured to from impact damper, read high frequency band voice signal S30 according to required time shift.Figure 26 a shows the schematic diagram of this embodiment D122 of lag line D120, and described lag line D120 comprises shift register SR1.Shift register SR1 is the impact damper with about length m that is configured to m the most recent sample of reception and storage high frequency band voice signal S30.Value m at least equals just (or " propelling ") and the summation of bearing (or " retardance ") time shift of the maximum of supporting.Value m equals the frame of high-frequency band signals S30 or the length of subframe may be more convenient.

Lag line D122 is configured to the deviation post OL output time deviation high-frequency band signals S30a from shift register SR1.The location of deviation post OL changes near reference position (zero time shift) according to the current time shift by for example regularization data-signal SD10 indication.Lag line D122 can be configured to support equal propelling and retardance restriction, or a restriction is greater than another restriction, and making can be in one direction than carry out larger skew on other direction.Figure 26 a shows that the positive time shift of supporting is greater than the particular instance of negative time shift.Lag line D122 can be configured to once export one or more samples (for example, depending on output bus width).

The regularization time shift with the value that is greater than several milliseconds can cause the illusion of hearing in decoded signal.Conventionally, the value of the regularization time shift of being carried out by narrowband encoder A124 will be no more than several milliseconds, make the time shift of being indicated by regularization data-signal SD10 by limited.Yet, may need in such cases lag line D122 to be configured to align and/or negative direction on time shift force maximum constraints (for example,, to follow the more strict restriction of restriction of forcing than narrowband encoder).

Figure 26 b shows the schematic diagram of the embodiment D124 of lag line D122, and lag line D122 comprises displacement window SW.In this example, the location of the deviation post OL window SW that is shifted limits.Although Figure 26 b shows buffer length m and be greater than the situation of the width of displacement window SW, lag line D124 also can equal m through the width of the window SW that implements to make to be shifted.

In other embodiments, lag line D120 is configured to, according to required time shift, high frequency band voice signal S30 is written to impact damper.Figure 27 shows the schematic diagram of the embodiment D130 of lag line D120, and described embodiment D130 comprises and is configured to receive and two shift register SR2 and the SR3 of storage high frequency band voice signal S30.Lag line D130 is configured to according to for example by the time shift of regularization data-signal SD10 indication, the frame from shift register SR2 or subframe being written to shift register SR3.Shift register SR3 is configured to fifo buffer, and it is configured to output time deviation high-frequency band signals S30a.

In the particular instance shown in Figure 27, shift register SR2 comprises frame buffer part FB1 and delay buffer part DB, and shift register SR3 comprises frame buffer part FB2, advances bumper portion AB and retardance bumper portion RB.Advance the length of impact damper AB and retardance impact damper RB to equate, or wherein one can be greater than another one, makes the skew in a supported direction be greater than the skew on supported other direction.Delay buffer DB and retardance bumper portion RB can be configured to have equal length.Or, the comparable retardance impact damper of delay buffer DB RB is short transfers to the shift register SR3 required time interval by sample from frame buffer FB1 to consider, described transfer can comprise other processing operation that for example first made sample bias before storing shift register SR3 into.

In the example of Figure 27, frame buffer FB1 is configured to have the length equal in length with a frame of high-frequency band signals S30.In another example, frame buffer FB1 is configured to have the length equal in length with a subframe of high-frequency band signals S30.In the case, lag line D130 for example can be configured to comprise, for identical (, average) delay is applied to the logic of all subframes of frame to be offset.Lag line D130 also can comprise the logic for the value to be rewritten with blocking impact damper RB or propelling impact damper AB of the value from frame buffer FB1 averaged.In another example, shift register SR3 can be configured to only via frame buffer FB1, receive the value of high-frequency band signals S30, and in the case, lag line D130 can comprise for being written to the logic of carrying out interpolation on the successive frame of shift register SR3 or the gap between subframe.In other embodiments, lag line D130 can be configured to before the sample from frame buffer FB1 is written to shift register SR3 described sample implementation deviation operation (for example,, according to the function of being described by regularization data-signal SD10).

Lag line D120 may need apply based on but be not equal to the time deviation by the deviation of regularization data-signal SD10 appointment.Figure 28 shows the block scheme of the embodiment AD12 of wideband speech coding device AD10, and wideband speech coding device AD10 comprises length of delay mapper D110.It is through mapping length of delay SD10a that length of delay mapper D110 is configured to the Preference-Deviation Mapping by regularization data-signal SD10 indication.Lag line D120 is configured to according to carrying out generation time deviation high frequency band voice signal S30a by the deviation of indicating through mapping length of delay SD10a.

Can expect that the time shift of narrowband encoder application makes progress in time smoothly.Therefore, conventionally computing voice is applied to the average narrow-band time shift of subframe image duration and is offset the respective frame of high frequency band voice signal S30 according to this mean value enough.In this type of example, length of delay mapper D110 is configured to calculate the mean value of the subframe delay value of each frame, and lag line D120 is configured to the mean value calculating to be applied to the respective frame of high-frequency band signals S30.In other example, for example can calculate and apply, for example, compared with the mean value in short period (, two subframes, or half frame) or longer cycle (, two frames).In the situation that mean value is the non integer value of sample, length of delay mapper D110 can be configured to before described value is outputed to lag line D120, described value is rounded to the integer number of sample.

The regularization time shift of the sample that narrowband encoder A124 can be configured to comprise non-integer number in encoded narrow-band pumping signal.In the case, length of delay mapper D110 may need to be configured to narrow-band time shift to be rounded to the integer number of sample, and lag line D120 may need the time shift through rounding up to be applied to high frequency band voice signal S30.

In some embodiments of wideband speech coding device AD10, narrow-band voice signal S20 may be different from the sampling rate of high frequency band voice signal S30.In such cases, length of delay mapper D110 can be configured to regulate the time shift amount of indicating in regularization data-signal SD10, to consider the difference between narrow-band voice signal S20 (or narrow-band pumping signal S80) and the sampling rate of high frequency band voice signal S30.For instance, length of delay mapper D110 can be configured to the ratio convergent-divergent time shift amount according to sampling rate.In a particular instance mentioned above, with 8kHz, narrow-band voice signal S20 is sampled, and with 7kHz, high frequency band voice signal S30 is sampled.In the case, length of delay mapper D110 is configured to each side-play amount to be multiplied by 7/8.The embodiment of length of delay mapper D110 also can be configured to carry out this convergent-divergent computing and integer described herein rounds up and/or time shift average calculating operation.

In other embodiments, lag line D120 is configured to otherwise revise the time scale (for example,, by compressing a part and expanding another part) of frame or other sample sequence.For instance, narrowband encoder A124 can be configured to carry out executing rule according to the function of for example tone contour or track.In the case, regularization data-signal SD10 can comprise the corresponding description (for example one group of parameter) to described function, and lag line D120 can comprise and is configured to the logic to the frame of high frequency band voice signal S30 or subframe implementation deviation according to described function.In other embodiments, length of delay mapper D110 be configured to before described function is applied to high frequency band voice signal S30 by lag line D120 to described function average, convergent-divergent and/or round up.For instance, length of delay mapper D110 can be configured to calculate one or more length of delays according to described function, each length of delay indication number of samples, its then by lag line D120 application with to one or more respective frame of high frequency band voice signal S30 or subframe execution time deviation.

Figure 29 shows according to the time deviation comprising in corresponding encoded narrow-band pumping signal the process flow diagram to the method MD100 of high frequency band voice signal execution time deviation.Task TD100 processes broadband voice signal to obtain narrow-band voice signal and high frequency band voice signal.For instance, task TD100 can be configured to use the bank of filters (for example embodiment of bank of filters A110) with low-pass filter and Hi-pass filter to carry out filtering to broadband voice signal.Task TD200 is at least encoded to encoded narrow-band pumping signal and a plurality of narrow band filter parameter by narrow-band voice signal.Described encoded narrow-band pumping signal and/or filter parameter can quantize, and encoded narrow-band pumping signal also can comprise for example other parameter of speech pattern parameter.Task TD200 also comprises the time deviation in encoded narrow-band pumping signal.

Task TD300 produces high band excitation signal based on narrow-band pumping signal.In the case, the narrow-band pumping signal of narrow-band pumping signal based on encoded.Task TD400, according at least described high band excitation signal, is at least encoded to a plurality of high band filter parameters by high frequency band voice signal.For instance, can be configured to high frequency band speech signal coding be a plurality of quantification LSF to task TD400.Task TD500 is applied to high frequency band voice signal by time shift, the information that the time deviation of described time shift based on comprising in encoded narrow-band pumping signal is relevant.

Task TD400 can be configured to high frequency band voice signal to carry out spectrum analysis (for example, lpc analysis), and/or calculates the gain envelope of high frequency band voice signal.In such cases, task TD500 can be configured to, before described analysis and/or the calculating of gain envelope, time shift is applied to high frequency band voice signal.

Other embodiment of wideband speech coding device A100 is configured to the time deviation of high band excitation signal S120 that the time deviation comprising in encoded narrow-band pumping signal is caused and reverses.For instance, high band excitation generator A300 can be through implementing to comprise the embodiment of lag line D120, the described embodiment of lag line D120 is configured to receive regularization data-signal SD10 or through mapping length of delay SD10a, and the time shift of reversing is accordingly applied to narrow-band pumping signal S80, and/or be applied to the signal subsequently (for example, harmonic wave extends signal S160 or high band excitation signal S120) based on described narrow-band pumping signal S80.

Other wideband speech coding device embodiment can be configured to independently of one another narrow-band voice signal S20 and high frequency band voice signal S30 be encoded, and makes high frequency band voice signal S30 be encoded as the representation of high frequency band spectrum envelope and high band excitation signal.This embodiment can be configured to the basis information relevant with the time deviation comprising in encoded narrow-band pumping signal to high frequency band residual signal execution time deviation, or otherwise time deviation is included in encoded high band excitation signal.For instance, high band encoder can comprise the embodiment that is configured to time deviation to be applied to lag line D120 and/or the length of delay mapper D110 of high frequency band residual signal described herein.The potential advantage of this operation comprise to high frequency band residual signal compared with efficient coding, and better coupling between synthesis of narrow frequency band and high frequency band voice signal.

As mentioned above, embodiment described herein comprises the embodiment that can be used for carrying out the compatible of embedded encoded, support and narrow band system and avoid needing code conversion.The support of high frequency band coding also be can be used on cost distinguishing and there is chip, chipset, device and/or the network of broadband support and backward compatibility and only there is chip, chipset, device and/or the network that narrow-band is supported.As described herein to the support of high frequency band coding also can with for supporting the technology of low-frequency band coding be combined with, and according to the system of this embodiment, method or equipment can support to for example approximately 50 or 100Hz until approximately 7 or the coding of the frequency component of 8kHz.

As mentioned above, to speech coder, add high frequency band support and can improve sharpness, especially aspect fricative differentiation.Although this differentiation may be derived according to specific context by human listener conventionally, high frequency band support can for example, be served as and enable feature in speech recognition and other machine decipher application (system of processing for automated voice menu navigation and/or automatic call).

According to the equipment of an embodiment, for example can be embedded into, in portable radio communication device (, cellular phone or personal digital assistant (PDA)).Or this equipment can be included in another communicator, for example VoIP mobile phone, be configured to support the personal computer of VoIP communication or be configured to routing telephone or the network equipment of VoIP communication.For instance, according to the equipment of an embodiment, may be implemented in the chip or chipset of communicator.Depending on application-specific, this device also can comprise for example following characteristics: to the conversion of the analog to digital of voice signal and/or digital-to-analog, amplify and/or the circuit of other signal processing operations for voice signal is carried out, and/or for launching and/or receive the radio circuit of encoded voice signal.

Clearly expection and announcement embodiment can comprise the 60/667th, No. 901 and the 60/673rd, No. 965 U.S. Provisional Patent Application cases, US publication are any one in the further feature disclosing in 2006/0282263,2007/0088558,2007/0088541,2006/0277042,2007/0088542,2006/0277038,2006/0271356 and 2008/0126086 patented claim or one is above and/or therewith use, and advocate the rights and interests of described temporary patent application case in the application's case.This category feature comprises removal and betides the non-existent high-energy burst with the short duration substantially in high frequency band and in narrow-band.This category feature comprises for example fixing or adaptive smooth of the coefficient representation of high frequency band LSF.Fixing or the self-adaptation that this category feature comprises the noise being associated with the quantification of the coefficient representation of for example LSF is shaped.This category feature also comprises the fixing or adaptive smooth of gain envelope, and the adaptive attenuation of gain envelope.

Provide the above introduction of described embodiment to make those skilled in the art can manufacture or use the present invention.May make various modifications to these embodiment, and General Principle provided herein also can be applicable to other embodiment.For instance, embodiment can be embodied as to a part or whole part hard-wired circuit, be embodied as the Circnit Layout being fabricated onto in special IC, or be embodied as machine readable code and be loaded into the firmware program in Nonvolatile memory devices or load or be loaded into the software program data storage medium from data storage medium, described code is the instruction that can be carried out by the array of logic elements of for example microprocessor or other digital signal processing unit.Data storage medium can be memory element array, semiconductor memory (it can comprise (being not limited to) dynamically or static RAM (SRAM) (random access memory), ROM (ROM (read-only memory)) and/or quick flashing RAM) for example, or ferroelectric, magnetic resistance, two-way switch semiconductor, polymkeric substance or phase transition storage; Or the disc type media of disk or CD for example.Any one or above instruction group or sequence that term " software " is interpreted as comprising source code, assembly language code, machine code, binary code, firmware, macrocode, microcode, can be carried out by array of logic elements, and any combination of this type of example.

The various elements of the embodiment of high band excitation generator A300 and B300, high band encoder A100, high band decoder B200, wideband speech coding device A200 and broadband Voice decoder B100 can be embodied as (for example) and reside on electronics and/or the optical devices on the same chip in chipset or between two or more chips, but also expection exists other configuration that there is no this restriction.One or more elements of this equipment can be embodied as one or more instruction groups in whole or in part, described instruction group be configured to one or more fix or programmable logic element (for example, transistor, door) to carry out on array, described element is for example microprocessor, embedded processor, the IP kernel heart, digital signal processor, FPGA (field programmable gate array), ASSP (Application Specific Standard Product) and ASIC (special IC).One or more these class components also (for example may have common structure, for carrying out processor corresponding to the code section of different elements at different time, through carrying out to carry out the instruction group corresponding to the task of different elements at different time, or the configuration for electronics and/or the optical devices of different elements executable operations at different time).In addition, one or more these class components may be for carrying out the task not directly related with the operation of equipment or other instruction group, for example to be embedded with the device of described equipment or system another operate relevant task.

Figure 30 shows according to an embodiment having narrow-band partly and the process flow diagram of the method M100 that the described highband part of the voice signal of highband part is encoded.Task X100 calculates one group of filter parameter of the feature of the spectrum envelope that represents highband part.Task X200 calculates frequency spectrum by the signal that nonlinear function is applied to partly derive from narrow-band and extends signal.Task X300 basis (A) described group of filter parameter and (B) high band excitation signal based on described frequency spectrum extension signal produce synthetic high-frequency band signals.Relation between the energy of task X400 based on (C) highband part and the energy of the signal of (D) partly deriving from narrow-band is carried out calculated gains envelope.

Figure 31 a shows the process flow diagram that produces the method M200 of high band excitation signal according to an embodiment.Task Y100 is calculated through harmonic wave and is extended signal by the narrow-band pumping signal that nonlinear function is applied to partly derive from the narrow-band of voice signal.Task Y200 will extend signal and mix to produce high band excitation signal through zoop signal through harmonic wave.Figure 31 b shows the process flow diagram that produces the method M210 of high band excitation signal according to another embodiment that comprises task Y300 and Y400.Task Y300 calculates temporal envelope according to the energy in time of narrow-band pumping signal and the one in harmonic wave extension signal.Task Y400 according to temporal envelope zoop signal to produce through zoop signal.

Figure 32 shows according to an embodiment having narrow-band partly and the process flow diagram of the method M300 that the described highband part of the voice signal of highband part is decoded.Task Z100 receives one group of filter parameter of the spectrum envelope that characterizes described highband part and characterizes one group of gain factor of the temporal envelope of described highband part.Task Z200 calculates frequency spectrum by the signal that nonlinear function is applied to partly derive from narrow-band and extends signal.Task Z300 basis (A) described group of filter parameter and (B) high band excitation signal based on described frequency spectrum extension signal produce synthetic high-frequency band signals.Task Z400 modulates the gain envelope of described synthetic high-frequency band signals based on described group of gain factor.For instance, task Z400 can be configured to pumping signal by described group of gain factor is applied to partly derive from narrow-band, be applied to frequency spectrum extends signal, is applied to high band excitation signal, or be applied to synthetic high-frequency band signals, modulate the gain envelope of described synthetic high-frequency band signals.

Embodiment also comprise as herein (for example) by be configured to carry out additional voice decoding, Code And Decode method structure embodiment description and clearly disclose as described in additional voice decoding, Code And Decode method.Each of these methods also (for example can positively be implemented, as one or more data storage mediums of above enumerating in) be one or more instruction groups that can for example, read and/or carry out by the machine that comprises logic element (, processor, microprocessor, microcontroller or other finite state machine) array.Therefore, the present invention is without wishing to be held to the embodiment above showing, but should meet with herein principle and the consistent widest range of novel feature that (in the appended claims of a part that comprises submitted to the original disclosure of formation) discloses by any way.

Claims (22)

1. coding has low-frequency band partly and a method for the described highband part of the voice signal of highband part, and described method comprises:

A plurality of filter parameters of the spectrum envelope of highband part described in computational representation;

By extending the frequency spectrum of the signal of partly deriving from described low-frequency band, calculate frequency spectrum extension signal;

Frequency spectrum is extended to signal and the addition of modulated noise signal, and so that high band excitation signal to be provided, wherein said modulated noise signal is to carry out which amplitude modulation according to the temporal envelope of narrow-band signal, narrow-band pumping signal or harmonic wave extension signal; And

Relation based between described highband part and the signal based on described low-frequency band part is carried out calculated gains envelope.

2. method according to claim 1, wherein said calculating frequency spectrum extends signal and comprises that the signal by nonlinear function being applied to partly derive from described low-frequency band extends the frequency spectrum of described signal.

3. method according to claim 2, wherein said calculated gains envelope is energy based on described highband part and the relation between the energy of the described signal based on described low-frequency band part.

4. method according to claim 1, described gain envelope comprises a plurality of gain factors, the different subframes in a series of continuous subframes of each frame corresponding to described highband part in described a plurality of gain factors, and

Described calculated gains envelope comprise energy based on the corresponding subframe of (A) described highband part and (B) relation between the energy of the corresponding time portion of the signal based on described low-frequency band part calculate each in described a plurality of gain factor.

5. method according to claim 1, it comprise according to (A) described high band excitation signal and (B) described a plurality of filter parameter generate synthetic high-frequency band signals, and wherein said signal based on described low-frequency band part is described synthetic high-frequency band signals.

6. method according to claim 5, wherein said gain envelope comprises a plurality of gain factors, the different subframes in a series of continuous subframes of each frame corresponding to described highband part in described a plurality of gain factors, and

Described calculated gains envelope comprise energy based on the corresponding subframe of (A) described highband part and (B) relation between the energy of the corresponding time portion of synthetic high-frequency band signals calculate each in described a plurality of gain factor.

7. method according to claim 1, wherein said calculating frequency spectrum extends signal and comprises by extend the frequency spectrum of described signal at the described signal application nonlinear function of partly deriving from described low-frequency band.

8. decoding has low-frequency band partly and a method for the described highband part of the voice signal of highband part, and described method comprises:

Receive a plurality of gain factors of a plurality of filter parameters of the spectrum envelope that characterizes described highband part and the temporal envelope of the described highband part of sign;

By the signal spectrum extending based on low band excitation signal, calculate frequency spectrum extension signal;

Frequency spectrum is extended to signal and the addition of modulated noise signal, and so that high band excitation signal to be provided, wherein said modulated noise signal is to carry out which amplitude modulation according to the temporal envelope of narrow-band signal, narrow-band pumping signal or harmonic wave extension signal;

According to (A) described a plurality of filter parameters and (B) described high band excitation signal produce synthetic high-frequency band signals; And

According to described a plurality of gain factors, revise the gain envelope of described synthetic high-frequency band signals.

9. method according to claim 8, wherein said calculating frequency spectrum extends signal and comprises the frequency spectrum that extends described signal by nonlinear function being applied to the described signal based on low band excitation signal.

10. method according to claim 8, wherein said modification gain envelope comprises according to described a plurality of gain factors revises at least one amplitude in time in the following: the signal based on described low band excitation signal, described frequency spectrum extend signal, described high band excitation signal and described synthetic high-frequency band signals.

11. 1 kinds are configured to the equipment of described highband part that coding has the voice signal of low-frequency band part and highband part, and described equipment comprises:

Analysis module, it is configured to one group of filter parameter of the spectrum envelope of highband part described in computational representation;

Frequency spectrum stretcher, it is configured to calculate frequency spectrum extension signal by extending the frequency spectrum of the signal of partly deriving from described low-frequency band;

Combiner, it is configured to by described frequency spectrum extension signal and modulated noise signal are added, high band excitation signal is provided, and wherein said modulated noise signal is to carry out which amplitude modulation according to the temporal envelope of narrow-band signal, narrow-band pumping signal or harmonic wave extension signal; And

Gain factor counter, its time-varying relationship being configured to based between described highband part and the signal based on described low-frequency band part carrys out calculated gains envelope.

12. equipment according to claim 11, wherein said frequency spectrum stretcher is configured to the frequency spectrum that signal by nonlinear function being applied to partly derive from described low-frequency band extends described signal.

13. equipment according to claim 12, the time-varying relationship that wherein said gain factor counter is configured between energy based on described highband part and the energy of the described signal based on described low-frequency band part calculates described gain envelope.

14. equipment according to claim 11, wherein said gain envelope comprises a plurality of gain factors, the different subframes in a series of continuous subframes of each frame corresponding to described highband part in described a plurality of gain factors, and

Described gain factor counter be configured to energy based on the corresponding subframe of (A) described highband part and (B) relation between the energy of the corresponding time portion of the signal based on described low-frequency band part calculate each in described a plurality of gain factor.

15. equipment according to claim 13, it comprises:

Composite filter, it is configured to generate synthetic high-frequency band signals according to described high band excitation signal and described one group of filter parameter, and wherein said signal based on low-frequency band part is described synthetic high-frequency band signals.

16. equipment according to claim 15, wherein said gain envelope comprises a plurality of gain factors, the different subframes in a series of continuous subframes of each frame corresponding to described highband part in described a plurality of gain factors, and

Described gain factor counter be configured to energy based on the corresponding subframe of (A) described highband part and (B) relation between the energy of the corresponding time portion of synthetic high-frequency band signals calculate each in described a plurality of gain factor.

17. equipment according to claim 11, wherein said frequency spectrum stretcher is configured to by extend the frequency spectrum of described signal at the described signal application nonlinear function of partly deriving from described low-frequency band.

18. equipment according to claim 11 are embedded in cell phone.

19. 1 kinds of high frequency band Voice decoders, it is configured to receive a plurality of filter parameters and (B) the encoded low band excitation signal of the part of the low-frequency band based on described voice signal of spectrum envelope that (A) characterizes the highband part of voice signal, and described demoder comprises:

Frequency spectrum stretcher, its signal spectrum being configured to by extending based on described encoded low band excitation signal calculates frequency spectrum extension signal;

Combiner, it is configured to by described frequency spectrum extension signal and modulated noise signal are added, high band excitation signal is provided, and wherein said modulated noise signal is to carry out which amplitude modulation according to the temporal envelope of narrow-band signal, narrow-band pumping signal or harmonic wave extension signal;

Composite filter, it is configured to produce synthetic high-frequency band signals according to described a plurality of filter parameters and described high band excitation signal; And

Gain control element, it is configured to modulate according to a plurality of gain factors that characterize the temporal envelope of described highband part the gain envelope of described synthetic high-frequency band signals.

20. demoders according to claim 19, wherein said frequency spectrum stretcher is configured to extend by nonlinear function being applied to signal based on described encoded low band excitation signal the frequency spectrum of described signal.

21. demoders according to claim 19, wherein said gain control element is configured to revise at least one amplitude in time in the following according to described a plurality of gain factors: described encoded low band excitation signal, described frequency spectrum extend signal, described high band excitation signal and described synthetic high-frequency band signals.

22. demoders according to claim 19, wherein said gain control element comprises at least one in multiplier and amplifier.

Images