WO2013143221A1 - 信号编码和解码的方法和设备 - Google Patents

信号编码和解码的方法和设备 Download PDF

Info

Publication number
WO2013143221A1
WO2013143221A1 PCT/CN2012/075924 CN2012075924W WO2013143221A1 WO 2013143221 A1 WO2013143221 A1 WO 2013143221A1 CN 2012075924 W CN2012075924 W CN 2012075924W WO 2013143221 A1 WO2013143221 A1 WO 2013143221A1
Authority
WO
WIPO (PCT)
Prior art keywords
frequency domain
signal
domain signal
frequency
decoded
Prior art date
Application number
PCT/CN2012/075924
Other languages
English (en)
French (fr)
Inventor
刘泽新
苗磊
齐峰岩
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority to MYPI2014002473A priority Critical patent/MY189975A/en
Priority to SG11201405216SA priority patent/SG11201405216SA/en
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Priority to MX2014011605A priority patent/MX339652B/es
Priority to JP2015502053A priority patent/JP6006400B2/ja
Priority to EP17160983.7A priority patent/EP3249645B1/en
Priority to CA2866202A priority patent/CA2866202C/en
Priority to EP12873219.5A priority patent/EP2809009B1/en
Priority to EP19191869.7A priority patent/EP3664085B1/en
Priority to ES12873219.5T priority patent/ES2655832T3/es
Priority to BR112014023577A priority patent/BR112014023577B8/pt
Priority to KR1020147026193A priority patent/KR101621641B1/ko
Priority to RU2014142255/08A priority patent/RU2592412C2/ru
Publication of WO2013143221A1 publication Critical patent/WO2013143221A1/zh
Priority to ZA2014/06424A priority patent/ZA201406424B/en
Priority to US14/496,986 priority patent/US9537694B2/en
Priority to US15/358,649 priority patent/US9786293B2/en
Priority to US15/684,079 priority patent/US9899033B2/en
Priority to US15/864,147 priority patent/US10600430B2/en

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L27/00—Modulated-carrier systems
    • H04L27/26—Systems using multi-frequency codes
    • H04L27/2601—Multicarrier modulation systems
    • H04L27/2602—Signal structure
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
    • G10L19/002—Dynamic bit allocation
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
    • G10L19/02—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using spectral analysis, e.g. transform vocoders or subband vocoders
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
    • G10L19/02—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using spectral analysis, e.g. transform vocoders or subband vocoders
    • G10L19/0204—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using spectral analysis, e.g. transform vocoders or subband vocoders using subband decomposition
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
    • G10L19/02—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using spectral analysis, e.g. transform vocoders or subband vocoders
    • G10L19/032—Quantisation or dequantisation of spectral components
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
    • G10L19/04—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using predictive techniques
    • G10L19/06—Determination or coding of the spectral characteristics, e.g. of the short-term prediction coefficients
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
    • G10L19/04—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using predictive techniques
    • G10L19/16—Vocoder architecture
    • G10L19/167—Audio streaming, i.e. formatting and decoding of an encoded audio signal representation into a data stream for transmission or storage purposes
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L5/00—Arrangements affording multiple use of the transmission path
    • H04L5/02—Channels characterised by the type of signal
    • H04L5/06—Channels characterised by the type of signal the signals being represented by different frequencies
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
    • G10L19/04—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using predictive techniques
    • G10L19/26—Pre-filtering or post-filtering

Definitions

  • Embodiments of the present invention relate to the field of domain communications and, more particularly, to a method and apparatus for signal encoding and decoding. Background technique
  • encoding techniques are used at the transmitting end to compress the signals to be transmitted to improve transmission efficiency, and corresponding decoding techniques are used at the receiving end to recover the transmitted signals.
  • the signal can be time domain coded and/or frequency domain coded depending on the characteristics of the signal, the transmission conditions, and the like. According to certain rules, different coding ratios are allocated for the time domain signal or the frequency domain signal, and it is desirable to characterize the signal to be transmitted with as few coded bits as possible. Therefore, it is necessary to allocate the coded bits reasonably so that the output signal is recovered as little as possible by decoding at the receiving end.
  • the speech can have a better codec effect, but for music, the encoding and decoding effect is relatively poor.
  • the input signal is encoded by a time domain coding method using partial bits; and, based on the input signal.
  • the frequency domain signal is encoded using the remaining bits, the signal characteristics are usually not considered, and the frequency domain signal is bit-allocated in a uniform manner, which results in poor encoding of the partial frequency domain signal.
  • the frequency domain signal is simply recovered by using a decoding technique corresponding to the encoding technique, and the undecoded frequency domain signal is filled with noise, and then inverse frequency domain transform and time domain synthesis are performed. Processing to obtain an output signal.
  • the noise fill introduces additional noise in some of the signals, reducing the quality of the output signal.
  • Embodiments of the present invention provide a method and a device for encoding and decoding a signal, which can optimize bit allocation of a frequency domain signal during encoding to achieve a better coding effect by using the same bit, and can be decoded during decoding.
  • the frequency domain excitation signal is expanded to achieve a better output signal.
  • a method for signal encoding comprising: obtaining a frequency domain signal according to an input signal; assigning a predetermined bit to the frequency domain signal according to a predetermined allocation rule; and frequency domain signal having a bit allocation When the highest frequency is greater than a predetermined value, the bit allocation of the frequency domain signal is adjusted; the frequency domain signal is encoded according to the bit allocation of the frequency domain signal.
  • a method for signal decoding comprising: obtaining a decoded frequency domain signal from a received bitstream; and if the decoded frequency domain signal satisfies a predetermined condition, The undecoded frequency domain signal is predicted according to the decoded frequency domain signal; and the final output time domain signal is obtained according to the decoded frequency domain signal and the predicted undecoded frequency domain signal.
  • an apparatus for signal encoding comprising: a frequency domain transform unit that obtains a frequency domain signal according to an input signal; and a bit allocation unit that allocates a predetermined bit to the frequency domain according to a predetermined allocation rule a bit adjustment unit that adjusts a bit allocation of the frequency domain signal when a highest frequency of a frequency domain signal having a bit allocation is greater than or equal to a predetermined value; and a frequency domain coding unit that allocates a frequency domain according to a bit of the frequency domain signal The signal is encoded.
  • an apparatus for signal decoding comprising: a decoding unit, obtaining a decoded frequency domain signal from a received bitstream; and a spreading unit, configured to predict an undecoded frequency domain a signal, when the decoded frequency domain signal satisfies a predetermined condition, predicting an undecoded frequency domain signal according to the decoded frequency domain signal; and outputting the unit according to the decoded frequency domain signal and the predicted frequency The domain signal is used to obtain the final output time domain signal.
  • the bit allocation of the frequency domain signal is adjusted by encoding according to the highest frequency of the frequency domain signal with bit allocation, and the frequency domain coding is performed with the same number of bits. Good coding effect;
  • the decoded frequency domain signal is used as a guide to set the undecoded frequency domain signal, so that the output signal achieves better results.
  • FIG. 1 illustrates a method of encoding a signal in accordance with an embodiment of the present invention
  • FIG. 3 illustrates a method of decoding a signal in accordance with an embodiment of the present invention
  • FIG. 4 illustrates a method of obtaining a decoded frequency domain signal from a received bit stream in a time-frequency joint decoding method
  • Figure 5 illustrates an exemplary implementation of an encoding device and/or a decoding device in accordance with the present invention
  • FIG. 6 illustrates an encoding device that encodes a signal in accordance with an embodiment of the present invention
  • FIG. 8 illustrates an apparatus for decoding a signal in accordance with an embodiment of the present invention
  • Figure 9 illustrates a block diagram of a decoding unit in time-frequency joint decoding. detailed description
  • the coding technical solution and the decoding technical solution of the present invention can be applied to transmission and reception in various communication systems, such as: GSM, Code Division Multiple Access (CDMA), Wideband Code Division WCDMA (Wandaband Code Division Multiple Access Wireless), General Packet Radio Service (GPRS), Long Term Evolution (LTE), etc.
  • GSM Global System for Mobile Communications
  • CDMA Code Division Multiple Access
  • WCDMA Wideband Code Division WCDMA
  • GPRS General Packet Radio Service
  • LTE Long Term Evolution
  • Coding technology solutions and decoding technology solutions widely used in a variety of electronic devices, such as: mobile phones, wireless devices, personal data assistants (PDA), handheld or portable computers, GPS receivers / navigators, cameras, audio / video Players, camcorders, video recorders, surveillance equipment, etc. usually,
  • PDA personal data assistants
  • Such an electronic device includes an audio encoder or an audio decoder, and the audio encoder or decoder can be directly implemented by a digital circuit or a chip such as a DSP (digital signal processor), or the software code drives the processor to execute a process in the software code. achieve.
  • DSP digital signal processor
  • an audio time domain signal is first transformed into a frequency domain signal, and then a coded bit is allocated to an audio frequency domain signal for encoding, and the encoded signal is transmitted to a decoding end through a communication system.
  • the decoding end decodes and recovers the encoded signal.
  • FIG. 1 illustrates a method 100 of encoding a signal in accordance with an embodiment of the present invention. As shown in Figure 1, the method includes:
  • the frequency domain signal is obtained according to the input signal.
  • the input signals may be of various types such as image signals, data signals, audio signals, video signals, text signals, and the like.
  • the frequency domain signal can be obtained by frequency-domain transforming the input signal by using an algorithm such as Fast Fourier Transform (FFT) or Discrete Cosine Transform (DCT).
  • FFT Fast Fourier Transform
  • DCT Discrete Cosine Transform
  • the type of input signal and the frequency domain transform algorithm do not constitute a limitation of the present invention.
  • the predetermined bit tot-bit is a bit to be used for frequency domain coding of a frequency domain signal.
  • the predetermined allocation rule may be, for example, allocating a larger number of bits in a predetermined bit to a low frequency band signal in a frequency domain signal, and allocating remaining bits in a predetermined bit to a larger energy than the low frequency band signal Frequency band.
  • the more bits may be allocated in the low band signal either equally for all low frequency bands or according to the energy distribution of the low band signals.
  • the reason for allocating more bits for low-band signals is that, in speech-audio signals, for example, low-band signals typically contain more sensitive information from the human ear.
  • the frequency domain signal is usually divided into subbands at equal intervals in frequency, or subbands are divided according to frequency domain coefficients, for example, one subband per 16 frequency domain coefficients. For example, for a wideband signal of 20 ms-frame, 160 coefficients in the frequency range of 0 ⁇ 4 kHz are divided into 10 sub-bands, wherein there are 5 sub-bands in the frequency range of 0 ⁇ 2 kHz and 5 sub-bands in the frequency range of 2 ⁇ 4 kHz. Then, bit allocation is performed for each subband.
  • the number of bits of the IF bit is allocated, and the predetermined bit tot bit is subtracted from the IF bit to obtain the remaining bit rest bit, and according to the frequency range of 2 ⁇ 4 kHz
  • the envelope size of each subband allocates the remaining bit rest bits to subbands in the frequency range of 2 ⁇ 4 kHz, each subband being 5 bits. Determine the number of subbands with bit allocation and bits according to the size of the rest_bits and the envelope of each subband
  • the subband of the highest frequency band is allocated last-bin, and the remainder that cannot be divisible by 5 is equally distributed to each subband in the range of 0 ⁇ 2 kHz.
  • the predetermined value B may be set according to an empirical value; in one embodiment, the number of bits of the predetermined bit tot_bit and the resolution of the frequency domain signal may be used (for example, there are 320 frequency domains in the 0 ⁇ 8 kHz bandwidth range) Coefficient) to determine the predetermined value B. In the case of a fixed bandwidth, the number of bits of the predetermined bit tot_bit is higher, and the predetermined value B is higher; the number of bits of the predetermined bit is fixed. When the bit is fixed, the higher the resolution of the frequency domain signal, the higher the predetermined value B.
  • the predetermined value B can be determined only based on the number of bits tot_bit of the predetermined bit, and the more the number of bits tot_bit of the predetermined bit, the higher the predetermined value B.
  • the predetermined value B is a preset upper limit frequency value. For example, empirically, after frequency domain transform of an input signal, a frequency domain signal having a frequency greater than the predetermined value is usually not allocated bits. Therefore, in a specific practice, the predetermined value B can be set to a frequency value lower than the highest frequency value of the frequency domain signal by a certain frequency, for example, set to 2.9 kHz, 3.2 kHz, 3.5 kHz, or the like. In other embodiments, the predetermined value B may also be determined based on other factors such as the frame length, the transform method employed, or the transform window length.
  • the predetermined value B may be an index number of 20 subbands in a frequency range of 0 to 8 kHz, and the highest frequency of the frequency domain signal having bit allocation is also It can be represented by the index number of the subband in which the highest frequency is located.
  • the predetermined value B and the highest frequency of the bit-assigned frequency domain signal are not limited to the frequency value, and may also be the index number of the sub-band. After reading the disclosure of the embodiments of the present invention, the engineering technician knows based on the practical conditions how to determine whether the highest frequency of the frequency domain signal having the bit allocation is greater than a predetermined value.
  • the adjustment of the bit allocation of the frequency domain signal is described below. Depending on the type of the signal or the frequency domain characteristics, etc., it is possible to reduce a portion of the frequency domain signal that contributes less to the output of the decoding side, and correspondingly increase the bit frequency of the highest frequency with bit allocation and the frequency domain signal in the vicinity thereof. That is, the adjusting the bit allocation of the frequency domain signal may include: reducing the allocation of the frequency band of the frequency domain signal to which more bits are allocated. The number of bits, and the maximum frequency with bit allocation and the number of bits allocated by the frequency domain signals in the vicinity thereof are increased. For audio signals, the frequency band to which more bits are allocated is, for example, a low frequency band of 0 to 2 kHz. The following is an illustration of the adjustment of the bit allocation to the frequency domain signal.
  • Adjustment example 1 The highest frequency with bit allocation is 4 kHz. If a sub-band bit in the range of 2 kHz to 4 kHz is allocated 0, 5 bits are allocated to this band until all sub-bands in the range of 2 kHz to 4 kHz are allocated to the number of bits. Assume that the additional number of bits in the range of 2 ⁇ 4 kHz is N blt . At this time, it is necessary to reduce N blt bits from the sub-bands in the range of 0 to 2 kHz.
  • the algorithm used is, for example: reducing 1 bit per subband from all subbands (5 subbands) in the range of 0 ⁇ 2 kHz; then subtracting one of the highest frequency subbands; from the remaining 4 subbands Each sub-band is further reduced by 1 bit, then reduced by a sub-high frequency sub-band, and so on, until the reduced number of bits equals N blt .
  • Adjustment example 2 Add J bits to all subbands of the allocated bits in the range of 2 kHz to 4 kHz. If the number of subbands with bit allocation in the range of 2 to 4 kHz is K, then additional bits in the range of 2 to 4 kHz are added at this time.
  • an algorithm that can be used is: averaging N blt /5 bits per subband from all subbands (5 subbands) in the range of 0 to 2 kHz.
  • the algorithm used can be: Adjust the algorithm in Example 1 and adjust any of the algorithms in Example 2.
  • FIG. 2 illustrates a time-frequency joint encoding method 200 in accordance with an embodiment of the present invention.
  • 220, 230, 240 are the same as 120, 130, 140 in Fig. 1, respectively. 2 differs from FIG. 1 in that steps 250, 260 are added, and 110 in FIG. 1 is replaced with 211 and 212.
  • steps 250, 260 are added, and 110 in FIG. 1 is replaced with 211 and 212.
  • the differences between Fig. 2 and Fig. 1 will be described hereinafter, and the same points will not be repeated.
  • obtaining a first time domain signal and a second time domain signal by performing time domain analysis on the input signal For example, linear inputive coding (LPC) analysis and processing of the input signal yields one of Line Spectral Frequency (LSF) parameters and Immittance Spectral Frequency (ISF) parameters, and is also disabled.
  • LSF Line Spectral Frequency
  • ISF Immittance Spectral Frequency
  • the difference signal res and the adaptive codebook contribute to exc_pit.
  • the LSF parameter or ISF parameter is used to indicate the frequency domain characteristics of the coefficients (i.e., LPC coefficients) used in the LPC analysis.
  • the residual signal res and the adaptive codebook contribution exc_pit are included in the first time domain signal, and the adaptive codebook contribution exc_pit is included in the second time domain signal.
  • the residual signal res and the adaptive codebook contribution exc_pit in the first time domain signal are respectively subjected to frequency domain transform, and then the residual signal f_res in the frequency domain and the adaptive codebook in the frequency domain contribute f_
  • the correlation of exc_pit is used to determine whether the adaptive codebook contribution contributes to the output signal. If the adaptive codebook contribution contributes to the output signal, the frequency domain adaptive codebook contribution f_exc_pit is subtracted from the frequency domain residual signal f_res to obtain the frequency domain difference signal f_diff, and The difference signal f_diff is used as the frequency domain signal. If the adaptive codebook contribution does not contribute to the output signal, the residual signal f_res in the frequency domain is directly used as the difference signal f-diff, that is, the frequency domain signal.
  • the frequency domain signals are encoded by the same 220, 230, 240 as 120, 130, 140 in Fig. 1, and the encoded frequency domain signals are obtained.
  • the time domain signal may be encoded using any time domain coding method such as predictive coding, Pulse Code Modulation (PCM) coding, etc., and the time domain coding method employed does not constitute a limitation of the present invention.
  • PCM Pulse Code Modulation
  • the adaptive codebook contribution contributes to the output signal, the adaptive codebook contribution needs to be obtained at the decoding end, so the adaptive codebook contribution exc_pit in the second time domain signal is encoded to be transmitted as a bit stream to the reception. end.
  • the adaptive codebook contribution does not contribute to the output signal, that is, the output of the decoder does not require an adaptive codebook contribution, then the time domain coding of the portion is not required, thereby improving coding efficiency.
  • the adaptive codebook contribution contributes to the output signal means decoding The terminal cannot obtain a high quality output signal based only on the encoded frequency domain signal.
  • the frequency domain signal to be subjected to frequency domain coding may include other signals, such as a flag flag indicating whether the adaptive codebook contribution contributes to the output signal, in addition to the difference signal f-diff.
  • the second time domain signal to be time domain coded may include, in addition to the adaptive codebook contribution exc_pit, other information needed for decoding.
  • the bit allocation of the frequency domain signal is adjusted according to the highest frequency of the frequency domain signal with bit allocation, and combined with the time domain coding, thereby achieving a better coding effect.
  • FIG. 3 illustrates a method 300 of decoding a signal in accordance with an embodiment of the present invention.
  • the method 300 includes:
  • the decoded frequency domain signal is obtained from the received bitstream by using a frequency domain decoding method corresponding to the frequency domain coding method.
  • the decoded frequency domain signal is obtained from the received bitstream by performing frequency domain decoding on the frequency domain information in the bitstream to obtain a first frequency domain signal; according to the first frequency domain The signal determines whether there is a time domain coded signal contributing to the output signal in the bit stream; when it is determined that there is a time domain coded signal contributing to the output signal in the bit stream, time domain coded signal is time domain decoded and frequency domain Transforming to obtain a second frequency domain signal, and synthesizing the first frequency domain signal and the second frequency domain signal to obtain the decoded frequency domain signal, which will be described in further detail below in conjunction with FIG.
  • the decoded frequency domain signal When the decoded frequency domain signal satisfies a predetermined condition, predicting the undecoded frequency domain signal according to the decoded frequency domain signal.
  • the decoded frequency domain signal satisfies predetermined conditions, including: the highest frequency of the decoded frequency domain signal is greater than a predetermined value, and the decoded frequency domain signal includes a frequency domain transformed contribution to the output signal. At least one of the time domain coded signals.
  • the decoded frequency domain signal may be applied first, including a frequency domain transformed time domain coded signal that contributes to the output signal, and then the highest frequency of the decoded frequency domain signal is greater than
  • the judgment condition of the predetermined value, or the order of the reverse order, may also be used only as described above in connection with 130 of FIG. 1, the predetermined value is a reservation according to frequency domain coding.
  • the number of bits is determined by the bit-to-bit and resolution of the frequency domain signal. According to practical needs, the predetermined value can be set to a frequency value lower than the highest frequency value of the frequency domain signal by a certain frequency.
  • the predetermined value may be an index number of the subband, and the highest frequency of the frequency domain signal having the bit allocation at this time also uses the subband of the highest frequency domain.
  • the index number indicates.
  • the value of the predetermined value of the decoding end may be the same as or different from the value of the predetermined value of the encoding end.
  • the bit stream is decoded at 310 to obtain a time domain coded signal that may include a frequency domain transform and contributes to an output signal, and the frequency domain transformed
  • the time domain coded signal contributing to the output signal is, for example, a signal obtained by time domain decoding and frequency domain transform of time domain coded information contained in the bit stream, such as contribution to an adaptive codebook.
  • the time domain coded signal that contributes to the output signal after the frequency domain transformation may be in addition to the adaptive codebook. Other signals than contributions.
  • the decoded frequency domain signal includes an adaptive codebook contribution
  • whether the decoded frequency domain signal includes a frequency may be learned according to whether the adaptive codebook contributes a flag flag that contributes to the output signal.
  • the decoded frequency domain signal includes a time domain coded signal that contributes to the output signal after frequency domain transformation, which means that it is difficult to obtain high quality output by frequency domain decoding only, and the characteristics of the audio signal are Simply setting the undecoded frequency domain signal to noise degrades the output signal quality, requiring prediction of the undecoded frequency domain signal.
  • a frequency domain signal of a frequency band selected from a highest frequency of the decoded frequency domain signal to a low frequency may be selected according to the selected
  • the frequency domain signal is used to predict the undecoded frequency domain signal. For example, for a signal with a frame length of 20 ms and a sample rate of 12.8 kHz, the frequency domain coefficient is 256 and the bandwidth is 6.4 kHz. At 7.6 kbps, there are 16 subbands for every 16 coefficients, and a total of 16 subbands are reserved. The value is set to 10 (4 kHz).
  • the undecoded frequency domain coefficients in the range of 4 to 6.4 kHz are predicted by the frequency domain coefficients decoded in the range of 1.6 to 4 kHz.
  • the undecoded frequency domain signal can be predicted by performing normalization processing, envelope processing, or the like on the selected frequency domain signal.
  • the implementation of the normalization process and the envelope process is a means known to those skilled in the art and will not be described in detail herein.
  • the method can predict the undecoded frequency domain signal.
  • the undecoded frequency domain signal can also be predicted according to the frequency domain signal of the fixed frequency band in the decoded frequency domain signal.
  • the ISF parameter or the LSF parameter from the encoding end may be adopted.
  • the predicted undecoded frequency domain coefficients are corrected.
  • the formant position is estimated by LSF parameters or ISF parameters; the frequency domain coefficients with larger amplitudes are scaled at each estimated formant position.
  • a threshold which may be set according to characteristics of the time domain analysis of the encoding end
  • decrease near the position of the formant The magnitude of the predicted frequency domain coefficient.
  • noise is used to predict the undecoded frequency domain signal.
  • the decoded frequency domain signal is obtained by decoding, and the undecoded frequency domain signal is predicted, thereby obtaining the frequency domain signal in the entire frequency band, by performing, for example, Inverse Fast Fourier Transform (IFFT)
  • IFFT Inverse Fast Fourier Transform
  • the inverse frequency transform or the like is processed to obtain an output signal in the time domain.
  • the ISF parameter or the LSF parameter is transformed to obtain an LPC coefficient, and the LPC coefficient is used to perform time domain synthesis on the signal obtained after inverse frequency domain transformation to obtain a time domain of the final output. signal.
  • IFFT Inverse Fast Fourier Transform
  • the output signal is better. effect.
  • the decoding method according to an embodiment of the present invention is applied in a time-frequency joint decoding scheme.
  • the subsequent operations are the same as those described in connection with Fig. 3 except for the step of obtaining the decoded frequency domain signal (310) from the received bit stream. Therefore, only how to obtain the decoded frequency domain signal in the time-frequency joint decoding method will be described below.
  • the method 410 includes: 411: Demultiplex the bitstream into a first set of bits and a second set of bits. Upon decoding at the receiving end, upon receiving the bitstream, the bitstream is demultiplexed into a first set of bits and a second set of bits using a demultiplexing technique corresponding to the multiplexing technique of 260 of FIG.
  • the first set of bits includes frequency domain information to be subjected to frequency domain decoding as described below
  • the second set of bits includes a time domain coded signal that contributes to the output signal to be subjected to the following time domain decoding.
  • the first set of bits includes, for example, a difference signal f-diff, a flag flag indicating whether the adaptive codebook contribution contributes to the output signal, and the like.
  • the second set of bits includes an adaptive codebook contribution, for example, when the adaptive codebook contribution contributes to the output signal. It is noted that the first set of bits and the second set of bits may also include other signals corresponding to the encoding of the signals.
  • the fourth12 Perform frequency domain decoding on the first group of bits to obtain a first frequency domain signal, and determine, according to the first frequency domain signal, whether a time domain coded signal that contributes to the output signal exists in the bit stream.
  • the first set of bits is decoded by a decoding method corresponding to the frequency domain encoding method at the encoding end to obtain a first frequency domain signal.
  • the first frequency domain signal includes, for example, a decoded difference signal f-diff, and a flag flag indicating whether the adaptive codebook contribution contributes to the output signal.
  • the second set of bits is decoded by a decoding method corresponding to the time domain encoding method of the encoding end to obtain a decoded time domain signal. Specifically, when it is determined that there is a time domain coded signal contributing to the output signal in the bit stream, the time domain coded signal in the second group of bits is time domain decoded.
  • the frequency is synthesized by adding the difference signal f_diff in the first frequency domain signal and the adaptive codebook contribution in the second frequency domain signal. Domain signal.
  • the difference signal f_diff in the first frequency domain signal is directly used as the frequency domain signal.
  • the present invention further provides an encoding device and a decoding device, which may be located in a terminal device, a network device, or a testing device.
  • the compilation The code device or the decoding device may be implemented by a hardware circuit or by software in conjunction with hardware.
  • Figure 5 illustrates an exemplary implementation of an encoding device and/or a decoding device in accordance with the present invention.
  • the encoding device or decoding device 530 is invoked by a processor 510 via the input/output interface 520 to effect encoding or decoding of the audio signal with the aid of the memory 540.
  • the encoding device or decoding device 530 can perform various methods and processes in the above method embodiments.
  • FIG. 6 illustrates an encoding device 600 that encodes a signal in accordance with an embodiment of the present invention.
  • the encoding device 600 includes: a frequency domain transform unit 610 that obtains a frequency domain signal according to an input signal; a bit allocation unit 620 that allocates a predetermined bit to the frequency domain signal according to a predetermined allocation rule; and a bit adjustment unit 630 that has a bit allocation When the highest frequency of the frequency domain signal is greater than or equal to a predetermined value, the bit allocation of the frequency domain signal is adjusted; the frequency domain encoding unit 640 encodes the frequency domain signal according to the adjusted bit allocation.
  • the frequency domain transform unit 610 can obtain a frequency domain signal based on the input signal.
  • the input signal can be various types of signals such as image signals, data signals, audio signals, video signals, text signals, and the like.
  • the frequency domain signal can be obtained by performing frequency domain transform on the input signal by using an algorithm such as FFT or DCT.
  • the type of input signal and the frequency domain transform algorithm do not constitute a limitation of the present invention.
  • Bit allocation unit 620 can assign a predetermined bit tot-bit to the frequency domain signal in accordance with a predetermined allocation rule.
  • the tot bit is the number of bits to be used to encode the frequency domain signal.
  • the predetermined allocation rule may be, for example, allocating a larger number of the predetermined bits to a low frequency band signal in a frequency domain signal, and allocating remaining bits in the predetermined bit to energy other than the low frequency band signal Large frequency band.
  • the more bits may be allocated in the low frequency band signal for all low frequency bands or according to the energy distribution of the low frequency band signals.
  • the reason for allocating more bits for low-band signals is that audio signals such as speech are mainly concentrated in the low frequency range in the frequency domain, and allocating more bits to them can improve the efficiency of frequency domain coding.
  • the frequency domain signal in the frequency range of 0 to 4 kHz is divided into 10 subbands, wherein the frequency range of 0 to 2 kHz is as described above in connection with 120 of FIG. There are 5 sub-bands inside, and there are 5 sub-bands in the frequency range of 2 ⁇ 4kHz. Then, bit allocation is performed for each subband. A number of bits of the IF-bit are allocated for the low-frequency frequency domain signals in the frequency range of 0 to 2 kHz.
  • the remaining bits rest-bit (tot-bit minus IF-bit) are allocated to sub-bands in the frequency range of 2 ⁇ 4 kHz according to the envelope of each sub-band in the frequency range of 2 ⁇ 4 kHz. Specifically, according to rest_bits The subband of the high frequency band is last-bin, and the remainder that cannot be divisible by 5 is equally distributed to each subband in the range of 0 ⁇ 2 kHz.
  • the bit adjustment unit 630 may adjust the bit allocation of the frequency domain signal when the highest frequency of the frequency-domain signal having the bit allocation is greater than or equal to a predetermined value B.
  • the predetermined value B is determined based on the number of bits tot_bit of the predetermined bit and the resolution of the frequency domain signal (e.g., 4 kHz).
  • the predetermined value is a preset upper limit frequency value.
  • the predetermined value B may be a frequency value lower than a highest frequency value (e.g., 4 kHz) of the frequency domain signal by a certain frequency, for example, 2.9 kHz, 3.2 kHz, 3.5 kHz, or the like.
  • the predetermined value B may be an index number (for example, 7 or 8) of 10 subbands in a frequency range of 0 to 4 kHz,
  • the highest frequency of the frequency domain signal having the bit allocation is also represented by the index number iridex of the subband in which the highest frequency is located.
  • the bit adjustment unit 630 may adjust the bit allocation of the frequency domain signal by the bit allocation unit 620 according to a predetermined allocation rule when the highest frequency is greater than or equal to a predetermined value. Depending on the type of the input signal or the frequency domain characteristics of the frequency domain signal, etc., it is possible to reduce a portion of the frequency domain signal that contributes less to the output of the decoding end, and correspondingly increase the highest frequency with bit allocation and the frequency in the vicinity thereof. Bit allocation of the domain signal. As an example, the bit adjustment unit 630 may reduce the number of bits allocated by the frequency band to which more bits are allocated in the frequency domain signal, and increase the highest frequency with bit allocation and the number of bits allocated by the frequency domain signal in the vicinity thereof. . For audio signals, the frequency band to which more bits are allocated is, for example, a low frequency band of 0 to 2 kHz.
  • the frequency domain encoding unit 640 encodes the frequency domain signal according to the adjusted bit allocation.
  • the method of encoding the frequency domain signal may be, for example, transform coding, subband coding, or the like. Further, when the highest frequency is less than a predetermined value, the bit adjustment unit 630 does not adjust the bit allocation of the frequency domain signal. at this time, code.
  • the time-frequency joint coding apparatus 700 includes: a time domain analysis unit 711, which obtains a first time domain signal and a second time domain signal by performing time domain analysis on the input signal; and the frequency domain transform unit 712 performs the first time domain signal Frequency domain transform and processing to obtain a frequency domain signal; bit allocation unit 720, assigning a predetermined bit to the frequency domain signal according to a predetermined allocation rule; and bit adjusting unit 730, the highest frequency of the frequency domain signal having bit allocation is greater than or equal to a predetermined value Adjusting the bit allocation of the frequency domain signal; the frequency domain encoding unit 740 encodes the frequency domain signal according to the adjusted bit allocation; the time domain encoding unit 750 encodes the second time domain signal; the bit multiplexing unit 760.
  • the coded frequency domain signal and the encoded second time domain signal are multiplexed into a bit stream.
  • bit allocation unit 720, the bit adjustment unit 730, and the frequency domain coding unit 740 are the same as the bit allocation unit 620, the bit adjustment unit 630, and the frequency domain coding unit 640 in FIG. 6, respectively.
  • 7 is different from FIG. 6 in that the time domain coding unit 750, the bit multiplexing unit 760 are added, and the frequency domain transform unit 610 in FIG. 6 is replaced by the time domain analysis unit 711 and the frequency domain transform unit 712.
  • the differences between Fig. 7 and Fig. 6 will be described hereinafter, and the same points will not be repeated.
  • the time domain analyzing unit 711 obtains the first time domain signal and the second time domain signal by performing time domain analysis on the input signal. For example, LPC analysis and processing of the input signal yields ISF parameters (or LSF parameters), residual signal res, and adaptive codebook contribution exc_pit. The residual signal res and the adaptive codebook contribution exc_pit are used as the first time domain signal, and the adaptive codebook contribution exc_pit is used as the second time domain signal.
  • ISF parameters or LSF parameters
  • residual signal res and the adaptive codebook contribution exc_pit are used as the first time domain signal
  • the adaptive codebook contribution exc_pit is used as the second time domain signal.
  • the frequency domain transform unit 712 can obtain the frequency domain signal by performing frequency domain transform and processing on the first time domain signal.
  • the residual signal res and the adaptive codebook contribution exc_pit in the first time domain signal are respectively subjected to frequency domain transform, and then the residual signal f_ res in the frequency domain and the adaptive codebook contribution in the frequency domain are f – The correlation of exc_pit to determine if the adaptive codebook contribution contributes to the output signal. If the adaptive codebook contribution contributes to the output signal, the frequency domain adaptive codebook contribution f_exc_pit is subtracted from the frequency domain residual signal f_res to obtain the frequency domain difference signal f_diff, and The difference signal f_diff is included in the Frequency domain signal.
  • the residual signal f_res of the frequency domain is directly used as the difference signal f_diff for transmission as a frequency domain signal.
  • the frequency domain signal may include other signals in addition to the difference signal f-diff, such as a flag flag indicating whether the adaptive codebook contribution contributes to the output signal.
  • bit allocation unit 720 and the bit adjustment unit in FIG. 7 are utilized.
  • the frequency domain coding unit 740 encodes the frequency domain signal to obtain the encoded frequency domain signal.
  • the time domain encoding unit 750 can encode the second time domain signal.
  • the time domain signal may be encoded using a time domain coding method such as predictive coding, pulse code modulation, and the like.
  • a time domain coding method such as predictive coding, pulse code modulation, and the like.
  • an adaptive codebook contribution is required at the decoding end, so the adaptive codebook contribution exc_pit in the second time domain signal is encoded for transmission to the receiving end.
  • the bit multiplexing unit 760 can multiplex the encoded frequency domain signal and the encoded second time domain signal into a bit stream.
  • the bit allocation of the frequency domain signal is adjusted according to the highest frequency of the frequency domain signal having the bit allocation, and combined with the time domain coding, thereby achieving better coding. effect.
  • FIG. 8 illustrates a decoding device 800 that decodes a signal in accordance with an embodiment of the present invention.
  • the decoding device 800 includes: a decoding unit 810, which obtains a decoded frequency domain signal from a received bitstream; a spreading unit 820, configured to predict an undecoded frequency domain signal, where the decoded frequency domain signal satisfies In the case of a predetermined condition, the undecoded frequency domain signal is predicted based on the decoded frequency domain signal; and the output unit 830 obtains the final output time domain signal based on the decoded frequency domain signal and the predicted frequency domain signal.
  • the decoding unit 810 can obtain the decoded frequency domain signal from the received bit stream.
  • the decoded frequency domain signal is obtained from the received bit stream by using a frequency domain decoding method corresponding to the frequency domain coding method.
  • the decoding unit 810 can obtain the decoded frequency domain signal from the received bit stream by performing frequency domain decoding on the frequency domain information in the bitstream to obtain the first frequency domain signal;
  • a frequency domain signal determines whether there is a time domain coded signal contributing to the output signal in the bit stream; when it is determined that there is a time domain coded signal contributing to the output signal in the bit stream, time domain coded signal is time domain decoded And frequency domain transform to obtain a second frequency domain signal, and synthesizing the first frequency domain signal and the second frequency domain signal to obtain the decoded frequency domain signal, which This will be described in detail below in conjunction with FIG.
  • the spreading unit 820 can be used to predict undecoded frequency domain signals.
  • the spreading unit 820 can predict the undecoded frequency domain signal based on the decoded frequency domain signal.
  • the decoded frequency domain signal satisfies predetermined conditions, including: the highest frequency of the decoded frequency domain signal is greater than a predetermined value, and the decoded frequency domain signal includes a frequency domain transformed contribution to the output signal. At least one of the time domain coded signals. In practice, you can make choices as needed. The resolution of the tot-bit and frequency domain signals is determined.
  • the predetermined value can be set to a frequency value lower than the highest frequency value of the frequency domain signal by a certain frequency according to practical needs.
  • the predetermined value may be an index number of the subband, and the highest frequency of the frequency domain signal having the bit allocation at this time also uses the subband of the highest frequency domain. The index number indicates.
  • the decoded frequency domain signal obtained by decoding the bit stream in the decoding unit 810 may include obtaining time domain decoding and frequency domain transform in the time domain information included in the bit stream.
  • the signal which for example contributes to the adaptive codebook.
  • Whether the frequency domain signal includes a time domain coded signal that contributes to the output signal after the frequency domain transformation may be known according to whether the adaptive codebook contributes a flag flag that contributes to the output signal.
  • the time domain coded signal that contributes to the output signal after the frequency domain transformation may be other signals.
  • the decoded frequency domain signal includes a signal obtained by performing time domain decoding and frequency domain transform on the time domain information included in the bit stream, which indicates that the undecoded frequency domain signal includes information useful for output, thereby requiring Predicting the undecoded frequency domain signal and simply setting the undecoded frequency domain signal to noise degrades the output signal quality.
  • the spreading unit 820 may set the undecoded frequency domain signal to noise when the decoded frequency domain signal does not satisfy the predetermined condition.
  • the spreading unit 820 may start a frequency domain signal of a frequency band selected from a highest frequency of the decoded frequency domain signal to a low frequency. And selecting the selected frequency domain signal as described above to predict the undecoded frequency domain signal based on the selected frequency domain signal.
  • the output frequency domain signal may, for example, predict the undecoded frequency domain signal based on the frequency domain signal of the fixed frequency band in the decoded frequency domain signal.
  • the output unit 830 can obtain the final output time domain signal based on the decoded frequency domain signal and the predicted frequency domain signal. After the undecoded frequency domain signal is predicted, the frequency domain signal in the entire frequency band is obtained, and the frequency domain inverse transform of the entire bandwidth is performed by using an inverse transform of the frequency domain transform used in the encoding, Thereby the output signal of the time domain is obtained. As described above, the output unit can obtain the final output time domain signal for output by performing time domain synthesis on the signal after inverse frequency domain transformation using the LPC coefficient obtained from the ISF parameter (or LSF parameter).
  • the undecoded frequency domain signal is set by using the decoded frequency domain signal as a guide, so that the output signal reaches the output signal. Good results.
  • Figure 9 illustrates a block diagram of decoding unit 910 in time-frequency joint decoding.
  • the decoding unit 910 includes: a demultiplexing unit 911, which demultiplexes the bit stream into a first group of bits and a second group of bits; and a frequency domain decoding unit 912 that performs frequency domain decoding on the first group of bits to obtain a first frequency domain signal.
  • the time domain decoding unit 913 when determining that there is a time domain coded signal contributing to the output signal in the bit stream Performing time domain decoding on the second group of bits; the frequency domain transforming unit 914 performs frequency domain transform on the decoded time domain signal to obtain a second frequency domain signal; the synthesizing unit 915, the first frequency domain signal and the second The frequency domain signal is synthesized to obtain a decoded frequency domain signal.
  • the disclosed apparatus and method may be implemented in other ways.
  • the device embodiments described above are only schematic.
  • the division of the unit is only a logical function division.
  • there may be another division manner for example, multiple units or components may be combined or Can be integrated into another system, or some features can be ignored, or not executed.
  • each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
  • the functions, if implemented in the form of software functional units and sold or used as separate products, may be stored in a computer readable storage medium.
  • the technical solution of the present invention which is essential or contributes to the prior art, or a part of the technical solution, may be embodied in the form of a software product, which is stored in a storage medium, including
  • the instructions are used to cause a computer device (which may be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention.
  • the foregoing storage medium includes: a U disk, a removable hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and the like, which can store program codes. .

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Signal Processing (AREA)
  • Computational Linguistics (AREA)
  • Health & Medical Sciences (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Human Computer Interaction (AREA)
  • Acoustics & Sound (AREA)
  • Multimedia (AREA)
  • Spectroscopy & Molecular Physics (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Compression, Expansion, Code Conversion, And Decoders (AREA)

Abstract

本发明实施例提供了用于信号编码和解码的方法和设备。所述用于信号编码的方法包括：根据输入信号得到频域信号；按照预定分配规则将预定比特分配给所述频域信号；在有比特分配的频域信号的最高频率大于预定值的情况下，调整频域信号的比特分配；根据频域信号的比特分配对频域信号进行编码。

Description

信号编码和解码的方法和设备
本.申请要求于 2012 年 03 月 29 日提交中国专利局、 申请号为 201210087702.9, 发明名称为 "信号编码和解码的方法和设备" 的中国专利申 请的优先权， 其全部内容通过引用结合在本申请中。 技术领域
本发明实施例涉及领域通信领域， 并且更具体地， 涉及一种信号编码和解 码的方法和设备。 背景技术
在诸如移动通信和光纤通信之类的通信领域，在发送端釆用编码技术， 来 对要传送的信号进行压缩以提高传输效率，并且在接收端釆用对应的解码技术 来恢复所传送的信号。 根据信号的特征、 传输条件等， 可以对信号进行时域编 码和 /或频域编码。 根据一定的规则为时域信号或频域信号分配不同的编码比 输效率， 希望用尽量少的编码比特来表征要传送的信号。 因此， 需要合理地分 配编码比特， 从而在接收端通过解码来失真尽量少地恢复输出信号。
在现有的用于音频信号的编码器中，在较低码率时， 一般对语音能有较好 的编解码效果，但对音乐，编解码效果比较差。为了提升低码率时音乐的质量， 利用部分比特通过时域编码方法对输入信号进行编码； 并且，根据所述输入信 码。 在利用剩余的比特对所述频域信号进行编码时， 通常没考虑信号特性， 用 统一的方式对频域信号进行比特分配， 这导致对部分频域信号的编码效果不 佳。在现有的用于音频信号的解码器中， 简单地利用与编码技术对应的解码技 术恢复频域信号，对经未解码出的频域信号填充噪声， 然后进行频域逆变换和 时域综合处理而得到输出信号。 所述噪声填充在某些信号中引入了额外噪声， 降低了输出信号的质量。
因此，现有的在频域编码算法中进行统一的比特分配方案， 导致了对某些 信号编码效果不佳；而在现有的频域解码算法中的上述噪声填充处理降低了输 出信号的质量。 发明内容
本发明实施例提供了一种对信号进行编码和解码的方法和设备，其在编码 时可以对频域信号的比特分配进行优化， 以用相同的比特达到更好的编码效 果，在解码时可以利用频域解码所解码出的信息做指导， 进行频域激励信号的 扩展， 使输出信号达到更好的效果。
一方面， 提供了一种用于信号编码的方法， 所述方法包括： 根据输入信号 得到频域信号； 按照预定分配规则将预定比特分配给所述频域信号； 在有比特 分配的频域信号的最高频率大于预定值的情况下， 调整频域信号的比特分配； 根据频域信号的比特分配对频域信号进行编码。
另一方面， 提供了一种用于信号解码的方法， 所述方法包括： 从接收的比 特流中获得解码出的频域信号；在所述解码出的频域信号满足预定条件的情况 下，根据该解码出的频域信号来预测未解码出的频域信号； 根据解码出的频域 信号和预测的未解码出的频域信号来获得最终输出的时域信号。
又一方面， 提供了一种用于信号编码的设备， 所述设备包括： 频域变换单 元， 根据输入信号得到频域信号； 比特分配单元， 按照预定分配规则将预定比 特分配给所述频域信号； 比特调整单元，在有比特分配的频域信号的最高频率 大于等于预定值时， 调整所述频域信号的比特分配； 频域编码单元， 根据所述 频域信号的比特分配对频域信号进行编码。
又一方面， 提供了一种用于信号解码的设备， 所述设备包括： 解码单元， 从接收的比特流中获得解码出的频域信号； 扩频单元， 用于预测未解码出的频 域信号，在所述解码出的频域信号满足预定条件的情况下，根据该解码出的频 域信号来预测未解码出的频域信号； 输出单元，根据解码出的频域信号和预测 的频域信号来获得最终输出的时域信号。
在本发明实施例的上述技术方案中，在编码时通过根据有比特分配的频域 信号的最高频率来调整频域信号的比特分配，在利用相同数目的比特进行频域 编码的情况下达到了更好的编码效果； 在解码时以该解码出的频域信号为指 导， 来设置未解码出的频域信号， 使输出信号达到更好的效果。 附图说明
为了更清楚地说明本发明实施例的技术方案，下面将对实施例或现有技术 描述中所需要使用的附图作简单地介绍，显而易见地， 下面描述中的附图仅仅 是本发明的一些实施例，对于本领域普通技术人员来讲， 在不付出创造性劳动 的前提下， 还可以根据这些附图获得其他的附图。
图 1图示了根据本发明实施例的对信号进行编码的方法； 图 3图示了根据本发明实施例的对信号进行解码的方法；
图 4图示了在时频联合解码方法中从接收的比特流中获得解码出的频域信 号的方法；
图 5图示了根据本发明的编码设备和 /或解码设备的示例性实现；
图 6图示了根据本发明实施例的对信号进行编码的编码设备； 图 8图示了根据本发明实施例的对信号进行解码的设备；
图 9图示了在时频联合解码中的解码单元的框图。 具体实施方式
下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清 楚、 完整地描述， 显然， 所描述的实施例是本发明一部分实施例， 而不是全部 的实施例。基于本发明中的实施例， 本领域普通技术人员在没有作出创造性劳 动前提下所获得的所有其他实施例， 都属于本发明保护的范围。
本发明的编码技术方案和解码技术方案，可以应用于各种通信系统中的发 送和接收，所述通信系统例如： GSM,码分多址（ CDMA, Code Division Multiple Access ) 系统， 宽带码分多址（WCDMA, Wideband Code Division Multiple Access Wireless ), 通用分组无线业务（ GPRS, General Packet Radio Service ), 长期演进 ( LTE, Long Term Evolution )等。
编码技术方案和解码技术方案， 广泛应用于各种电子设备中， 例如： 移动 电话， 无线装置， 个人数据助理（PDA ), 手持式或便携式计算机， GPS接收 机 /导航器， 照相机， 音频 /视频播放器， 摄像机， 录像机， 监控设备等。 通常， 这类电子设备中包括音频编码器或音频解码器，音频编码器或者解码器可以直 接由数字电路或芯片例如 DSP ( digital signal processor )实现， 或者由软件代码 驱动处理器执行软件代码中的流程而实现。
作为示例， 在一种音频编码技术方案中， 首先将音频时域信号变换为频域 信号，再将编码比特分配给音频频域信号进行编码，将编码后的信号通过通信 系统传输给解码端， 解码端对编码后的信号解码恢复。
图 1图示了根据本发明实施例的对信号进行编码的方法 100。 如图 1所示， 该方法包括：
110: 根据输入信号得到频域信号。 所述输入信号可以为各种类型， 诸如 图像信号、 数据信号、 音频信号、 视频信号、 文本信号等。 可釆用诸如快速傅 立叶变换（ FFT, Fast Fourier Transform )、 离散余弦变换 ( DCT, Discrete Cosine Transform )等算法对输入信号进行频域变换而得到频域信号。 输入信号的类 型和频域变换算法不构成对本发明的限制。
120: 按照预定分配规则将预定比特分配给所述频域信号。 所述预定比特 tot— bit是用于对频域信号进行频域编码所要使用的比特。 所述预定分配规则例 如可以是将预定比特中的较多比特分配给频域信号中的低频段信号、并将预定 比特中的剩余比特分配给除了所述低频段信号之外的能量较大的频段。可以针 对所有低频段相同地、或者根据低频段信号的能量分布， 将所述较多比特分配 在低频段信号中。 为低频段信号分配较多比特的原因在于，在诸如语音频信号 中， 低频段信号通常包含人耳更敏感的信息。
下面以音频信号的频域编码为例进行说明。在进行频域编码时，通常按照 频率等间隔地将频域信号划分为子带， 或者根据频域系数来划分子带， 例如每 16个频域系数一个子带。 例如， 针对 20ms—帧的宽带信号， 将 0 ~ 4kHz频率范 围的 160个系数划分为 10个子带，其中 0 ~ 2kHz频率范围内有 5个子带， 2 ~ 4kHz 频率范围内有 5个子带。 然后， 针对每个子带进行比特分配。 对 0 ~ 2kHz频率 范围内的低频频域信号分配数目为 IF— bit的较多比特， 将预定比特 tot— bit减去 IF— bit得到剩余比特 rest— bit, 并根据 2 ~ 4kHz频率范围内的每个子带的包络大 小将剩余比特 rest— bit分配给 2 ~ 4kHz频率范围内的子带， 每个子带 5比特。 根 据 rest— bits和每个子带的包络的大小来确定有比特分配的子带的数目和有比特 分配的最高频带的子带 last— bin, 同时将不能被 5整除的余数平均分配给 0~2kHz 范围内的每个子带。
130: 在有比特分配的频域信号的最高频率大于预定值的情况下， 调整频 域信号的比特分配。 所述预定值 B可以根据经验值设定； 一个实施例中， 可以 根据所述预定比特的比特数目 tot— bit和频域信号的分辨率（例如， 0~8kHz带宽 范围内有 320个频域系数） 来确定预定值 B。 在固定带宽的情况下， 预定比特 的比特数目 tot— bit越多， 预定值 B越高； 预定比特的比特数目 tot— bit固定时， 频 域信号的分辨率越高， 预定值 B越高。 在带宽固定、 频域信号的分辨率也是固 定的情况下， 预定值 B可以仅根据预定比特的比特数目 tot— bit来确定， 预定比 特的比特数目 tot— bit越多，预定值 B越高。 该预定值 B是一预先设置的上限频率 值。 例如， 根据经验估计， 对输入信号进行频域变换后， 频率大于该预定值的 频域信号通常分配不到比特。 因此， 在具体实践中， 可以将预定值 B设置为比 频域信号的最高频率值低一定频率的频率值， 例如， 设置为 2.9kHz、 3.2kHz, 3.5kHz等。 在其他实施例中， 预定值 B也可以根据诸如帧长、 所釆用的变换方 法、 或变换窗长之类的其他因素来确定。
在将频域信号划分为子带以进行编码的情况中， 所述预定值 B可以为在 0 ~ 8kHz频率范围内的 20个子带的索引号， 而有比特分配的频域信号的最高频 率也可以用该最高频率所位于的子带的索引号来表示。 例如， 对釆样率为 16kHz的宽带信号，帧长为 20ms,如果传输速率为 6.8kbps,根据子带的总数（20 个）和待分配的预定比特的数目 （ 6.8kbps X 20ms = 136 bits ), 将 B设置为 6; 在传输速率为 7.6kbps时， 根据子带的总数（20个）和待分配的预定比特的数 目 （ 7.6kbps X 20ms = 152 bits ), 将 B设置为 8。 总之， 所述预定值 B和有比特分 配的频域信号的最高频率不局限于为频率数值，还可以为子带的索引号。在阅 读了本发明实施例的公开之后 ,工程技术人员根据实践条件知道如何确定有比 特分配的频域信号的最高频率是否大于预定值。
下面描述频域信号的比特分配的调整。根据信号的类型或频域特征等， 可 以减少频域信号中对解码端的输出贡献较少的部分的比特，并相应地增加有比 特分配的最高频率和其附近的频域信号的比特分配。也就是说， 所述调整频域 信号的比特分配可包括：减少所述频域信号中被分配较多比特的频段所分配的 比特数， 并增加有比特分配的最高频率和其附近的频域信号所分配的比特数。 对于音频信号， 所述被分配较多比特的频段为例如 0 ~ 2kHz的低频段。 下面举 例说明对频域信号的比特分配的调整。
调整示例 1 : 有比特分配的最高频率为 4kHz, 如果 2kHz~4kHz范围内的某 子带比特分配为 0, 给此频带分配 5比特， 直到 2kHz~4kHz范围的所有子带都分 配到比特数，假定 2~4kHz范围内额外增加的比特数为 Nblt。此时，需要从 0~2kHz 范围的子带减少 Nblt个比特。 所釆用的算法例如为： 从 0~2kHz范围内的所有子 带（5个子带）中每个子带减少 1个比特； 然后减去一个最高频的子带； 从剩余 的 4个子带中每个子带再减少 1比特，再减少一个次高频的子带， ... ...以此类推， 直到所减少的比特数等于 Nblt时停止。
调整示例 2: 对 2kHz~4kHz范围内的已分配比特的所有子带增加 J个比特， 假如 2~4kHz范围有比特分配的子带数为 K, 则此时 2~4kHz范围内额外增加的 比特数 Nblt = J X K, 需要从 0~2kHz范围的子带减少 Nblt = J X K个比特。 例如可 釆用的算法为： 从 0~2kHz范围内的所有子带（5个子带） 中平均每个子带减少 Nblt/5个比特。
调整示例 3: 对 2kHz~4kHz范围内没有分配比特数的子带各分 5比特， 然后 对 2~4kHz范围内的所有子带再增加 J个比特，假如 2~4kHz范围有比特分配的子 带数为 K, 则此时 2~4kHz范围内额外增加的比特数 Nblt = 5 ( 5-K ) + 5 J, 需要从 0~2kHz范围的子带减少 Nblt个比特。 所釆用的算法可以为： 调整示例 1 中的算法和调整示例 2中的算法中的任何一个。
此外，如果所述有比特分配的频域信号的最高频率小于预定值， 则保持在
140: 根据调整后的比特分配对频域信号进行编码。 在实践中， 可以根据 需要釆用任何频域编码方法。 所选择的频域编码方法不构成对本发明的限制。
通过以上对信号进行编码的方法，通过根据有比特分配的频域信号的最高 频率来调整频域信号的比特分配，在利用相同数目的比特进行频域编码的情况 下达到了更好的编码效果。 图 2图示了根据本发明实施例的时频联合编码方法 200。在图 2中， 220、 230、 240分别于与图 1中的 120、 130、 140相同。 图 2与图 1的不同之处在于增加了步 骤 250、 260, 并且用 211和 212代替了图 1中的 110。 下文中将描述图 2与图 1的不 同之处， 并对于相同之处不进行重复。
211 : 通过对输入信号进行时域分析得到第一时域信号和第二时域信号。 例如， 对输入信号进行线性预测编码（LPC, linear predictive coding )分析和 处理得到线语频率 （Line Spectral Frequency , LSF ) 参数和导抗语频率 ( Immittance Spectral Frequency, ISF )参数之一 , 还得到残差信号 res和自适 应码书贡献 exc_pit。所述 LSF参数或 ISF参数用于表示在 LPC分析中所使用的系 数（即 LPC系数） 的频域特性。 将所述残差信号 res和自适应码书贡献 exc_pit 包括在第一时域信号中， 自适应码书贡献 exc_pit被包括在第二时域信号中。
212: 通过对第一时域信号进行频域变换和处理得到频域信号。作为示例， 对第一时域信号中的残差信号 res和自适应码书贡献 exc_pit分别进行频域变换， 然后根据频域的残差信号 f— res和频域的自适应码书贡献 f— exc_pit的相关性来 判断自适应码书贡献对输出信号是否有贡献。如果自适应码书贡献对输出信号 有贡献， 则从频域的残差信号 f— res减去频域的自适应码书贡献 f— exc_pit而得到 频域的差值信号 f— diff, 并将该差值信号 f— diff作为所述频域信号。 如果自适应 码书贡献对输出信号没有贡献， 则将频域的残差信号 f— res直接作为差值信号 f— diff, 即频域信号。
在得到频域信号之后， 利用与图 1中的 120、 130、 140相同的 220、 230、 240 对频域信号进行编码， 而得到编码后的频域信号。
250: 对第二时域信号进行编码。 作为示例， 与频域信号的编码同时地执 行 260。可以利用任意时域编码方法（诸如预测编码、脉冲编码调制（ Pulse Code Modulation, PCM )编码等）对所述时域信号进行编码， 所釆用的时域编码方 法不构成对本发明的限制。在自适应码书贡献对输出信号有贡献时， 需要在解 码端得到自适应码书贡献，所以对该第二时域信号中的自适应码书贡献 exc_pit 进行编码， 以作为比特流传送到接收端。 然而， 如果自适应码书贡献对输出信 号没有贡献， 即解码端的输出不需要自适应码书贡献， 则不需要该部分的时域 编码，从而提高了编码效率。 自适应码书贡献对输出信号有贡献意味着在解码 端仅仅根据编码后的频域信号不能得到高质量的输出信号。
260: 将编码后的频域信号和编码后的第二时域信号复用为比特流。
需要说明的是，要进行频域编码的频域信号除了包括差值信号 f— diff之夕卜， 还可以包括其它信号，诸如指明自适应码书贡献对输出信号是否有贡献的标记 flag。 同样地， 要进行时域编码的第二时域信号除了可包括自适应码书贡献 exc_pit, 还可以包括解码需要的其它信息。
在上面结合图 2描述的时频联合编码中， 通过根据有比特分配的频域信号 的最高频率来调整频域信号的比特分配， 并且与时域编码相结合，从而达到了 更好的编码效果。
图 3图示了根据本发明实施例的对信号进行解码的方法 300。 该方法 300包 括：
310: 从接收的比特流中获得解码出的频域信号。 在仅釆用频域编码的情 况，通过釆用与频域编码方法对应的频域解码方法来从接收的比特流中获得解 码出的频域信号。对于时频联合编码的情况，通过如下操作来从接收的比特流 中获得解码出的频域信号：对比特流中的频域信息进行频域解码得到第一频域 信号；根据第一频域信号确定在比特流中是否存在对输出信号有贡献的时域编 码信号； 当确定在比特流中存在对输出信号有贡献的时域编码信号时，对时域 编码信号进行时域解码和频域变换而得到第二频域信号，并将第一频域信号和 第二频域信号合成而获得所述解码出的频域信号， 这将在下面结合图 4进一步 详细描述。
320: 在解码出的频域信号满足预定条件的情况下， 根据该解码出的频域 信号来预测未解码出的频域信号。作为示例， 所述解码出的频域信号满足预定 条件包括： 该解码出的频域信号的最高频率大于预定值、和该解码出的频域信 号包括频域变换后的对输出信号有贡献的时域编码信号中的至少一个。 要注 意，在实践中， 可以先适用该解码出的频域信号包括频域变换后的对输出信号 有贡献的时域编码信号的判断条件，然后适用该解码出的频域信号的最高频率 大于预定值的判断条件， 或者釆取相反的先后的顺序，还可以仅仅釆用两者之 如前面结合图 1的 130所描述的，所述预定值是根据频域编码所使用的预定 比特的数目 tot— bit和频域信号的分辨率来确定的。 根据实践需要， 可以将预定 值设置为比频域信号的最高频率值低一定频率的频率值。在将频域信号划分为 子带的情况中， 所述预定值可以为子带的索引号， 此时有比特分配的频域信号 的最高频率也用该最高频域所位于的子带的索引号表示。解码端的预定值的取 值可以与编码端的预定值的取值相同、 也可以不同。
在时频联合编码的情况中， 在 310中对比特流进行解码而得到解码出的频 域信号中可能包括频域变换后的对输出信号有贡献的时域编码信号 ,该频域变 换后的对输出信号有贡献的时域编码信号例如是通过对比特流中包含的时域 编码信息进行时域解码和频域变换而获得的信号， 诸如为自适应码书贡献。根 据被编码信号的类型的不同、和在编码时所釆用的时域分析方法不是 LPC分析 时，该频域变换后的对输出信号有贡献的时域编码信号还可以是除了自适应码 书贡献之外的其它信号。
在所述解码出的频域信号包括自适应码书贡献的情况中，可以根据前述的 自适应码书贡献对输出信号是否有贡献的标记 flag来获知所述解码出的频域信 号是否包括频域变换后的对输出信号有贡献的时域编码信号。在解码出的频域 信号中包括频域变换后的对输出信号有贡献的时域编码信号，这说明仅靠频域 解码难以得到高质量输出， 而且才艮据语音频信号的特性， 此时简单地将未解码 出的频域信号设置为噪声会使输出的信号质量恶化，从而需要预测未解码出的 频域信号。
作为根据该解码出的频域信号来预测未解码出的频域信号的示例，可以从 该解码出的频域信号的最高频率开始向低频率选择的一段频带的频域信号，并 根据所选择的频域信号来预测未解码出的频域信号。 例如， 对帧长为 20ms, 釆样率为 12.8kHz的信号，频域系数为 256个，带宽为 6.4kHz,在 7.6kbps码率时， 每 16个系数一个子带， 共有 16个子带， 预定值设为 10 ( 4kHz ), 当该解码出的 频域信号的最高频带大于 10时， 则 4~6.4kHz范围未解码出的频域系数通过 1.6~4kHz范围内解码出的频域系数预测得到。作为预测的实现的示例， 可通过 对所选择的频域信号进行归一化处理、 包络处理等来预测未解码出的频域信 号。 所述归一化处理、 包络处理的实现是本领域技术人员已知的手段， 这里不 进行详细描述。 此外， 根据输出信号的类型， 本领域的技术人员可能选择其它 的方式来预测未解码出的频域信号，例如还可以根据已解码出的频域信号中固 定频段的频域信号来预测未解码出的频域信号。
需要说明的是， 在根据解码出的频域信号得到未解码出的频域信号之后， 为了防止所预测的频域信号中有能量过大的频点， 可以通过来自编码端的 ISF 参数或 LSF参数来对所预测的未解码出的频域系数进行修正。 例如， 通过 LSF 参数或 ISF参数估计共振峰位置； 在每个估计的共振峰位置， 对幅值较大的频 域系数进行缩放。作为示例， 当在该共振峰位置附近的预测的频域系数的幅值 大于一阔值（该阔值可以根据编码端的时域分析的特征来设置）时， 减小在该 共振峰位置附近的预测的频域系数的幅值。
此外，在频域信号不满足所述预定条件时， 则利用噪声来预测该未解码出 的频域信号。
330: 根据解码出的频域信号和预测的频域信号来获得最终输出的时域信 号。在通过解码得到了解码出的频域信号，并预测了未解码出的频域信号之后， 从而获得整个频带内的频域信号， 则通过进行诸如快速傅立叶逆变换（IFFT, Inverse Fast Fourier Transform )之类的频域逆变换等处理来得到时域的输出信 号。 作为示例， 对于时频联合编码的情况， 对所述 ISF参数或 LSF参数进行变 换得到 LPC系数， 利用该 LPC系数对频域逆变换之后得到的信号进行时域综 合， 来得到最终输出的时域信号。 在实践中， 本领域的工程技术人员知道如何 根据频域信号得到时域的输出信号的方案， 这里不进行详细描述。
在上面结合图 3描述的根据本发明实施例的对信号进行解码的方法中， 通 过以该解码出的频域信号为指导， 来设置未解码出的频域信号，使输出信号达 到更好的效果。 据本发明实施例的解码方法在时频联合解码方案中的应用。在时频联合解码方 案中， 除了从接收的比特流中获得解码出的频域信号 （310 ) 的步骤之外， 随 后的操作与结合图 3的描述的 320、 330相同。 因此， 下面仅描述在时频联合解 码方法中如何得到解码出的频域信号。
图 4图示了在时频联合解码方法中从接收的比特流中获得解码出的频域信 号的方法 410。 该方法 410包括： 411 : 将比特流解复用为第一组比特和第二组比特。 在接收端解码时， 当 接收到了比特流后，釆用与图 2的 260中的复用技术对应的解复用技术将比特流 解复用为第一组比特和第二组比特。该第一组比特包括要进行下述的频域解码 的频域信息， 而该第二组比特包括要进行下述的时域解码的、对输出信号有贡 献的时域编码信号。
对于音频信号的时域联合解码， 该第一组比特例如包括差值信号 f— diff、 指明自适应码书贡献对输出信号是否有贡献的标记 flag等。 该第二组比特例如 在自适应码书贡献对输出信号有贡献时包括自适应码书贡献。要注意， 该第一 组比特和第二组比特与信号的编码对应地还可以包括其它的信号。
412: 对第一组比特进行频域解码得到第一频域信号， 并根据第一频域信 号确定在比特流中是否存在对输出信号有贡献的时域编码信号。通过与编码端 的频域编码方法对应的解码方法对对第一组比特进行解码 ,以得到第一频域信 号。 该第一频域信号例如包括解码后的差值信号 f— diff、 和指明自适应码书贡 献对输出信号是否有贡献的标记 flag。
413: 对第二组比特进行时域解码。 通过与编码端的时域编码方法对应的 解码方法对对第二组比特进行解码， 以得到解码后的时域信号。 具体地， 当确 定在比特流中存在对输出信号有贡献的时域编码信号时，对第二组比特中的时 域编码信号进行时域解码。
414: 对解码后的时域信号中的自适应码书贡献进行频域变换得到第二频 域信号。
415: 将第一频域信号和第二频域信号合成获得解码出的频域信号。 作为 示例， 当自适应码书贡献对输出信号有贡献时，通过相加第一频域信号中的差 值信号 f— diff和第二频域信号中的自适应码书贡献来合成所述频域信号。 当自 适应码书贡献对输出信号没有贡献时， 直接将第一频域信号中的差值信号 f— diff作为所述频域信号。
在获得解码出的频域信号之后， 釆用与图 3的 320、 330相同的步骤来获得 最终输出的时域信号。
与上述方法实施例相关联， 本发明还提供一种编码设备和一种解码设备， 该编码设备或解码设备可以位于终端设备， 网络设备， 或测试设备中。 所述编 码设备或解码设备可以由硬件电路来实现， 或者由软件配合硬件来实现。
图 5图示了根据本发明的编码设备和 /或解码设备的示例性实现。 如图 5所 示， 由一个处理器 510经由输入 /输出接口 520来调用编码设备或解码设备 530 , 在内存 540的协助下实现音频信号的编码或解码处理。 该编码设备或解码设备 530可以执行上述方法实施例中的各种方法和流程。
图 6图示了根据本发明实施例的对信号进行编码的编码设备 600。该编码设 备 600包括：频域变换单元 610,根据输入信号得到频域信号；比特分配单元 620, 按照预定分配规则将预定比特分配给所述频域信号； 比特调整单元 630, 在有 比特分配的频域信号的最高频率大于等于预定值时，调整所述频域信号的比特 分配； 频域编码单元 640, 根据调整后的比特分配对频域信号进行编码。
频域变换单元 610可根据输入信号得到频域信号。 该输入信号可以为诸如 图像信号、 数据信号、 音频信号、 视频信号、 文本信号等的各种类型的信号。 可釆用诸如 FFT、 DCT等算法对输入信号进行频域变换而得到频域信号。 输入 信号的类型和频域变换算法不构成对本发明的限制。
比特分配单元 620可按照预定分配规则将预定比特 tot— bit分配给所述频域 信号。 所述 tot— bit是用于对频域信号进行编码所要使用的比特数目。 所述预定 分配规则例如可以是将所述预定比特中的较多比特分配给频域信号中的低频 段信号，并将预定比特中的剩余比特分配给除了所述低频段信号之外的能量较 大的频段。 对于低频段的频域信号的分配， 可以针对所有低频段相同地、 或者 根据低频段信号的能量分布 ,将所述较多比特分配在低频段信号中。 为低频段 信号分配较多比特的原因在于，诸如语音之类的音频信号在频域中主要集中在 低频范围内， 为其分配较多比特能提高频域编码的效率。
作为示例， 在对音频信号进行频域编码的示例性情况中， 如前面结合图 1 的 120所描述的， 将 0 ~ 4kHz频率范围的频域信号划分为 10个子带， 其中 0 ~ 2kHz频率范围内有 5个子带， 2 ~ 4kHz频率范围内有 5个子带。 然后， 针对每个 子带进行比特分配。 为 0 ~ 2kHz频率范围内的低频频域信号分配数目为 IF— bit 的较多比特。 根据 2 ~ 4kHz频率范围内的每个子带的包络将剩余比特 rest— bit ( tot— bit减去 IF— bit )分配给 2 ~ 4kHz频率范围内的子带。具体地，根据 rest— bits 高频带的子带 last— bin, 同时将不能被 5整除的余数平均分配给 0~2kHz范围内的 每个子带。
比特调整单元 630可以在有比特分配的频域信号的最高频率大于等于预定 值 B时， 调整所述频域信号的比特分配。 所述预定值 B是根据所述预定比特的 比特数目 tot— bit和频域信号的分辨率（例如， 4kHz )来确定的。 该预定值是一 预先设置的上限频率值。 在具体实践中， 预定值 B可以为比频域信号的最高频 率值（例如， 4kHz )低一定频率的频率值， 例如， 为 2.9kHz、 3.2kHz, 3.5kHz 等。 如前所述， 在将频域信号划分为子带以进行编码的情况中， 所述预定值 B 可以为在 0 ~ 4kHz频率范围内的 10个子带的索引号 （例如， 7或 8 ), 此时， 有 比特分配的频域信号的最高频率也要用该最高频率所位于的子带的索引号 iridex来表示。
如果所述有比特分配的频域信号的最高频率（例如， index = 7 )小于预定 述预定比特的分配。
比特调整单元 630可在所述最高频率大于等于预定值时， 调整所述比特分 配单元 620按照预定分配规则进行的频域信号的比特分配。 根据所述输入信号 的类型或所述频域信号的频域特征等，可以减少频域信号中对解码端的输出贡 献较少的部分，并相应地增加有比特分配的最高频率和其附近的频域信号的比 特分配。 作为示例， 所述比特调整单元 630可减少所述频域信号中被分配较多 比特的频段所分配的比特数，并增加有比特分配的最高频率和其附近的频域信 号所分配的比特数。 对于音频信号， 所述被分配较多比特的频段为例如 0 ~ 2kHz的低频段。
关于调整频域信号的比特分配的实现， 可以参照上面描述的调整示例 1 ~ 3 , 这里不再详细说明。
所述频域编码单元 640根据调整后的比特分配对频域信号进行编码。 对频 域信号进行编码的方法例如可以为变换编码、 子带编码等。 此外， 在所述最高 频率小于预定值时， 所述比特调整单元 630不调整频域信号的比特分配。 此时， 码。
在以上对信号进行编码的设备 600中， 通过根据有比特分配的频域信号的 最高频率来调整频域信号的比特分配， 达到了更好的编码效果。 合编码中的应用为例进行示例性说明。 该时频联合编码设备 700包括： 时域分析单元 711 , 通过对输入信号进行时 域分析得到第一时域信号和第二时域信号； 频域变换单元 712, 通过对第一时 域信号进行频域变换和处理得到频域信号； 比特分配单元 720, 按照预定分配 规则将预定比特分配给所述频域信号； 比特调整单元 730, 在有比特分配的频 域信号的最高频率大于等于预定值时， 调整所述频域信号的比特分配； 频域编 码单元 740 , 根据调整后的比特分配对频域信号进行编码； 时域编码单元 750, 对第二时域信号进行编码； 比特复用单元 760 , 将编码后的频域信号和编码后 的第二时域信号复用为比特流。
在图 7中， 比特分配单元 720、 比特调整单元 730、 频域编码单元 740分别于 与图 6中的比特分配单元 620、 比特调整单元 630、 频域编码单元 640相同。 图 7 与图 6的不同之处在于增加了时域编码单元 750、 比特复用单元 760 , 并且用时 域分析单元 711和频域变换单元 712代替了图 6中的频域变换单元 610。下文中将 描述图 7与图 6的不同之处， 并对于相同之处不进行重复。
时域分析单元 711通过对输入信号进行时域分析得到第一时域信号和第二 时域信号。 例如， 对输入信号进行 LPC分析和处理得到 ISF参数（或 LSF参数）、 残差信号 res和自适应码书贡献 exc_pit。 将所述残差信号 res和自适应码书贡献 exc_pit作为第一时域信号， 将自适应码书贡献 exc_pit作为第二时域信号。
频域变换单元 712可通过对第一时域信号进行频域变换和处理得到频域信 号。作为示例，对第一时域信号中的残差信号 res和自适应码书贡献 exc_pit分别 进行频域变换， 并然后根据频域的残差信号 f— res和频域的自适应码书贡献 f— exc_pit的相关性来判断自适应码书贡献对输出信号是否有贡献。 如果自适应 码书贡献对输出信号有贡献，则从频域的残差信号 f— res减去频域的自适应码书 贡献 f— exc_pit而得到频域的差值信号 f— diff, 并将该差值信号 f— diff包括在所述 频域信号。 如果自适应码书贡献对输出信号没有贡献， 则将频域的残差信号 f— res直接作为差值信号 f— diff, 以作为频域信号传送。 该频域信号除了包括差 值信号 f— diff之外， 还可以包括其它信号， 例如指明自适应码书贡献对输出信 号是否有贡献的标记 flag。
在得到频域信号之后， 利用与图 7中的比特分配单元 720、 比特调整单元
730、 频域编码单元 740对频域信号进行编码， 而得到编码后的频域信号。
时域编码单元 750可以对第二时域信号进行编码。可以使用诸如预测编码、 脉冲编码调制等时域编码方法对所述时域信号进行编码。在自适应码书贡献对 输出信号有贡献时， 需要在解码端得到自适应码书贡献， 所以对该第二时域信 号中的自适应码书贡献 exc_pit进行编码， 以传送到接收端。 然而， 如果自适应 码书贡献对输出信号没有贡献， 则不需要对自适应码书贡献进行编码和传输 , 从而提高了编码效率。 比特复用单元 760可将编码后的频域信号和编码后的第 二时域信号复用为比特流。
在上面结合图 7描述的时频联合编码设备中， 通过根据有比特分配的频域 信号的最高频率来调整频域信号的比特分配， 并且与时域编码相结合，从而达 到了更好的编码效果。
图 8图示了根据本发明实施例的对信号进行解码的解码设备 800。该解码设 备 800包括： 解码单元 810, 从接收的比特流中获得解码出的频域信号； 扩频单 元 820 , 用于预测未解码出的频域信号， 在所述解码出的频域信号满足预定条 件的情况下，根据该解码出的频域信号来预测未解码出的频域信号； 输出单元 830, 根据解码出的频域信号和预测的频域信号来获得最终输出的时域信号。
解码单元 810可从接收的比特流中获得解码出的频域信号。 在仅釆用频域 编码的情况，通过釆用与频域编码方法对应的频域解码方法来从接收的比特流 中获得解码出的频域信号。 对于时频联合编码的情况， 解码单元 810可通过如 下操作从接收的比特流中获得解码出的频域信号：对比特流中的频域信息进行 频域解码得到第一频域信号；根据第一频域信号确定在比特流中是否存在对输 出信号有贡献的时域编码信号；当确定在比特流中存在对输出信号有贡献的时 域编码信号时， 对时域编码信号进行时域解码和频域变换而得到第二频域信 号， 并将第一频域信号和第二频域信号合成而获得所述解码出的频域信号， 这 将在下面结合图 9详细描述。
扩频单元 820可用于预测未解码出的频域信号。 在所述解码出的频域信号 满足预定条件的情况下， 扩频单元 820可根据解码出的频域信号来预测未解码 出的频域信号。 作为示例， 所述解码出的频域信号满足预定条件包括： 该解码 出的频域信号的最高频率大于预定值、和该解码出的频域信号包括频域变换后 的对输出信号有贡献的时域编码信号中的至少一个。在实践中， 可根据需要进 行选择。 tot— bit和频域信号的分辨率来确定的。 根据实践需要， 可以将预定值设置为比 频域信号的最高频率值低一定频率的频率值。在将频域信号划分为子带的情况 中， 所述预定值可以为子带的索引号， 此时有比特分配的频域信号的最高频率 也用该最高频域所位于的子带的索引号表示。
在使用时频联合解码技术的情况中， 在解码单元 810对比特流进行解码得 到的解码出的频域信号中可能包括对比特流中包含的时域信息进行时域解码 和频域变换而获得的信号， 其例如为自适应码书贡献。可以根据前述的自适应 码书贡献对输出信号是否有贡献的标记 flag来获知所述频域信号是否包括频域 变换后的对输出信号有贡献的时域编码信号。 根据被编码信号的类型的不同、 和在编码时所釆用的时域分析方法不是 LPC分析时，该频域变换后的对输出信 号有贡献的时域编码信号还可以是其它信号。
在解码出的频域信号中包括对比特流中包含的时域信息进行时域解码和 频域变换而获得的信号， 这说明未解码出的频域信号中包括对输出有用的信 息，从而需要预测未解码出的频域信号， 简单地将未解码出的频域信号设置为 噪声会使输出的信号质量恶化。
此外， 在该解码出的频域信号不满足预定条件时， 所述扩频单元 820可以 将该未解码出的频域信号设置为噪声。
作为根据解码出的频域信号来预测未解码出的频域信号的示例，所述扩频 单元 820可以从该解码出的频域信号的最高频率开始向低频率选择的一段频带 的频域信号， 并如上所述对所选择的频域信号进行处理， 以根据所选择的频域 信号来预测未解码出的频域信号。此外，还可以釆用其它的方式来预测未解码 出的频域信号，例如还可以根据已解码出的频域信号中固定频段的频域信号来 预测未解码出的频域信号。
输出单元 830可根据解码出的频域信号和预测的频域信号来获得最终输出 的时域信号。在预测了未解码出的频域信号之后， 获得了整个频带内的频域信 号，则通过使用与编码时釆用的频域变换的逆变换对整个带宽的频域信号进行 频域逆变换， 从而得到时域的输出信号。 如前所述， 该输出单元可以通过使用 根据 ISF参数（或 LSF参数 )得到的 LPC系数对频域逆变换之后的信号进行时域 综合， 来得到最终输出的时域信号以输出。
对于解码设备 800中的各个单元的更详细操作，可以参照前面结合图 3所描 述的各个步骤。
在上面结合图 8描述的根据本发明实施例的对信号进行解码的解码设备 800中， 通过以所解码出的频域信号为指导， 来设置未解码出的频域信号， 使 输出信号达到更好的效果。 发明实施例的解码设备在时频联合解码方案中的应用。 在时频联合解码方案 中， 除了解码单元 810的操作之外， 其它构成单元的操作与扩频单元 820、 输出 单元 830的操作相同。 因此， 下面仅描述解码单元 810在时频联合解码方法中的 具体实现。
图 9图示了在时频联合解码中的解码单元 910的框图。该解码单元 910包括： 解复用单元 911 , 将比特流解复用为第一组比特和第二组比特； 频域解码单元 912, 对第一组比特进行频域解码得到第一频域信号， 并根据第一频域信号确 定在比特流中是否存在对输出信号有贡献的时域编码信号；时域解码单元 913 , 当确定在比特流中存在对输出信号有贡献的时域编码信号时，对第二组比特中 的进行时域解码； 频域变换单元 914 , 对解码后的时域信号进行频域变换得到 第二频域信号； 合成单元 915 , 将第一频域信号和第二频域信号合成获得解码 出的频域信号。
为描述的方便和简洁， 对于所述解复用单元 911、 频域解码单元 912、 时域 解码单元 913、频域变换单元 914、合成单元 915的具体操作，请参见图 4中的 411、 412、 413、 414、 和 415 , 在此不再赘述。 本领域普通技术人员可以意识到，结合本文中所公开的实施例描述的各示 例的单元及算法步骤， 能够以电子硬件、或者计算机软件和电子硬件的结合来 实现。这些功能究竟以硬件还是软件方式来执行，取决于技术方案的特定应用 和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现 所描述的功能， 但是这种实现不应认为超出本发明的范围。
在本申请所提供的几个实施例中， 应该理解到， 所揭露的设备和方法， 可 以通过其它的方式实现。 例如， 以上所描述的设备实施例仅仅是示意性的， 例 如， 所述单元的划分， 仅仅为一种逻辑功能划分， 实际实现时可以有另外的划 分方式， 例如多个单元或组件可以结合或者可以集成到另一个系统， 或一些特 征可以忽略， 或不执行。 以位于一个地方， 或者也可以分布到多个网络单元上。可以根据实际的需要选 择其中的部分或者全部单元来实现本实施例方案的目的。
另外， 在本发明各个实施例中的各功能单元可以集成在一个处理单元中， 也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个单元 中。
所述功能如果以软件功能单元的形式实现并作为独立的产品销售或使用 时， 可以存储在一个计算机可读取存储介质中。 基于这样的理解， 本发明的技 术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的部分可以 以软件产品的形式体现出来， 该计算机软件产品存储在一个存储介质中， 包括 若干指令用以使得一台计算机设备（可以是个人计算机， 服务器， 或者网络设 备等）执行本发明各个实施例所述方法的全部或部分步骤。 而前述的存储介质 包括： U盘、 移动硬盘、 只读存储器（ROM, Read-Only Memory )、 随机存取 存储器（RAM, Random Access Memory ), 磁碟或者光盘等各种可以存储程序 代码的介质。
以上所述，仅为本发明的具体实施方式，但本发明的保护范围并不局限于 此，任何熟悉本技术领域的技术人员在本发明揭露的技术范围内， 可轻易想到 变化或替换， 都应涵盖在本发明的保护范围之内。 因此， 本发明的保护范围应 所述以权利要求的保护范围为准。

Claims

权 利 要 求
1. 一种用于信号编码的方法， 其特征在于， 所述方法包括：
根据输入信号得到频域信号；
按照预定分配规则将预定比特分配给所述频域信号；
在有比特分配的频域信号的最高频率大于预定值的情况下，调整频域信号 的比特分配；
根据频域信号的比特分配对频域信号进行编码。
2. 根据权利要求 1所述的方法， 其特征在于， 所述调整频域信号的比特 分配包括： 减少所述频域信号中被分配较多比特的频段所分配的比特数， 并增 加有比特分配的最高频率和其附近的频域信号所分配的比特数。
3. 根据权利要求 1所述的方法， 其特征在于， 其中所述预定值是根据所 述预定比特的数目和频域信号的分辨率来确定的。
4. 根据权利要求 1所述的方法， 其特征在于， 其中， 所述预定分配规则 是： 将所述预定比特中的较多比特分配给频域信号中的低频段信号， 并将预定 比特中的剩余比特分配给除了所述低频段信号之外的能量较大的频段。
5. 根据权利要求 1所述的方法， 其特征在于， 还包括：
如果所述最高频率小于预定值，则保持按照预定分配规则进行的所述频域 信号的比特分配。
6. 一种用于信号解码的方法， 其特征在于， 所述方法包括：
从接收的比特流中获得解码出的频域信号；
在所述解码出的频域信号满足预定条件的情况下，根据该解码出的频域信 号来预测未解码出的频域信号；
根据解码出的频域信号和预测的频域信号来获得最终输出的时域信号。
7. 根据权利要求 6的方法， 其特征在于， 所述解码出的频域信号满足预 定条件包括满足下列条件中的至少一个：
该解码出的频域信号的最高频率大于预定值； 和
该解码出的频域信号包括频域变换后的对输出信号有贡献的时域编码信 号。
8. 根据权利要求 6或 7的方法， 其特征在于， 其中， 所述从接收的比特 流中获得解码出的频域信号的步骤包括： 对比特流中的频域信息进行频域解码得到第一频域信号；
根据第一频域信号确定在比特流中是否存在对输出信号有贡献的时域编 码信号；
当确定在比特流中存在对输出信号有贡献的时域编码信号时，对时域编码 信号进行时域解码和频域变换而得到第二频域信号，并将第一频域信号和第二 频域信号合成而获得所述解码出的频域信号。
9. 根据权利要求 7的方法， 其特征在于， 其中， 所述预定值是根据频域 编码所使用的预定比特的数目和解码出的频域信号的分辨率来确定的。
10. 根据权利要求 6的方法， 其特征在于， 其中所述根据该解码出的频域 信号来预测未解码出的频域信号包括：从解码出的频域信号中选择某一频段的 频域信号， 并根据所选择的频域信号来预测未解码出的频域信号。
11. 根据权利要求 6的方法， 其特征在于， 其中所述预测未解码出的频域 信号包括： 根据线谱频率 LSF或导抗谱频率 ISF估计频域信号的共振峰位置， 当在共振峰位置附近的预测的频域系数的幅值大于一阔值时，减小在该共振峰 位置附近的预测的频域系数的幅值。
12. 根据权利要求 6的方法， 其特征在于， 其中， 在该解码出的频域信号 不满足预定条件的情况下， 利用噪声来预测未解码出的频域信号。
13. 一种用于信号编码的设备， 其特征在于， 所述设备包括：
频域变换单元， 根据输入信号得到频域信号；
比特分配单元， 按照预定分配规则将预定比特分配给所述频域信号； 比特调整单元， 在有比特分配的频域信号的最高频率大于等于预定值时， 调整所述频域信号的比特分配；
频域编码单元， 根据所述频域信号的比特分配对频域信号进行编码。
14. 根据权利要求 13所述的设备， 其特征在于， 所述比特调整单元通过 减少所述频域信号中被分配较多比特的频段所分配的比特数、并增加有比特分 配的最高频率和其附近的频域信号所分配的比特数，来调整频域信号的比特分 配。
15. 根据权利要求 13所述的设备， 其特征在于， 其中所述预定值是根据 所述预定比特的数目和频域信号的分辨率来确定的。
16. 根据权利要求 13所述的设备， 其特征在于， 其中， 所述预定分配规 则为将所述预定比特中的较多比特分配给频域信号中的低频段信号，并将预定 比特中的剩余比特分配给除了所述低频段信号之外的能量较大的频段。
17. 根据权利要求 13所述的设备， 其特征在于， 其中， 在所述最高频率 小于预定值时， 比特调整单元不调整所述频域信号的比特分配， 所述频域编码
18. 一种用于信号解码的设备， 其特征在于， 所述设备包括：
解码单元， 从接收的比特流中获得解码出的频域信号；
扩频单元， 用于预测未解码出的频域信号，在所述解码出的频域信号满足 预定条件的情况下， 根据该解码出的频域信号来预测未解码出的频域信号； 输出单元，根据解码出的频域信号和预测的频域信号来获得最终输出的时 域信号。
19. 根据权利要求 18的设备， 其特征在于， 其中， 所述解码出的频域信 号满足预定条件包括： 该解码出的频域信号的最高频率大于预定值、和该解码 出的频域信号包括频域变换后的对输出信号有贡献的时域编码信号中的至少 一个。
20. 根据权利要求 18或 19的设备， 其特征在于， 其中， 所述解码单元通 过如下操作从接收的比特流中获得解码出的频域信号：
对比特流中的频域信息进行频域解码得到第一频域信号；
根据第一频域信号确定在比特流中是否存在对输出信号有贡献的时域编 码信号；
当确定在比特流中存在对输出信号有贡献的时域编码信号时，对时域编码 信号进行时域解码和频域变换而得到第二频域信号，并将第一频域信号和第二 频域信号合成而获得所述解码出的频域信号。
21. 根据权利要求 19的设备， 其特征在于， 其中， 所述预定值是根据频 域编码所使用的预定比特的数目和该解码出的频域信号的分辨率来确定的。
22. 根据权利要求 18的设备， 其特征在于， 其中， 在所述解码出的频域 信号满足预定条件的情况下，所述扩频单元从解码出的频域信号中选择某一频 段的频域信号， 并根据所选择的频域信号来预测未解码出的频域信号。
23. 根据权利要求 18 的方法， 其特征在于， 其中， 所述扩频单元在根据 该解码出的频域信号来预测未解码出的频域信号之后， 根据线谱频率 LSF或 导抗谱频率 ISF估计频域信号的共振峰位置，并且当在共振峰位置附近的预测 的频域系数的幅值大于一阔值时，减小在该共振峰位置附近的预测的频域系数 的幅值。
24. 根据权利要求 18的设备， 其特征在于， 其中， 在所述该解码出的频 域信号不满足预定条件时， 所述扩频单元利用噪声来预测未解码出的频域信 号。
25. 计算机可读存储介质， 其特征在于， 该存储介质中存储有计算机指 令， 所述指令在处理器驱动下能够执行权利要求 1至 12任一方法中的步骤。
PCT/CN2012/075924 2012-03-29 2012-05-23 信号编码和解码的方法和设备 WO2013143221A1 (zh)

Priority Applications (17)

Application Number Priority Date Filing Date Title
ES12873219.5T ES2655832T3 (es) 2012-03-29 2012-05-23 Método y dispositivo de codificación y descodificación de señal
EP19191869.7A EP3664085B1 (en) 2012-03-29 2012-05-23 Signal coding and decoding methods and devices
MX2014011605A MX339652B (es) 2012-03-29 2012-05-23 Metodos y dispositivos de codificacion y descodificacion de señal.
JP2015502053A JP6006400B2 (ja) 2012-03-29 2012-05-23 信号符号化の方法および装置
EP17160983.7A EP3249645B1 (en) 2012-03-29 2012-05-23 Signal coding and decoding methods and devices
CA2866202A CA2866202C (en) 2012-03-29 2012-05-23 Signal coding and decoding methods and devices
EP12873219.5A EP2809009B1 (en) 2012-03-29 2012-05-23 Signal encoding and decoding method and device
MYPI2014002473A MY189975A (en) 2012-05-21 2012-05-23 Signal coding and decoding methods and devices
SG11201405216SA SG11201405216SA (en) 2012-03-29 2012-05-23 Signal coding and decoding methods and devices
KR1020147026193A KR101621641B1 (ko) 2012-03-29 2012-05-23 신호 코딩 및 디코딩 방법 및 장치
BR112014023577A BR112014023577B8 (pt) 2012-03-29 2012-05-23 Método e dispositivo de codificação de sinal de áudio e método e dispositivo de decodificação de sinal de áudio.
RU2014142255/08A RU2592412C2 (ru) 2012-03-29 2012-05-23 Способы и устройства кодирования и декодирования сигналов
ZA2014/06424A ZA201406424B (en) 2012-03-29 2014-09-01 Signal coding and decoding methods and devices.
US14/496,986 US9537694B2 (en) 2012-03-29 2014-09-25 Signal coding and decoding methods and devices
US15/358,649 US9786293B2 (en) 2012-03-29 2016-11-22 Signal coding and decoding methods and devices
US15/684,079 US9899033B2 (en) 2012-03-29 2017-08-23 Signal coding and decoding methods and devices
US15/864,147 US10600430B2 (en) 2012-03-29 2018-01-08 Signal decoding method, audio signal decoder and non-transitory computer-readable medium

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201210087702.9A CN103368682B (zh) 2012-03-29 2012-03-29 信号编码和解码的方法和设备
CN201210087702.9 2012-03-29

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US14/496,986 Continuation US9537694B2 (en) 2012-03-29 2014-09-25 Signal coding and decoding methods and devices

Publications (1)

Publication Number Publication Date
WO2013143221A1 true WO2013143221A1 (zh) 2013-10-03

Family

ID=49258139

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2012/075924 WO2013143221A1 (zh) 2012-03-29 2012-05-23 信号编码和解码的方法和设备

Country Status (15)

Country Link
US (4) US9537694B2 (zh)
EP (3) EP2809009B1 (zh)
JP (2) JP6006400B2 (zh)
KR (1) KR101621641B1 (zh)
CN (3) CN103368682B (zh)
BR (1) BR112014023577B8 (zh)
CA (2) CA2866202C (zh)
ES (3) ES2655832T3 (zh)
MX (1) MX339652B (zh)
PL (1) PL3664085T3 (zh)
PT (1) PT3249645T (zh)
RU (1) RU2592412C2 (zh)
SG (2) SG11201405216SA (zh)
WO (1) WO2013143221A1 (zh)
ZA (1) ZA201406424B (zh)

Families Citing this family (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10923204B2 (en) 2010-08-20 2021-02-16 Attopsemi Technology Co., Ltd Fully testible OTP memory
US10916317B2 (en) 2010-08-20 2021-02-09 Attopsemi Technology Co., Ltd Programmable resistance memory on thin film transistor technology
US9818478B2 (en) 2012-12-07 2017-11-14 Attopsemi Technology Co., Ltd Programmable resistive device and memory using diode as selector
US10586832B2 (en) 2011-02-14 2020-03-10 Attopsemi Technology Co., Ltd One-time programmable devices using gate-all-around structures
CN103368682B (zh) 2012-03-29 2016-12-07 华为技术有限公司 信号编码和解码的方法和设备
CN105374363B (zh) * 2014-08-25 2019-06-04 广东美的集团芜湖制冷设备有限公司 音频信号编码方法和系统
US11062786B2 (en) 2017-04-14 2021-07-13 Attopsemi Technology Co., Ltd One-time programmable memories with low power read operation and novel sensing scheme
US10726914B2 (en) 2017-04-14 2020-07-28 Attopsemi Technology Co. Ltd Programmable resistive memories with low power read operation and novel sensing scheme
US11615859B2 (en) 2017-04-14 2023-03-28 Attopsemi Technology Co., Ltd One-time programmable memories with ultra-low power read operation and novel sensing scheme
US10535413B2 (en) 2017-04-14 2020-01-14 Attopsemi Technology Co., Ltd Low power read operation for programmable resistive memories
JP6934648B2 (ja) * 2017-07-03 2021-09-15 東日本旅客鉄道株式会社 トロリ線曲げ工具
RU2744485C1 (ru) * 2017-10-27 2021-03-10 Фраунхофер-Гезелльшафт Цур Фердерунг Дер Ангевандтен Форшунг Е.Ф. Ослабление шума в декодере
US10770160B2 (en) 2017-11-30 2020-09-08 Attopsemi Technology Co., Ltd Programmable resistive memory formed by bit slices from a standard cell library
CN118053437A (zh) * 2022-11-17 2024-05-17 抖音视界有限公司 音频编码方法、解码方法、装置、设备及存储介质

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101055720A (zh) * 2005-12-07 2007-10-17 三星电子株式会社 对音频信号编码和解码的方法和设备
CN101494054A (zh) * 2009-02-09 2009-07-29 深圳华为通信技术有限公司 一种音频码率控制方法及系统
CN101523485A (zh) * 2006-10-02 2009-09-02 卡西欧计算机株式会社 音频编码装置、音频解码装置、音频编码方法、音频解码方法和信息记录介质

Family Cites Families (47)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5517511A (en) * 1992-11-30 1996-05-14 Digital Voice Systems, Inc. Digital transmission of acoustic signals over a noisy communication channel
JP3131542B2 (ja) * 1993-11-25 2001-02-05 シャープ株式会社 符号化復号化装置
KR970011727B1 (en) 1994-11-09 1997-07-14 Daewoo Electronics Co Ltd Apparatus for encoding of the audio signal
JP3521596B2 (ja) 1996-01-30 2004-04-19 ソニー株式会社 信号符号化方法
JP3519859B2 (ja) * 1996-03-26 2004-04-19 三菱電機株式会社 符号器及び復号器
FI970553A (fi) * 1997-02-07 1998-08-08 Nokia Mobile Phones Ltd Audiokoodausmenetelmä ja -laite
US6356211B1 (en) 1997-05-13 2002-03-12 Sony Corporation Encoding method and apparatus and recording medium
KR100335609B1 (ko) * 1997-11-20 2002-10-04 삼성전자 주식회사 비트율조절이가능한오디오부호화/복호화방법및장치
KR100304092B1 (ko) * 1998-03-11 2001-09-26 마츠시타 덴끼 산교 가부시키가이샤 오디오 신호 부호화 장치, 오디오 신호 복호화 장치 및 오디오 신호 부호화/복호화 장치
US6226616B1 (en) 1999-06-21 2001-05-01 Digital Theater Systems, Inc. Sound quality of established low bit-rate audio coding systems without loss of decoder compatibility
US6621935B1 (en) * 1999-12-03 2003-09-16 Microsoft Corporation System and method for robust image representation over error-prone channels
JP2001255882A (ja) 2000-03-09 2001-09-21 Sony Corp 音声信号処理装置及びその信号処理方法
SE0001926D0 (sv) * 2000-05-23 2000-05-23 Lars Liljeryd Improved spectral translation/folding in the subband domain
ATE320651T1 (de) 2001-05-08 2006-04-15 Koninkl Philips Electronics Nv Kodieren eines audiosignals
US7333929B1 (en) * 2001-09-13 2008-02-19 Chmounk Dmitri V Modular scalable compressed audio data stream
CN1127054C (zh) * 2001-11-02 2003-11-05 北京阜国数字技术有限公司 用于知觉音频编码的信号处理方法
US20030187663A1 (en) * 2002-03-28 2003-10-02 Truman Michael Mead Broadband frequency translation for high frequency regeneration
DE10328777A1 (de) * 2003-06-25 2005-01-27 Coding Technologies Ab Vorrichtung und Verfahren zum Codieren eines Audiosignals und Vorrichtung und Verfahren zum Decodieren eines codierten Audiosignals
US7349842B2 (en) * 2003-09-29 2008-03-25 Sony Corporation Rate-distortion control scheme in audio encoding
US7672838B1 (en) * 2003-12-01 2010-03-02 The Trustees Of Columbia University In The City Of New York Systems and methods for speech recognition using frequency domain linear prediction polynomials to form temporal and spectral envelopes from frequency domain representations of signals
US7586924B2 (en) * 2004-02-27 2009-09-08 Fraunhofer-Gesellschaft Zur Foerderung Der Angewandten Forschung E.V. Apparatus and method for coding an information signal into a data stream, converting the data stream and decoding the data stream
EP1873753A1 (en) * 2004-04-01 2008-01-02 Beijing Media Works Co., Ltd Enhanced audio encoding/decoding device and method
KR100723400B1 (ko) 2004-05-12 2007-05-30 삼성전자주식회사 복수의 룩업테이블을 이용한 디지털 신호 부호화 방법 및장치
JP4809234B2 (ja) * 2004-09-17 2011-11-09 パナソニック株式会社 オーディオ符号化装置、復号化装置、方法、及びプログラム
KR20070084002A (ko) 2004-11-05 2007-08-24 마츠시타 덴끼 산교 가부시키가이샤 스케일러블 복호화 장치 및 스케일러블 부호화 장치
PL1866912T3 (pl) * 2005-03-30 2011-03-31 Koninl Philips Electronics Nv Kodowanie wielokanałowego sygnału audio
EP1949369B1 (en) * 2005-10-12 2012-09-26 Samsung Electronics Co., Ltd. Method and apparatus for encoding/decoding audio data and extension data
JP2007149151A (ja) * 2005-11-24 2007-06-14 Funai Electric Co Ltd 光ディスク再生装置、音声信号出力装置及びａｖシステム
BRPI0707135A2 (pt) 2006-01-18 2011-04-19 Lg Electronics Inc. aparelho e método para codificação e decodificação de sinal
JP2007264154A (ja) * 2006-03-28 2007-10-11 Sony Corp オーディオ信号符号化方法、オーディオ信号符号化方法のプログラム、オーディオ信号符号化方法のプログラムを記録した記録媒体及びオーディオ信号符号化装置
ES2312142T3 (es) * 2006-04-24 2009-02-16 Nero Ag Aparato avanzado para codificar datos de audio digitales.
KR20070115637A (ko) * 2006-06-03 2007-12-06 삼성전자주식회사 대역폭 확장 부호화 및 복호화 방법 및 장치
KR101565919B1 (ko) * 2006-11-17 2015-11-05 삼성전자주식회사 고주파수 신호 부호화 및 복호화 방법 및 장치
RU2406165C2 (ru) * 2007-02-14 2010-12-10 ЭлДжи ЭЛЕКТРОНИКС ИНК. Способы и устройства для кодирования и декодирования объектно-базированных аудиосигналов
EP2571024B1 (en) * 2007-08-27 2014-10-22 Telefonaktiebolaget L M Ericsson AB (Publ) Adaptive transition frequency between noise fill and bandwidth extension
KR100970446B1 (ko) 2007-11-21 2010-07-16 한국전자통신연구원 주파수 확장을 위한 가변 잡음레벨 결정 장치 및 그 방법
CN101436407B (zh) 2008-12-22 2011-08-24 西安电子科技大学 音频编解码方法
UA99878C2 (ru) * 2009-01-16 2012-10-10 Долби Интернешнл Аб Гармоническое преобразование, усовершенствованное перекрестным произведением
EP2273493B1 (en) * 2009-06-29 2012-12-19 Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V. Bandwidth extension encoding and decoding
CN101958119B (zh) * 2009-07-16 2012-02-29 中兴通讯股份有限公司 一种改进的离散余弦变换域音频丢帧补偿器和补偿方法
CA2778368C (en) 2009-10-20 2016-01-26 Fraunhofer-Gesellschaft Zur Foerderung Der Angewandten Forschung E.V. Audio encoder, audio decoder, method for encoding an audio information, method for decoding an audio information and computer program using an iterative interval size reduction
CN102081927B (zh) * 2009-11-27 2012-07-18 中兴通讯股份有限公司 一种可分层音频编码、解码方法及系统
MY153845A (en) * 2010-01-12 2015-03-31 Fraunhofer Ges Forschung Audio encoder, audio decoder, method for encoding and audio information, method for decoding an audio information and computer program using a hash table describing both significant state values and interval boundaries
US8751225B2 (en) * 2010-05-12 2014-06-10 Electronics And Telecommunications Research Institute Apparatus and method for coding signal in a communication system
US9047875B2 (en) * 2010-07-19 2015-06-02 Futurewei Technologies, Inc. Spectrum flatness control for bandwidth extension
CN102208188B (zh) 2011-07-13 2013-04-17 华为技术有限公司 音频信号编解码方法和设备
CN103368682B (zh) 2012-03-29 2016-12-07 华为技术有限公司 信号编码和解码的方法和设备

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101055720A (zh) * 2005-12-07 2007-10-17 三星电子株式会社 对音频信号编码和解码的方法和设备
CN101523485A (zh) * 2006-10-02 2009-09-02 卡西欧计算机株式会社 音频编码装置、音频解码装置、音频编码方法、音频解码方法和信息记录介质
CN101494054A (zh) * 2009-02-09 2009-07-29 深圳华为通信技术有限公司 一种音频码率控制方法及系统

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP2809009A4 *

Also Published As

Publication number Publication date
KR101621641B1 (ko) 2016-05-16
EP3249645B1 (en) 2019-11-13
JP6006400B2 (ja) 2016-10-12
RU2592412C2 (ru) 2016-07-20
US9899033B2 (en) 2018-02-20
CN110706715B (zh) 2022-05-24
PL3664085T3 (pl) 2022-11-14
JP2017016141A (ja) 2017-01-19
ES2655832T3 (es) 2018-02-21
CA2866202C (en) 2018-04-03
EP2809009B1 (en) 2017-11-29
MX2014011605A (es) 2014-10-17
BR112014023577A8 (pt) 2021-05-25
RU2014142255A (ru) 2016-05-20
BR112014023577B1 (pt) 2021-08-24
ES2770831T3 (es) 2020-07-03
EP2809009A1 (en) 2014-12-03
CA2866202A1 (en) 2013-10-03
PT3249645T (pt) 2020-02-05
EP3664085A1 (en) 2020-06-10
CN110706715A (zh) 2020-01-17
US20170352359A1 (en) 2017-12-07
KR20140130182A (ko) 2014-11-07
JP6323881B2 (ja) 2018-05-16
CA2994705C (en) 2020-06-30
BR112014023577A2 (pt) 2017-06-20
CN106409299A (zh) 2017-02-15
US10600430B2 (en) 2020-03-24
MX339652B (es) 2016-06-02
BR112014023577B8 (pt) 2021-10-05
ZA201406424B (en) 2019-04-24
US20180130481A1 (en) 2018-05-10
US9537694B2 (en) 2017-01-03
US20170076733A1 (en) 2017-03-16
EP2809009A4 (en) 2015-07-15
US20150010021A1 (en) 2015-01-08
EP3664085B1 (en) 2022-07-06
ES2927563T3 (es) 2022-11-08
US9786293B2 (en) 2017-10-10
CN103368682B (zh) 2016-12-07
CN103368682A (zh) 2013-10-23
EP3249645A1 (en) 2017-11-29
CN106409299B (zh) 2019-11-05
CA2994705A1 (en) 2013-10-03
SG11201405216SA (en) 2014-11-27
JP2015517119A (ja) 2015-06-18
SG10201701275XA (en) 2017-03-30

Similar Documents

Publication Publication Date Title
US10600430B2 (en) Signal decoding method, audio signal decoder and non-transitory computer-readable medium
JP7010885B2 (ja) 音声または音響符号化装置、音声または音響復号装置、音声または音響符号化方法及び音声または音響復号方法
JP6574820B2 (ja) 高周波帯域信号を予測するための方法、符号化デバイス、および復号デバイス
CN105874534B (zh) 编码装置、解码装置、编码方法、解码方法及程序

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 12873219

Country of ref document: EP

Kind code of ref document: A1

REEP Request for entry into the european phase

Ref document number: 2012873219

Country of ref document: EP

WWE Wipo information: entry into national phase

Ref document number: 2012873219

Country of ref document: EP

ENP Entry into the national phase

Ref document number: 2866202

Country of ref document: CA

ENP Entry into the national phase

Ref document number: 20147026193

Country of ref document: KR

Kind code of ref document: A

ENP Entry into the national phase

Ref document number: 2015502053

Country of ref document: JP

Kind code of ref document: A

WWE Wipo information: entry into national phase

Ref document number: MX/A/2014/011605

Country of ref document: MX

NENP Non-entry into the national phase

Ref country code: DE

REG Reference to national code

Ref country code: BR

Ref legal event code: B01A

Ref document number: 112014023577

Country of ref document: BR

WWE Wipo information: entry into national phase

Ref document number: IDP00201406385

Country of ref document: ID

ENP Entry into the national phase

Ref document number: 2014142255

Country of ref document: RU

Kind code of ref document: A

ENP Entry into the national phase

Ref document number: 112014023577

Country of ref document: BR

Kind code of ref document: A2

Effective date: 20140923