WO2009121298A1 - 编码方法、解码方法、系统及装置 - Google Patents

编码方法、解码方法、系统及装置 Download PDF

Info

Publication number
WO2009121298A1
WO2009121298A1 PCT/CN2009/071123 CN2009071123W WO2009121298A1 WO 2009121298 A1 WO2009121298 A1 WO 2009121298A1 CN 2009071123 W CN2009071123 W CN 2009071123W WO 2009121298 A1 WO2009121298 A1 WO 2009121298A1
Authority
WO
WIPO (PCT)
Prior art keywords
harmonic
harmonic structure
interval
signal
starting position
Prior art date
Application number
PCT/CN2009/071123
Other languages
English (en)
French (fr)
Inventor
张德明
张琦
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
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Publication of WO2009121298A1 publication Critical patent/WO2009121298A1/zh

Links

Classifications

    • GPHYSICS
    • G10MUSICAL INSTRUMENTS; ACOUSTICS
    • G10LSPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L19/00Speech 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
    • G10L19/04Speech 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
    • G10L19/08Determination or coding of the excitation function; Determination or coding of the long-term prediction parameters
    • G10L19/093Determination or coding of the excitation function; Determination or coding of the long-term prediction parameters using sinusoidal excitation models
    • GPHYSICS
    • G10MUSICAL INSTRUMENTS; ACOUSTICS
    • G10LSPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L19/00Speech 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
    • G10L19/04Speech 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
    • G10L19/08Determination or coding of the excitation function; Determination or coding of the long-term prediction parameters
    • G10L19/09Long term prediction, i.e. removing periodical redundancies, e.g. by using adaptive codebook or pitch predictor

Definitions

  • the present invention relates to the field of communications technologies, and in particular, to an encoding method, a decoding method, a system, and a device.
  • the speech codec has gradually expanded to broadband and ultra-wideband.
  • the International Telecommunication Union ITU
  • broadband voice codec standards such as G.722, G.722.1, G.722.2, and G.729.1.
  • 3GPP has introduced the AMR-WB broadband voice codec standard, 3GPP2.
  • VMR-WB broadband voice codec standard
  • the ITU recently proposed G.729.1 & G.EV.VBR combined with ultra-wideband, G.711 WB&G.722 combined with ultra-wideband.
  • the narrowband range is usually 0 ⁇ 4kHz
  • the wideband is usually in the 0 ⁇ 8kHz band
  • the ultra-wideband is usually in the 0 ⁇ 16kHz band.
  • the signal band can be divided into three parts, the narrowband signal part (0 ⁇ 4kHz), Wideband signal part (4 ⁇ 8kHz) and ultra-wideband signal (8 ⁇ 16kHz) part, but in the process of encoding and decoding audio signal, usually (0 ⁇ 3.5kHz) is called narrowband signal part, (4 ⁇ 7kHz) is called broadband
  • the signal part, (8 ⁇ 14kHz) is called the ultra-wideband part signal, and the signals in the 7 ⁇ 8kHz band are continuously put into the ultra-wideband part for processing together.
  • the narrowband signal part is also called the core layer, which is generally code-excited linear-precision (CELP) coding, while the wideband/ultra-wideband signal part uses transform coding, such as modified discrete cosine transform (MDCT, Modified Discrete Cosine Transform ), TCX and other technologies.
  • transform coding such as modified discrete cosine transform (MDCT, Modified Discrete Cosine Transform ), TCX and other technologies.
  • the number of bits that can be used for encoding the signal of the ultra-wideband signal is relatively small, and usually only the key parameters of the time domain envelope, the spectral envelope and the partial spectral coefficient of the ultra-wideband signal part can be encoded, and then The decoder performs bandwidth expansion through these parameters to reconstruct the signal of the ultra-wideband signal portion.
  • the technical problem to be solved by the embodiments of the present invention is to provide an encoding method, a decoding method, a system and a device, which can reconstruct a harmonic structure.
  • an embodiment of the present invention provides an encoding method, including: acquiring feature information of a harmonic structure by using a pitch information of a narrowband signal portion;
  • the feature information of the harmonic structure is quantized and transmitted to the decoding end.
  • a decoding method including:
  • a codec system including:
  • the encoding end is configured to obtain feature information of the harmonic structure through the pitch information of the narrowband signal part; quantize and encode the feature information of the harmonic structure, and send the information;
  • a decoding end characteristic information for the harmonic structure; reconstructing a harmonic structure according to characteristic information of the harmonic structure and a voiced gain of the narrowband signal portion.
  • an encoding end including:
  • An acquiring unit configured to acquire feature information of a harmonic structure by using a pitch information of a narrowband signal portion; and a coding unit, configured to quantize and encode feature information of the harmonic structure;
  • a sending unit configured to send, to the decoding end, feature information of the harmonic structure after the coding unit is quantized.
  • a decoding end including:
  • a receiving unit configured to receive feature information of a harmonic structure transmitted by the encoding end
  • a reconstruction unit configured to reconstruct the harmonic structure according to the feature information of the harmonic structure and the voiced gain of the narrowband signal portion.
  • the coding end is based on the narrowband signal part.
  • the pitch information acquires the characteristic information of the harmonic structure, and sends the characteristic information of the harmonic structure to the decoding end, so that the decoding end can reconstruct the harmonic structure according to the characteristic information of the harmonic structure, thereby avoiding the inability to recover the harmonic at the decoding end.
  • Wave structure damage to the final hearing effect.
  • Embodiment 1 is a flowchart of Embodiment 1 of an encoding method provided by the present invention
  • Embodiment 1 of a decoding method provided by the present invention is a flowchart of Embodiment 1 of a decoding method provided by the present invention
  • FIG. 3 is a flowchart of Embodiment 2 of an encoding method provided by the present invention.
  • Embodiment 4 is a flowchart of Embodiment 2 of a decoding method provided by the present invention.
  • FIG. 5 is a flowchart of Embodiment 3 of an encoding method provided by the present invention.
  • FIG. 6 is a diagram of an embodiment of a codec system provided by the present invention.
  • FIG. 7 is a diagram of an embodiment of a coding end provided by the present invention.
  • FIG. 8 is a diagram of an embodiment of a decoding end provided by the present invention.
  • Embodiments of the present invention provide an encoding method, a decoding method, a system, and a device, which can reconstruct a harmonic structure.
  • the MDCT coefficients of the specific position can be processed before the MDCT inverse transform is performed on the decoding side, and the inversely transformed MDCT is performed on the processed coefficients.
  • these specific positions are determined by the starting position and period.
  • characteristic information that characterizes part of the harmonic structure of the ultra-wideband signal such as the harmonic interval and the harmonic starting position
  • the UWB signal portion can be reconstructed well at the decoding end by using these characteristic information.
  • the harmonic structure achieves the purpose of bandwidth expansion and improves the overall auditory effect.
  • the pitch information of the narrowband signal part is obtained, and the voiced gain G is obtained; the ultra-wideband signal part of the input signal passes.
  • the MDCT transform a set of MDCT coefficients ⁇ y _ swb(0 y _swb(l ......, _w6(319) ⁇ ) is obtained.
  • the input signal has obvious harmonic structural characteristics, its MDCT coefficient will show obvious period. Sexuality, this period is the harmonic interval.
  • the pitch period of the narrowband signal part 7 ⁇ has been able to roughly describe the harmonic spacing of the ultra-wideband signal part, which can be simplified by the Average Magnitude Difference Function (AMDF).
  • AMDF Average Magnitude Difference Function
  • the method searches for a more accurate harmonic interval of the UWB signal in the vicinity of 7 ⁇ . After determining the partial harmonic interval ⁇ of the UWB signal, the harmonic starting position P of the MDCT coefficient of the UWB signal portion is searched in the range of [0 ⁇ ]. .
  • the encoding end acquires the feature information of the harmonic structure through the pitch information of the narrowband signal portion; the feature information of the harmonic structure is quantized and encoded, and is sent to the decoding end. After receiving the decoding end, the decoding end can decode according to the feature information of the harmonic structure, and reconstruct the harmonic structure.
  • the decoding end receives the characteristic information of the harmonic structure transmitted by the encoding end; and reconstructs the harmonic structure according to the characteristic information of the harmonic structure and the voiced gain of the narrowband signal portion.
  • the first embodiment of the encoding method provided by the present invention is as shown in FIG. 1:
  • Step 102 Obtain a harmonic interval according to the value interval of the harmonic interval and the MDCT coefficient.
  • the pitch period ⁇ of the narrowband signal part can be used directly. Harmonic interval A as part of the ultra-wideband signal.
  • Step 103 Obtain a harmonic structure starting position according to the position of the first peak of the harmonic interval and the spectral coefficient, that is, the position where the first harmonic of the ultra-wideband signal portion appears.
  • the [y - ) ⁇ is integrated in the interval [ ⁇ 2 - 1 to obtain the partial harmonic structure starting position P of the ultra-wideband signal.
  • Step 104 Perform quantization coding on the harmonic interval and the start position of the harmonic structure, and send the same.
  • N can take a value of 3
  • M can take a value of 7.
  • the encoding end obtains the harmonic interval, the harmonic structure starting position according to the pitch information and the spectral coefficient of the narrowband signal portion, and sends the harmonic interval and the harmonic structure starting position to the decoding end. So that the decoding end can reconstruct the harmonic structure according to the harmonic interval and the starting position of the harmonic structure.
  • the first embodiment of the coding method provided by the present invention is equally applicable to the narrowband signal portion and the wideband signal portion.
  • the first embodiment of the decoding method provided by the present invention is as shown in FIG. 2 when the encoding side uses the first embodiment of the encoding method provided by the present invention.
  • Step 201 Receive a harmonic interval and a starting position of the harmonic structure.
  • the decoding end decodes from the code stream to obtain the harmonic interval ⁇ , the harmonic structure starting position ⁇ , and the voiced gain G of the narrowband signal portion.
  • Step 202 Perform harmonic reconstruction according to the harmonic interval, the start position of the harmonic structure, and the voiced gain of the narrowband signal portion. Since the signals in the 7-8 kHz band are put into the ultra-wideband portion for processing in order to be continuous in the frequency band, the starting position of the partial harmonic structure of the ultra-wideband signal, that is, the harmonic structure above 7 kHz, needs to be calculated in this step. The starting position.
  • the starting position of the harmonic structure above 7 kHz is calculated. Since P represents the starting position of the harmonic structure in the signal above 8 kHz, the number of MDCT coefficients in the 1 kHz frequency range of 7 to 8 kHz is 40, so the starting position of the harmonic structure above 7 kHz can be expressed by the formula. for:
  • the ultra-wideband signal portion at 7 kHz to 14 kHz at intervals of ⁇ is used for harmonic reconstruction as follows:
  • the MDCT coefficient at the ⁇ '+" ⁇ position is filled by a number with a random phase controlled by the voiced gain G.
  • the specific filling method is: y-swb ( i ) at the position ⁇ '+" ⁇ ⁇
  • n (- 1) M *G* I*2
  • M is a random integer.
  • n is the sequential number of the MDCT coefficients, taking the first few MDCT coefficients, and n is a few.
  • Step 203 Perform decoding according to a normal step to obtain an ultra-wideband signal partial signal with enhanced harmonic structure.
  • the harmonic structure when the harmonic interval and the harmonic structure starting position transmitted by the encoding end are received, the harmonic structure can be reconstructed according to the harmonic interval and the starting position of the harmonic structure.
  • the first embodiment of the decoding method provided by the present invention is equally applicable to the narrowband signal portion and the wideband signal portion.
  • harmonic information may be generated according to the time domain signal of the UWB signal part and the pitch information of the narrowband signal part, and sent to the decoding end, so that the decoding end can reconstruct the harmonic structure.
  • the pitch information gain G of the narrow-band signal portion is obtained, and the time-domain signal of the ultra-wideband signal portion of the current frame input signal is ⁇ y_hi(0ly_hi(ll...,_y -/ '( 3 l 9 ) ⁇
  • the ultra-wideband signal part of the signal also exhibits significant periodicity, and the pitch period of this period and the narrow-band signal portion is very close. Therefore, the periodicity of the signal of the ultra-wideband signal portion can be approximated by the pitch period of the narrow-band signal portion. Through this period, the ultra-wideband signal portion can be approximately estimated.
  • the harmonic structure of the sub-signal is at the beginning of the frequency domain, but this estimate is relatively coarse.
  • the inaccurate starting position of the harmonic structure will lead to error accumulation during harmonic reconstruction at the decoding end, resulting in some
  • the frequency position of some harmonic structures should be neglected, which causes the reconstructed harmonic structure to deviate and impair the auditory experience. Therefore, based on the known pitch period of the narrowband signal portion 7 ⁇ , a more accurate signal period T can be found in the ultra-wideband signal portion by the AMDF method, and then this period is converted into the harmonic structure starting position information P in the MDCT coefficient. And pass this information to the decoder. Based on this information, the decoder can accurately establish harmonic structures in the MDCT coefficients for better harmonic reconstruction.
  • FIG. 3 The second embodiment of the encoding method provided by the present invention is shown in FIG. 3:
  • Step 301 Calculate a value interval of a signal period of the ultra-wideband signal by using pitch information of the narrowband signal portion.
  • Step 302 Obtain an ultra-wideband signal partial signal period according to a value interval of the signal period of the UWB signal portion and a time domain signal of the UWB signal portion.
  • n is the sequential number of the MDCT coefficients, taking the first few MDCT coefficients, n is a few; the maximum value of k is ⁇ + "', and the minimum value of k is "-".
  • Step 303 Obtain a harmonic structure starting position according to a partial signal period of the ultra-wideband signal.
  • Tx 25 shows the integer function.
  • Step 304 Perform quantization coding on the start position of the harmonic structure, and send.
  • the harmonic structure start position P is quantized and encoded by M bits, and this part of information is packaged and transmitted as a transmission code stream.
  • M can take a value of 7.
  • the encoding end obtains the starting position of the harmonic structure according to the pitch information of the narrowband signal part and the time domain signal of the ultra-wideband signal part, and sends the starting position of the harmonic structure to the decoding end to decode The end can reconstruct the harmonic structure according to the starting position of the harmonic structure.
  • the second embodiment of the coding method provided by the present invention is also applicable to the narrowband signal part. Part, broadband signal part.
  • the second embodiment of the decoding method provided by the present invention is as shown in FIG. 4:
  • Step 401 Receive a start position of a harmonic structure.
  • the decoding end decodes from the code stream to obtain a narrowband signal portion pitch period T. , harmonic structure starting position ⁇ , voiced gain G.
  • Step 402 Perform harmonic reconstruction according to a start position of the harmonic structure, a pitch period of the narrowband signal portion, and a voiced gain of the narrowband signal portion.
  • the partial harmonic interval of the ultra-wideband signal and the starting position of the harmonic structure need to be calculated in this step.
  • the harmonic interval of 7 kHz or more is calculated according to the pitch period 7 of the narrowband signal part:
  • the window reconstructs the harmonic structure according to the calculated harmonic interval ⁇ above 7 kHz, the harmonic structure starting position ⁇ , and the voiced gain G.
  • the ultra-wideband signal portion at 7 kHz to 14 kHz at intervals of ⁇ is harmonically reconstructed as follows:
  • the MDCT coefficient at the ⁇ '+" ⁇ position is filled by a number with a random phase controlled by the voiced gain G.
  • the specific filling method is: y-swb ( i ) at the position ⁇ '+" ⁇ ⁇
  • Step 403 Perform decoding according to normal steps to obtain an ultra-wideband signal partial signal with enhanced harmonic structure.
  • the harmonic structure when the harmonic interval and the harmonic structure starting position transmitted by the encoding end are received, the harmonic structure can be reconstructed according to the harmonic interval and the starting position of the harmonic structure.
  • the second embodiment of the decoding method provided by the present invention is equally applicable to the narrowband signal portion and the wideband signal portion.
  • the signal part time domain signal generates harmonic information and sends it to the decoding end so that the decoding end can reconstruct the harmonic structure.
  • the pitch information of the narrowband portion is obtained.
  • the ultra-wideband partial time domain signal of the current frame input signal is ⁇ y- hi (oiy_hi(il..., _y_ ( 3 l 9 ) ⁇ ), and the ultra-wideband portion of the previous frame signal stored in the encoding register
  • the time domain signal is ⁇ _ oW _ hi (° y_ ld_hi ⁇ ), , _ ⁇ / ⁇ / _ /z (319) ⁇ .
  • FIG. 5 The third embodiment of the encoding method provided by the present invention is shown in FIG. 5:
  • Step 501 Calculate a value interval of a signal period of the ultra-wideband signal by using pitch information of the narrowband signal portion.
  • Step 502 Obtain an ultra-wideband signal partial signal period according to the value interval of the partial signal period of the ultra-wideband signal, the current frame, and the time domain signal of the ultra-wideband signal part of the previous frame.
  • the ultra-wideband partial signal period ⁇ is found by the AMDF function.
  • the ultra-wideband partial signal period ⁇ can be expressed by the formula:
  • is the sequential number of the ultra-wideband partial time domain signal, taking the first ultra-wideband partial time domain signal, ⁇ is a few; the maximum value of k is the minimum value of k is ⁇ ".
  • y(n) ⁇ y_ old _ hi(0), y-old _ hi( ⁇ ... , y _old _hi(3 ⁇ 9), y _ hi(0), y _ hi(l),... , y_hi(3 ⁇ 9) ⁇
  • Step 503 Obtain a harmonic structure starting position according to a partial signal period of the ultra-wideband signal.
  • the harmonic structure starting position P can be expressed by the formula:
  • Step 504 Perform quantization coding on the start position of the harmonic structure, and send the same.
  • the M-bit is used to quantize the start position P of the harmonic structure, and this part of the information is transmitted.
  • the code stream is packaged for transmission.
  • M can take a value of 7.
  • the encoding end obtains the starting position of the harmonic structure according to the pitch information of the narrowband signal part, the current frame, and the time domain signal of the ultra-wideband signal part of the previous frame, and the starting position of the harmonic structure Sended to the decoder, so that the decoder can reconstruct the harmonic structure according to the starting position of the harmonic structure.
  • the third embodiment of the encoding method provided by the present invention is equally applicable to the narrowband signal portion and the wideband signal portion.
  • the decoding end can perform the harmonic reconstruction using the decoding method embodiment 2 provided by the present invention.
  • the feature information of the harmonic structure is obtained through the pitch information of the narrowband signal portion; the feature information of the harmonic structure is quantized and encoded, and sent to the decoding end.
  • the harmonic structure is reconstructed according to the characteristic information of the harmonic structure and the voiced gain of the narrowband signal portion. It will be understood by those skilled in the art that all or part of the steps of implementing the foregoing embodiments may be performed by a program to instruct related hardware, and the program may be stored in a computer readable storage medium. , including the following steps:
  • An encoding method including:
  • the feature information of the harmonic structure is quantized and transmitted to the decoding end.
  • a decoding method including:
  • the harmonic structure is reconstructed according to the characteristic information of the harmonic structure and the voiced gain of the narrowband signal portion.
  • the above-mentioned storage medium may be a read only memory, a magnetic disk or an optical disk or the like.
  • a codec system implementation provided by the present invention, as shown in FIG. 6, includes:
  • the encoding end 601 is configured to obtain feature information of the harmonic structure through the pitch information of the narrowband signal part; quantize and encode the feature information of the harmonic structure, and send the information;
  • the decoding end 602 is configured to receive feature information of the harmonic structure, and reconstruct a harmonic structure according to the feature information of the harmonic structure and the voiced gain of the narrowband signal portion.
  • the encoding end obtains the feature information of the harmonic structure according to the pitch information of the narrowband signal part, and sends the feature information of the harmonic structure to the decoding end, so that the decoding end can be based on the harmonic
  • the characteristic information of the wave structure reconstructs the harmonic structure.
  • An encoding end implementation provided by the present invention is as shown in FIG. 7, and includes:
  • the obtaining unit 710 is configured to obtain, by using the narrowband signal part pitch information, feature information of the harmonic structure;
  • a coding unit 720 configured to quantize and encode feature information of the harmonic structure
  • the transmitting unit 720 is configured to send, to the decoding end, the feature information of the quantized harmonic structure of the coding unit 720.
  • the acquiring unit 710 includes:
  • a first harmonic spacing unit 711 for obtaining a harmonic interval by a pitch period of the narrowband signal portion; a first starting position unit 712, a harmonic interval obtained according to the first harmonic spacing unit and a modified discrete cosine The position of the first peak in the transform coefficient is taken to obtain the starting position of the harmonic structure.
  • the first harmonic spacing unit 711 includes:
  • a second harmonic spacing unit 711a for using the pitch period as the harmonic interval; and/or a third harmonic spacing unit 711b for calculating the harmonic interval by a pitch period of the narrowband signal portion
  • the value interval is obtained according to the value interval of the harmonic interval and the modified discrete cosine transform coefficient.
  • the characteristic information of the harmonic structure includes: a start position of the harmonic structure
  • the obtaining unit 710 includes:
  • the value interval unit 713 of the signal period is used to calculate a value interval of the signal period by using the pitch information
  • a signal period unit 714 configured to obtain a signal period according to the value interval of the signal period and the time domain signal
  • the second starting position unit 715 is configured to obtain a harmonic structure starting position according to a signal period.
  • a decoding end implementation provided by the present invention, as shown in FIG. 8, includes:
  • the receiving unit 810 is configured to receive feature information of a harmonic structure sent by the encoding end.
  • a reconstruction unit 820 configured to: according to the feature information of the harmonic structure, the voiced sound of the narrowband signal portion Gain, reconstruct harmonic structure.
  • the characteristic information of the harmonic structure includes: a harmonic interval, a harmonic structure starting position; the reconstruction unit 820 includes:
  • a third starting position unit 821 configured to obtain a starting position of a part of the harmonic structure of the ultra-wideband signal according to the harmonic interval and the starting position of the harmonic structure;
  • the first reconstruction unit 822 is configured to reconstruct a harmonic structure according to a starting position of the partial harmonic structure of the ultra-wideband signal, the harmonic interval, and the voiced gain.
  • the characteristic information of the harmonic structure includes: a start position of the harmonic structure
  • the reconstruction unit 820 includes:
  • a fourth harmonic spacing unit 823 configured to obtain a partial harmonic interval of the ultra-wideband signal according to a pitch period of the narrowband signal portion
  • the second reconstruction unit 824 is configured to reconstruct the harmonic structure according to the partial harmonic interval of the ultra-wideband signal, the starting position of the harmonic structure, and the voiced gain.
  • the encoding end acquires the feature information of the harmonic structure according to the pitch information of the narrowband signal part, and sends the feature information of the harmonic structure to the decoding end.
  • the decoding end can reconstruct the harmonic structure according to the characteristic information of the harmonic structure, thereby avoiding the damage that cannot be recovered at the decoding end and causing the final auditory effect.

Landscapes

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

Description

编码方法、 解码方法、 系统及装置
本申请要求于 2008 年 4 月 3 日提交中国专利局、 申请号为 200810089863.5、 发明名称为 "编码方法、 解码方法、 系统及装置" 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
本发明涉及通信技术领域, 尤其涉及一种编码方法、 解码方法、 系统及装 置。
背景技术
随着承载技术的发展,人们越来越不满足于窄带语音编解码器的质量, 因 此语音编解码器已逐步向宽带、 超宽带扩展。 例如国际电信联盟 (ITU, International Telecommunication Union )推出了 G.722、 G.722.1、 G.722.2、 G.729.1 等宽带语音编解码标准, 3GPP推出了 AMR-WB这一宽带语音编解码标准, 3GPP2则推出了 VMR-WB。 此外 ITU最近又提出了 G.729.1&G.EV.VBR联合 超宽带, G.711 WB&G.722联合超宽带等。 其中, 窄带的范围通常为 0~4kHz 频带, 宽带的范围通常为 0~8kHz频带, 超宽带的范围通常为 0~16kHz频带。
这些标准都是从窄带扩展而来的,在这些标准中,根据信号通常釆用的釆 样率 8khz、 16kHz , 32kHz , 可将信号频带分为三个部分, 窄带信号部分 ( 0~4kHz )、 宽带信号部分(4~8kHz )和超宽带信号 (8~16kHz )部分, 但是 在音频信号的编解码过程中通常将( 0 ~ 3.5kHz )称为窄带信号部分, ( 4~7kHz ) 称为宽带信号部分, (8~14kHz)称为超宽带部分信号, 为了频带连续会将 7 ~ 8kHz频带内的信号放入超宽带部分进行一起处理。 其中, 窄带信号部分也被 称为核心层, 一般为码激励线性预测 (CELP, Code-Excited Linear-Prediction ) 编码,而宽带 /超宽带信号部分使用变换编码,如修正的离散余弦变换(MDCT, Modified Discrete Cosine Transform ), TCX等技术。
在超宽带信号部分能够用于超宽带信号部分信号进行编码的比特数比较 少, 通常只能对超宽带信号部分的时域包络、 频谱包络、 部分谱系数这些关键 参数进行编码, 然后在解码端通过这些参数进行带宽扩展, 重构超宽带信号部 分的信号。
在对现有技术的研究和实践过程中, 发明人发现现有技术存在以下问题: 在超宽带信号部分资源受限制的情况下,通常只能对超宽带信号部分的时 域包络、 频谱包络、 部分谱系数这些关键参数进行编码, 其他一些信息, 如表 征谐波结构特征的谐波间隔、 谐波起始位置等参数在编码过程中都已经丟失, 不能在解码端恢复出这些描述超宽带信号部分信号的谐波信息,会给最终的听 觉效果带来一定的损害。
发明内容
本发明实施例要解决的技术问题是提供一种编码方法、解码方法、 系统及 装置, 可以对谐波结构进行重新构建。
为解决上述技术问题,本发明实施例一方面,提供了一种编码方法, 包括: 通过窄带信号部分基音信息获取谐波结构的特征信息;
将所述谐波结构的特征信息进行量化编码, 向解码端发送。
另一方面, 提供了一种解码方法, 包括:
接收编码端发送的谐波结构的特征信息;
根据所述谐波结构的特征信息、窄带信号部分的浊音增益,重构谐波结构。 另一方面, 提供了一种编解码系统, 包括:
编码端, 用于通过窄带信号部分基音信息获取谐波结构的特征信息; 将所 述谐波结构的特征信息进行量化编码, 并发送;
解码端, 用于所述谐波结构的特征信息; 根据所述谐波结构的特征信息、 窄带信号部分的浊音增益, 重构谐波结构。
另一方面, 提供了一种编码端, 包括:
获取单元, 用于通过窄带信号部分基音信息获取谐波结构的特征信息; 编码单元, 用于将所述谐波结构的特征信息进行量化编码;
发送单元,用于向解码端发送所述编码单元量化编码后的谐波结构的特征 信息。
另一方面, 提供了一种解码端, 包括:
接收单元, 用于接收编码端发送的谐波结构的特征信息;
重构单元 ,用于根据所述谐波结构的特征信息、窄带信号部分的浊音增益, 重构谐波结构。
由以上技术方案可以看出,在本发明实施例中, 编码端根据窄带信号部分 的基音信息获取谐波结构的特征信息, 并将谐波结构的特征信息发送到解码 端,使解码端可以根据谐波结构的特征信息重构谐波结构,避免了不能在解码 端恢复出谐波结构, 给最终的听觉效果带来的损害。
附图说明
图 1为本发明提供的编码方法实施例一流程图;
图 2为本发明提供的解码方法实施例一流程图;
图 3为本发明提供的编码方法实施例二流程图;
图 4为本发明提供的解码方法实施例二流程图;
图 5为本发明提供的编码方法实施例三流程图;
图 6为本发明提供的一种编解码系统实施例图;
图 7为本发明提供的一种编码端实施例图;
图 8为本发明提供的一种解码端实施例图。
具体实施方式
本发明实施例提供了一种编码方法、 解码方法、 系统及装置, 可以对谐波 结构进行重新构建。
当窄带信号部分出现了强烈的谐波结构时,在超宽带信号部分也会有很明 显的谐波结构,如果在解码端能正确地重构出超宽带信号部分的谐波结构,对 听觉效果会有很好的提升。 由于超宽带信号部分的信号进行了 MDCT变换之 后, 其谐波结构在变换域就表现为周期性的峰值, 如果不受编码资源限制, 能 够将所有的 MDCT系数编码传输到解码端, 就可以保留这一周期性峰值的特 性,从而进行精确地谐波重构。但是通常用于超宽带信号部分编码的资源是有 限的, 这一特性在编码传输的过程中就会丟失掉。 如果能得到这组峰值在 MDCT 系数中的起始位置和峰值出现的周期性, 就可以在解码端进行 MDCT 反变换之前,对特定位置的 MDCT系数进行处理,对处理过的系数进行 MDCT 反变换, 以达到较精确的重构超宽带信号部分谐波结构的目的, 而这些特定的 位置就由起始位置和周期来决定。本发明实施例中,通过提取表征超宽带信号 部分谐波结构的特征信息: 如谐波间隔和谐波起始位置, 利用这些特征信息在 解码端就能很好的重构出超宽带信号部分的谐波结构, 达到带宽扩展的目的, 改善整体听觉效果。 在核心编码层为 CELP编码的宽带 /超宽带语音编解码系统中, 输入信号 经过核心层 CELP编码之后,得到了窄带信号部分的基音信息 ^,浊音增益 G; 输入信号的超宽带信号部分经过了 MDCT 变换之后得到了一组 MDCT 系数 {y _ swb(0 y _swb(l ……, _w6(319)}。 当输入信号具有明显的谐波结构特性时, 其 MDCT系数会有表现出明显的周期性, 这一周期即是谐波间隔。 窄带信号 部分的基音周期7^已经能够大致的描述出超宽带信号部分的谐波间隔,可以通 过简化的平均幅度差函数 ( AMDF, Average Magnitude Difference Function )法 在7^附近寻找超宽带信号部分更精确的谐波间隔 。 确定了超宽带信号部分谐 波间隔 Δ后, 在 [0 Δ]范围内搜索超宽带信号部分 MDCT系数中谐波起始位置 P。
在本发明提供的编码方法中,编码端通过窄带信号部分基音信息获取谐波 结构的特征信息; 将所述谐波结构的特征信息进行量化编码, 向解码端发送。 解码端收到后可以根据所述谐波结构的特征信息进行解码, 重构谐波结构。
在本发明提供的解码方法中,解码端接收编码端发送的谐波结构的特征信 息;根据所述谐波结构的特征信息、窄带信号部分的浊音增益,重构谐波结构。
本发明提供的编码方法实施例一流程如图 1所示:
步骤 101、 通过窄带信号部分的基音信息计算出谐波间隔的取值区间。 由于窄带信号部分的基音周期7^为已知量, 基音周期在时域为7 则在频 域为 1 / 7S 转换到 MDCT域为 32Q/7% 其中 320 为和 8kHz频率范围对应的 MDCT系数的个数, 因此可以得出谐波间隔的初始值为 = 32()/
以谐波间隔△的波动范围为 ±△' , 由于不同的使用环境, 可能导致 不同, 因此 Δ'是一个根据实际经验获得的经验值, 在以前的使用经验中 Δ'为多少, 取 值即为多少, 此时, 谐波间隔 Δ的取值区间可以表述为。= ^。- Δ' Δ+ Δ']。
步骤 102、 根据谐波间隔的取值区间、 MDCT系数获得谐波间隔。
^=Δ0+Δ'
Α ΔΊ = arg min (D(k)) 在区间 Ω内找出 AMDF函数最小值的位置 A , 即 , 其中
D(k) = _ swb(n + k) - y _swb(n)
"=。 " " '为 AMDF函数; 其中, n为 MDCT系数的顺 序编号, 取第几个 MDCT系数, n即为几; k的最大取值为 4 k的最小 取值为△。_△'。 在区间 Ω内对 AMDF函数积分求得 Δ2;
由 和八2加权求得超宽带信号部分的谐波间隔 Δ = "χΔ1 + (1 - )χ Δ2
在一些特殊情况下, 或者对精度要求不高时, 可以直接使用窄带信号部分 的基音周期 Τ。作为超宽带信号部分的谐波间隔 A。
步骤 103、根据谐波间隔和谱系数中第一个峰值所在的位置获得谐波结构 起始位置, 也即超宽带信号部分第一个谐波出现的位置。
以谱系数是 MDCT谱系数为例, 在 [0 Δ]内搜索 MDCT系数中峰值的位 k=A P
^ Ρ(、 = are max (y swb(k)) 、 ^ ^ , . ^ ^ ρ , „, , rt P ^、 =
置 。 , 即 ° S =。 - 、 "; 这一位置对^于第 个谐波, 即 Δ 示取整函数;
推算出 [Q ^区间内超宽带信号部分第一个谐波出现的大概位置 即
Ρ2 ϋ χΔ ;
在 [ρ2 - 1 的区间内对 [y - )}求积分, 得到超宽带信号部分谐波结 构起始位置 P。
步骤 104、 对谐波间隔、 谐波结构起始位置进行量化编码, 并发送。
用 N比特对谐波间隔 Δ进行量化编码, 用 M比特对谐波结构起始位置 P 进行量化编码, 将这两部分信息作为传输码流进行打包传输。在本实施例中 N 可以取值为 3 , M可以取值为 7。
本发明提供的编码方法实施例一, 编码端根据窄带信号部分的基音信息、 谱系数获得谐波间隔、 谐波结构起始位置, 并将谐波间隔、 谐波结构起始位置 发送到解码端 ,使解码端可以根据谐波间隔、谐波结构起始位置重构谐波结构。
需要说明的是, 本发明提供的编码方法实施例一, 同样适用于窄带信号部 分、 宽带信号部分。
在编码端使用了本发明提供的编码方法实施例一时,本发明提供的解码方 法实施例一流程如图 2所示:
步骤 201、 接收谐波间隔、 谐波结构起始位置。
解码端从码流中解码得到谐波间隔 Δ、 谐波结构起始位置 Ρ、 以及窄带信 号部分的浊音增益 G。
步骤 202、 根据谐波间隔、 谐波结构起始位置、 以及窄带信号部分的浊音 增益进行谐波重构。 由于为了频带连续, 会将 7 ~ 8kHz频带内的信号放入超宽带部分进行一 起处理, 因此本步骤中需要计算出的超宽带信号部分谐波结构的起始位置,也 即 7kHz以上谐波结构的起始位置。
根据谐波间隔、 谐波结构起始位置, 计算出 7kHz以上谐波结构的起始位 置尸。 由于 P代表的是 8kHz以上的信号中谐波结构的起始位置, 7~8kHz这 1kHz频率范围内的 MDCT系数的个数为 40 , 因此 7kHz以上谐波结构的起始 位置尸可以使用公式表述为:
'= + 40 - ||( + 40) / Δ||χΔ . 从尸位置开始, 以 Δ为间隔在 7kHz~14kHz的超宽带信号部分按如下方法 进行谐波重构:
在 Ρ'+"χΔ位置处的 MDCT系数由一个用浊音增益 G控制的具有随机相位 的数来填充, 具体填充方法为: 在 Ρ'+" Χ Δ位置处的 y— swb ( i )
Q _ 240-i
=(- 1)M *G* I*2 其中对 8~14kHz部分, ― 240 , 对 7~8kHz部分 Q=l , M是一个随机的整数。 其中, n为 MDCT系数的顺序编号, 取第几个 MDCT 系数, n即为几。
步骤 203、 按照正常的步骤进行解码, 得到增强了谐波结构的超宽带信号 部分信号。
本发明提供的解码方法实施例一,在收到编码端发送的谐波间隔、谐波结 构起始位置时, 可以根据谐波间隔、 谐波结构起始位置重构谐波结构。
需要说明的是, 本发明提供的解码方法实施例一, 同样适用于窄带信号部 分、 宽带信号部分。
进一步,还可以根据超宽带信号部分时域信号、 窄带信号部分的基音信息 生成谐波信息, 发送给解码端, 以使解码端可以重构谐波结构。
当前帧输入信号经过核心层 CELP编码之后,得到了窄带信号部分的基音 信息 浊音增益 G , 当前帧输入信号的超宽带信号部分时域信号为 {y_hi(0ly_hi(ll……,_y -/ '(3 l9)}。 当输入信号在窄带信号部分有明显的谐波结构 时, 其超宽带信号部分信号也会表现出明显的周期性, 而这一周期和窄带信号 部分的基音周期 。很接近, 因此可以用窄带信号部分的基音周期近似的表示超 宽带信号部分信号的周期性。通过这一周期 。可以近似的估计出超宽带信号部 分信号的谐波结构在频域的起始位置,但这一估计是比较粗糙的, 不精确的谐 波结构起始位置会导致在解码端进行谐波重构时出现误差累计,而导致某些应 有谐波结构的频点位置被忽略掉,使得重构的谐波结构产生偏差,损害了听觉 感受。 因此可以在已知窄带信号部分基音周期7 ^的基础上, 通过 AMDF法在 超宽带信号部分找到更精确地信号周期 T, 然后将这一周期转换为 MDCT系 数中谐波结构起始位置信息 P, 并将这一信息编码传送给解码器。 解码器根据 这一信息就能在 MDCT系数中准确地建立谐波结构, 进行较好的谐波重构。
本发明提供的编码方法实施例二流程如图 3所示:
步骤 301、 通过窄带信号部分的基音信息计算出超宽带信号部分信号周期 的取值区间。
超宽带信号部分信号周期 T的取值区间为: Ω = [ - « 。 +^。
其中"一般为进行基音周期搜索时的分数延迟, 通常取值为《 = 1/3
步骤 302、根据超宽带信号部分信号周期的取值区间及超宽带信号部分时 域信号获得超宽带信号部分信号周期。
在此区间 Ω内通过 AMDF 函数找出超宽带信号部分信号周期 Τ , 即
Figure imgf000009_0001
为 AMDF函数。 其中, n 为 MDCT系数的顺序编号, 取第几个 MDCT系数, n即为几; k的最大取值 为^ +"' , k的最小取值为 。―"。
步骤 303、 根据超宽带信号部分信号周期获得谐波结构起始位置。
将超宽带信号部分信号周期 Τ转化为谐波结构在谱系数中的起始位置 Ρ,
16000
Ρ =
Tx 25 示取整函数。
步骤 304、 对谐波结构起始位置进行量化编码, 并发送。
用 M比特对谐波结构起始位置 P进行量化编码, 将这部分信息作为传输 码流进行打包传输。 在本实施例中 M可以取值为 7。
本发明提供的编码方法实施例二, 编码端根据窄带信号部分的基音信息、 超宽带信号部分时域信号获得谐波结构起始位置,并将谐波结构起始位置发送 到解码端, 使解码端可以根据谐波结构起始位置重构谐波结构。
需要说明的是, 本发明提供的编码方法实施例二, 同样适用于窄带信号部 分、 宽带信号部分。
在编码端使用了本发明提供的编码方法实施例二时,本发明提供的解码方 法实施例二流程如图 4所示:
步骤 401、 接收谐波结构起始位置。
解码端从码流中解码得到窄带信号部分基音周期 T。、谐波结构起始位置 Ρ、 浊音增益 G。
步骤 402、 根据谐波结构起始位置、 窄带信号部分基音周期、 以及窄带信 号部分的浊音增益进行谐波重构。
由于为了频带连续, 会将 7 ~ 8kHz频带内的信号放入超宽带部分进行一 起处理, 因此本步骤中需要计算出的超宽带信号部分谐波间隔、谐波结构的起 始位置, 也即 7kHz以上谐波间隔、 谐波结构的起始位置。
根据窄带信号部分基音周期7 计算出 7kHz以上谐波间隔: 窗 再根据计算出的 7kHz以上谐波间隔△、 谐波结构起始位置 Ρ、 浊音增益 G, 重构谐波结构。
根据 7kHz以上谐波间隔△、谐波结构起始位置 Ρ, 计算出 7kHz以上谐波 结构的起始位置: Ρ'= Ρ + 40 - |(Ρ + 40)/Δ|ΧΔ ;
从尸位置开始, 以 Δ为间隔在 7kHz~14kHz的超宽带信号部分按如下方法 进行^谐波重构:
在 Ρ'+"χΔ位置处的 MDCT系数由一个用浊音增益 G控制的具有随机相位 的数来填充, 具体填充方法为: 在 Ρ'+" Χ Δ位置处的 y— swb ( i )
Q _ 240-i
=(- 1)M *G* I*2 其中对 8~14kHz部分, ― 240 , 对 7~8kHz部分 Q=l , M是一个随机的整数。
步骤 403、 按照正常的步骤进行解码, 得到增强了谐波结构的超宽带信号 部分信号。
本发明提供的解码方法实施例二,在收到编码端发送的谐波间隔、谐波结 构起始位置时, 可以根据谐波间隔、 谐波结构起始位置重构谐波结构。
需要说明的是, 本发明提供的解码方法实施例二, 同样适用于窄带信号部 分、 宽带信号部分。
进一步,还可以根据窄带信号部分的基音信息、 当前帧及前一帧的超宽带 信号部分时域信号生成谐波信息,发送给解码端, 以使解码端可以重构谐波结 构。
当前帧输入信号经过核心层 CELP编码之后,得到了窄带部分的基音信息
Τ。 , 浊音增益 G , 当前帧输入信号的超宽带部分时域信号为 {y-hi(oiy_hi(il……,_y_ (3 l9)} ,保存在编码寄存器中的上一帧信号的超宽带部 分时域信号为 ^ _ oW _ hi(° y_ ld_hi{\), , _ ο/ί/ _ /z (319) }。
本发明提供的编码方法实施例三流程如图 5所示:
步骤 501、 通过窄带信号部分的基音信息计算出超宽带信号部分信号周期 的取值区间。
超宽带部分信号周期 τ所在区间为: Ω = [ - « +
其中"一般为进行基音周期搜索时的分数延迟, 通常取值为《 = 1/3
步骤 502、 根据超宽带信号部分信号周期的取值区间、 当前帧及前一帧的 超宽带信号部分时域信号获得超宽带信号部分信号周期。
在此区间 Ω内通过 AMDF函数找出超宽带部分信号周期 Τ, 此时超宽带 部分信号周期 Τ可以使用公式表示为:
Figure imgf000011_0001
其中, η为超宽带部分时域信号的顺序编号, 取第几个超宽带部分时域信 号, η即为几; k的最大取值为 k的最小取值为 ― "。
y(n) = {y_ old _ hi(0), y—old _ hi(\\…… ,y _old _hi(3 \9),y _ hi(0), y _ hi(l),…… ,y_hi(3 \9)}
步骤 503、 根据超宽带信号部分信号周期获得谐波结构起始位置。
将超宽带部分信号周期转化 T为谐波结构在 MDCT系数中的起始位置 P, 此时谐波结构起始位置 P可以使用公式表示为:
Figure imgf000011_0002
步骤 504、 对谐波结构起始位置进行量化编码, 并发送。
用 M比特对谐波结构起始位置 P进行量化编码, 将这部分信息作为传输 码流进行打包传输。 在本实施例中 M可以取值为 7。
本发明提供的编码方法实施例三, 编码端根据窄带信号部分的基音信息、 当前帧及前一帧的超宽带信号部分时域信号获得谐波结构起始位置 ,并将谐波 结构起始位置发送到解码端 ,使解码端可以根据谐波结构起始位置重构谐波结 构。
需要说明的是, 本发明提供的编码方法实施例三, 同样适用于窄带信号部 分、 宽带信号部分。
在编码端使用了本发明提供的编码方法实施例三时,解码端可以使用本发 明提供的解码方法实施例二进行谐波重构。
本发明提供的编解码谐波信息的方法实施例包括:
在编码端, 通过窄带信号部分基音信息获取谐波结构的特征信息; 将所述谐波结构的特征信息进行量化编码, 向解码端发送。
在解码端, 接收编码端发送的谐波结构的特征信息;
根据所述谐波结构的特征信息、窄带信号部分的浊音增益,重构谐波结构。 本领域普通技术人员可以理解实现上述实施例方法中的全部或部分步骤 是可以通过程序来指令相关的硬件完成,所述的程序可以存储于一种计算机可 读存储介质中, 该程序在执行时, 包括如下步骤:
一种编码方法, 包括:
通过窄带信号部分基音信息获取谐波结构的特征信息;
将所述谐波结构的特征信息进行量化编码, 向解码端发送。
一种解码方法, 包括:
接收编码端发送的谐波结构的特征信息;
根据所述谐波结构的特征信息、窄带信号部分的浊音增益,重构谐波结构。 上述提到的存储介质可以是只读存储器, 磁盘或光盘等。
本发明提供的一种编解码系统实施例如图 6所示, 包括:
编码端 601 , 用于通过窄带信号部分基音信息获取谐波结构的特征信息; 将所述谐波结构的特征信息进行量化编码, 并发送;
解码端 602, 用于接收所述谐波结构的特征信息; 根据所述谐波结构的特 征信息、 窄带信号部分的浊音增益, 重构谐波结构。 应用本发明提供的一种编解码系统实施例,编码端根据窄带信号部分的基 音信息、 获取谐波结构的特征信息, 并将谐波结构的特征信息发送到解码端, 使解码端可以根据谐波结构的特征信息重构谐波结构。
本发明提供的一种编码端实施例如图 7所示, 包括:
获取单元 710 , 用于通过窄带信号部分基音信息获取谐波结构的特征信 息;
编码单元 720 , 用于将所述谐波结构的特征信息进行量化编码;
发送单元 720 , 用于向解码端发送所述编码单元 720量化编码后的谐波结 构的特征信息。
在所述谐波结构的特征信息包括: 谐波间隔、 谐波结构起始位置时; 所述获取单元 710包括:
第一谐波间隔单元 711 ,用于通过窄带信号部分的基音周期获得谐波间隔; 第一起始位置单元 712, 用于根据所述第一谐波间隔单元获得的谐波间隔 和修正的离散余弦变换系数中第一个峰值所在的位置获得谐波结构起始位置。
其中, 所述第一谐波间隔单元 711包括:
第二谐波间隔单元 711a, 用于以所述基音周期作为所述谐波间隔; 和 /或, 第三谐波间隔单元 711b, 用于通过窄带信号部分的基音周期计算出所述 谐波间隔的取值区间,根据谐波间隔的取值区间、修正的离散余弦变换系数获 得所述谐波间隔。
在所述谐波结构的特征信息包括: 谐波结构起始位置时;
所述获取单元 710包括:
信号周期的取值区间单元 713 , 用于通过所述基音信息计算出信号周期的 取值区间;
信号周期单元 714, 用于根据所述信号周期的取值区间及时域信号获得信 号周期;
第二起始位置单元 715 , 用于根据信号周期获得谐波结构起始位置。
本发明提供的一种解码端实施例如图 8所示, 包括:
接收单元 810 , 用于接收编码端发送的谐波结构的特征信息;
重构单元 820 , 用于根据所述谐波结构的特征信息、 窄带信号部分的浊音 增益, 重构谐波结构。
在所述谐波结构的特征信息包括: 谐波间隔、 谐波结构起始位置时; 所述重构单元 820包括:
第三起始位置单元 821 , 用于根据所述谐波间隔、 谐波结构起始位置获得 超宽带信号部分谐波结构的起始位置;
第一重构单元 822, 用于根据所述超宽带信号部分谐波结构的起始位置、 所述谐波间隔、 所述浊音增益, 重构谐波结构。
在所述谐波结构的特征信息包括: 谐波结构起始位置时;
所述重构单元 820包括:
第四谐波间隔单元 823 , 用于根据窄带信号部分基音周期, 获得超宽带信 号部分谐波间隔;
第二重构单元 824, 用于根据所述超宽带信号部分谐波间隔、 所述谐波结 构的起始位置、 所述浊音增益, 重构谐波结构。
本发明实施例提供的编码端、解码端实施例, 具体工作方式可参考上文对 本发明提供的编码方法、 解码方法实施例, 在此不再重复。
在本发明提供的编解码系统实施例、 编码端、 解码端实施例中, 编码端根 据窄带信号部分的基音信息获取谐波结构的特征信息,并将谐波结构的特征信 息发送到解码端,使解码端可以根据谐波结构的特征信息重构谐波结构,避免 了不能在解码端恢复出谐波结构, 给最终的听觉效果带来的损害。
以上对本发明所提供的一种编码方法、解码方法、 系统及装置进行了详细 施例的说明只是用于帮助理解本发明的方法及其核心思想; 同时,对于本领域 的一般技术人员,依据本发明的思想, 在具体实施方式及应用范围上均会有改 变之处, 综上所述, 本说明书内容不应理解为对本发明的限制。

Claims

权 利 要 求
1、 一种编码方法, 其特征在于, 包括:
通过窄带信号部分基音信息获取谐波结构的特征信息;
将所述谐波结构的特征信息进行量化编码, 向解码端发送。
2、 如权利要求 1所述的编码方法, 其特征在于, 所述谐波结构的特征信 息包括: 谐波间隔、 谐波结构起始位置;
所述通过窄带信号部分基音信息获取谐波结构的特征信息包括:
通过窄带信号部分的基音周期获得谐波间隔;
根据谐波间隔和修正的离散余弦变换系数中第一个峰值所在的位置获得 谐波结构起始位置。
3、 如权利要求 2所述的编码方法, 其特征在于, 所述通过窄带信号部分 的基音周期获得谐波间隔包括:
以所述基音周期作为所述谐波间隔; 或者,
通过窄带信号部分的基音周期计算出所述谐波间隔的取值区间,根据谐波 间隔的取值区间、 修正的离散余弦变换系数获得所述谐波间隔。
4、 如权利要求 1所述的编码方法, 其特征在于, 所述谐波结构的特征信 息包括: 谐波结构起始位置;
所述通过窄带信号部分基音信息获取谐波结构的特征信息包括:
通过所述基音信息计算出信号周期的取值区间;
根据所述信号周期的取值区间及时域信号获得信号周期;
才艮据信号周期获得谐波结构起始位置。
5、 如权利要求 4所述的编码方法, 其特征在于, 所述时域信号包括: 当 前帧时域信号或当前帧与前一帧时域信号。
6、 一种解码方法, 其特征在于, 包括:
接收编码端发送的谐波结构的特征信息;
根据所述谐波结构的特征信息、窄带信号部分的浊音增益,重构谐波结构。
7、 如权利要求 6所述的解码方法, 其特征在于, 所述谐波结构的特征信 息包括: 谐波间隔、 谐波结构起始位置;
所述根据所述谐波结构的特征信息、 窄带信号部分的浊音增益, 重构谐波 结构包括:
根据所述谐波间隔、谐波结构起始位置获得超宽带信号部分谐波结构的起 始位置;
根据所述超宽带信号部分谐波结构的起始位置、所述谐波间隔、所述浊音 增益, 重构谐波结构。
8、 如权利要求 6所述的解码方法, 其特征在于, 所述谐波结构的特征信 息包括: 谐波结构起始位置;
所述根据所述谐波结构的特征信息、 窄带信号部分的浊音增益, 重构谐波 结构包括:
根据窄带信号部分基音周期, 获得超宽带信号部分谐波间隔;
根据所述超宽带信号部分谐波间隔、所述谐波结构的起始位置、 所述浊音 增益, 重构谐波结构。
9、 一种编解码系统, 其特征在于, 包括:
编码端, 用于通过窄带信号部分基音信息获取谐波结构的特征信息; 将所 述谐波结构的特征信息进行量化编码, 并发送;
解码端, 用于接收所述谐波结构的特征信息; 根据所述谐波结构的特征信 息、 窄带信号部分的浊音增益, 重构谐波结构。
10、 一种编码端, 其特征在于, 包括:
获取单元, 用于通过窄带信号部分基音信息获取谐波结构的特征信息; 编码单元, 用于将所述谐波结构的特征信息进行量化编码;
发送单元,用于向解码端发送所述编码单元量化编码后的谐波结构的特征 信息。
11、 如权利要求 10所述的编码端, 其特征在于, 所述谐波结构的特征信 息包括: 谐波间隔、 谐波结构起始位置;
所述获取单元包括: 第一谐波间隔单元, 用于通过窄带信号部分的基音周期获得谐波间隔; 第一起始位置单元,用于根据所述第一谐波间隔单元获得的谐波间隔和修 正的离散余弦变换系数中第一个峰值所在的位置获得谐波结构起始位置。
12、 如权利要求 11所述的编码端, 其特征在于, 所述第一谐波间隔单元 包括:
第二谐波间隔单元, 用于以所述基音周期作为所述谐波间隔; 和 /或, 第三谐波间隔单元,用于通过窄带信号部分的基音周期计算出所述谐波间 隔的取值区间,根据谐波间隔的取值区间、修正的离散余弦变换系数获得所述 谐波间隔。
13、 如权利要求 10所述的编码端, 其特征在于, 所述谐波结构的特征信 息包括: 谐波结构起始位置;
所述获取单元包括:
信号周期的取值区间单元,用于通过所述基音信息计算出信号周期的取值 区间;
信号周期单元,用于根据所述信号周期的取值区间及时域信号获得信号周 期;
第二起始位置单元,用于根据所述信号周期单元获得的信号周期获得谐波 结构起始位置。
14、 一种解码端, 其特征在于, 包括:
接收单元, 用于接收编码端发送的谐波结构的特征信息;
重构单元 ,用于根据所述谐波结构的特征信息、窄带信号部分的浊音增益, 重构谐波结构。
15、 如权利要求 14所述的解码端, 其特征在于, 所述谐波结构的特征信 息包括: 谐波间隔、 谐波结构起始位置;
所述重构单元包括:
第三起始位置单元, 用于根据所述谐波间隔、谐波结构起始位置获得超宽 带信号部分谐波结构的起始位置; 第一重构单元, 用于根据所述超宽带信号部分谐波结构的起始位置、所述 谐波间隔、 所述浊音增益, 重构谐波结构。
16、 如权利要求 14所述的解码端, 其特征在于, 所述谐波结构的特征信 息包括: 谐波结构起始位置;
所述重构单元包括:
第四谐波间隔单元, 用于根据窄带信号部分基音周期, 获得超宽带信号部 分谐波间隔;
第二重构单元, 用于根据所述超宽带信号部分谐波间隔、 所述谐波结构的 起始位置、 所述浊音增益, 重构谐波结构。
PCT/CN2009/071123 2008-04-03 2009-04-01 编码方法、解码方法、系统及装置 WO2009121298A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN200810089863.5 2008-04-03
CNA2008100898635A CN101552005A (zh) 2008-04-03 2008-04-03 编码方法、解码方法、系统及装置

Publications (1)

Publication Number Publication Date
WO2009121298A1 true WO2009121298A1 (zh) 2009-10-08

Family

ID=41134846

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2009/071123 WO2009121298A1 (zh) 2008-04-03 2009-04-01 编码方法、解码方法、系统及装置

Country Status (2)

Country Link
CN (1) CN101552005A (zh)
WO (1) WO2009121298A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
RU2744477C2 (ru) * 2012-03-29 2021-03-10 Телефонактиеболагет Л М Эрикссон (Пабл) Преобразующее кодирование/декодирование гармонических звуковых сигналов

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2011048815A1 (ja) 2009-10-21 2011-04-28 パナソニック株式会社 オーディオ符号化装置、復号装置、方法、回路およびプログラム
RU2725416C1 (ru) * 2012-03-29 2020-07-02 Телефонактиеболагет Лм Эрикссон (Пабл) Расширение полосы частот гармонического аудиосигнала
WO2016142002A1 (en) * 2015-03-09 2016-09-15 Fraunhofer-Gesellschaft Zur Foerderung Der Angewandten Forschung E.V. Audio encoder, audio decoder, method for encoding an audio signal and method for decoding an encoded audio signal

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH09237100A (ja) * 1996-02-29 1997-09-09 Matsushita Electric Ind Co Ltd 音声符号化・復号化装置
US6119082A (en) * 1998-07-13 2000-09-12 Lockheed Martin Corporation Speech coding system and method including harmonic generator having an adaptive phase off-setter
CN1455390A (zh) * 2002-04-30 2003-11-12 Lg电子株式会社 估算声音编码器的谐波的装置和方法
CN1639769A (zh) * 2002-06-27 2005-07-13 三星电子株式会社 利用谐波提取的音频编码方法和设备
CN1849648A (zh) * 2003-09-16 2006-10-18 松下电器产业株式会社 编码装置和译码装置
CN101044553A (zh) * 2004-10-28 2007-09-26 松下电器产业株式会社 可扩展编码装置、可扩展解码装置及其方法

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0994464A1 (fr) * 1998-10-13 2000-04-19 Koninklijke Philips Electronics N.V. Procédé destiné à génére un signal large bande a partir d'un signal en bande étroite, appareil pour realiser un tel procédé et equipement téléphonique comportant un tel appareil

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH09237100A (ja) * 1996-02-29 1997-09-09 Matsushita Electric Ind Co Ltd 音声符号化・復号化装置
US6119082A (en) * 1998-07-13 2000-09-12 Lockheed Martin Corporation Speech coding system and method including harmonic generator having an adaptive phase off-setter
CN1455390A (zh) * 2002-04-30 2003-11-12 Lg电子株式会社 估算声音编码器的谐波的装置和方法
CN1639769A (zh) * 2002-06-27 2005-07-13 三星电子株式会社 利用谐波提取的音频编码方法和设备
CN1849648A (zh) * 2003-09-16 2006-10-18 松下电器产业株式会社 编码装置和译码装置
CN101044553A (zh) * 2004-10-28 2007-09-26 松下电器产业株式会社 可扩展编码装置、可扩展解码装置及其方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
CHENG,M. ET AL.: "A 4kbit/s Hybrid Speech Coding System", JIANGSU COMMUNICATION TECHNOLOGY, vol. 18, no. 5, pages 23 - 26 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
RU2744477C2 (ru) * 2012-03-29 2021-03-10 Телефонактиеболагет Л М Эрикссон (Пабл) Преобразующее кодирование/декодирование гармонических звуковых сигналов
US11264041B2 (en) 2012-03-29 2022-03-01 Telefonaktiebolaget Lm Ericsson (Publ) Transform encoding/decoding of harmonic audio signals

Also Published As

Publication number Publication date
CN101552005A (zh) 2009-10-07

Similar Documents

Publication Publication Date Title
US10249313B2 (en) Adaptive bandwidth extension and apparatus for the same
KR101854297B1 (ko) 시간 도메인 여기 신호를 기초로 하는 오류 은닉을 사용하여 디코딩된 오디오 정보를 제공하기 위한 오디오 디코더 및 방법
KR101854296B1 (ko) 시간 도메인 여기 신호를 변형하는 오류 은닉을 사용하여 디코딩된 오디오 정보를 제공하기 위한 오디오 디코더 및 방법
JP5203929B2 (ja) スペクトルエンベロープ表示のベクトル量子化方法及び装置
EP2041745B1 (en) Adaptive encoding and decoding methods and apparatuses
US8532982B2 (en) Method and apparatus to encode and decode an audio/speech signal
US9218817B2 (en) Low-delay sound-encoding alternating between predictive encoding and transform encoding
EP3503098B1 (en) Apparatus and method decoding an audio signal using an aligned look-ahead portion
RU2636685C2 (ru) Решение относительно наличия/отсутствия вокализации для обработки речи
KR100603167B1 (ko) 시간 동기식 파형 보간법을 이용한 피치 프로토타입파형으로부터의 음성 합성
JP2002530705A (ja) 音声の無声セグメントの低ビットレート符号化
CN106605263B (zh) 确定用于编码lpd/fd过渡帧的预算
WO2009121298A1 (zh) 编码方法、解码方法、系统及装置
WO2010000179A1 (zh) 频带扩展的方法、系统和设备
JP2000132193A (ja) 信号符号化装置及び方法、並びに信号復号装置及び方法
KR101546793B1 (ko) 오디오 신호의 부호화/복호화 방법 및 장치
KR20080034817A (ko) 부호화/복호화 장치 및 방법

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: 09726671

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 09726671

Country of ref document: EP

Kind code of ref document: A1