An audio encoding, decoding method, apparatus, device and storage medium

Abstract

Embodiments of the present application provide an audio encoding and decoding method, device, equipment and storage medium. The method comprises: determining a category of an audio frame to be encoded as a target category; wherein the category of the audio frame is divided according to a numerical range of quantized spectral coefficients; selecting an encoding table corresponding to the target category from a plurality of generated encoding tables corresponding to categories as a first target encoding table; and encoding the quantized spectral coefficients of the audio frame to be encoded by using the first target encoding table. It can be seen that in the present scheme, different encoding tables are used to encode audio frames of different categories, and the category of the audio frame is divided according to the numerical range of the quantized spectral coefficients, that is, the present scheme takes into account the different characteristics of different quantized spectral coefficients, thereby improving the encoding efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of audio technology, in particular to an audio encoding and decoding method, device, apparatus and storage medium. BACKGROUND

[0002] Audio encoding can reduce the redundancy of audio data, facilitating the transmission and storage of audio data. In some related audio encoding schemes, a set of encoding tables is usually generated, for example, a codebook is used in Huffman encoding scheme, and a cumulative probability table is used in arithmetic coding scheme. The encoding table includes the probability of occurrence of each byte, and the quantized spectral coefficients of the audio are encoded according to the encoding table.

[0003] However, in this scheme, all quantized spectral coefficients of the audio are encoded using the same encoding table, without considering the different characteristics of different quantized spectral coefficients, resulting in low encoding efficiency. SUMMARY

[0004] The purpose of the embodiments of the present application is to provide an audio encoding and decoding method, device, apparatus and storage medium to improve the encoding efficiency.

[0005] To achieve the above purpose, the present application provides an audio encoding method, comprising:

[0006] determining the category of the audio frame to be encoded as a target category; wherein the category of the audio frame is divided according to the numerical range of the quantized spectral coefficients;

[0007] selecting the encoding table corresponding to the target category from the generated encoding tables corresponding to multiple categories as a first target encoding table;

[0008] encoding the quantized spectral coefficients of the audio frame to be encoded using the first target encoding table.

[0009] Optionally, the category division of the audio frame is also based on any one or more of the following: code rate, content, stationarity, and harmonicity;

[0010] The process of generating the encoding tables corresponding to the multiple categories comprises:

[0011] The encoding tables corresponding to the multiple categories are generated by statistically analyzing the probability distribution of the quantized spectral coefficients of the audio frames of different categories.

[0012] Optionally, the process of generating the encoding tables corresponding to the multiple categories by statistically analyzing the probability distribution of the quantized spectral coefficients of the audio frames of different categories comprises:

[0013] The M*K*N*P*L codebooks are generated by performing probability distribution statistics on N code rates, M types of signals, K types of stationarity, L types of harmonic signals and P ranges of quantized spectral coefficients.

[0014] Optionally, the method further comprises:

[0015] selecting the codebook corresponding to the target category from the pre-generated codebooks corresponding to multiple categories as a first candidate codebook;

[0016] updating the first candidate codebook by performing probability distribution statistics on the quantized spectral coefficients of the audio frames to be coded in the current time period to obtain a first target codebook.

[0017] Optionally, the method further comprises:

[0018] determining whether the interval between the current time and the time when the first candidate codebook was last updated reaches a preset time length condition;

[0019] if the interval does not reach the preset time length condition, determining the selected first candidate codebook as the first target codebook;

[0020] if the interval reaches the preset time length condition, performing the step of updating the first candidate codebook by performing probability distribution statistics on the quantized spectral coefficients of the audio frames to be coded in the current time period to obtain a first target codebook.

[0021] Optionally, the method further comprises:

[0022] selecting the codebook corresponding to the target category from the pre-generated codebooks corresponding to multiple categories as a first candidate codebook;

[0023] updating the first candidate codebook by performing probability distribution statistics on the quantized spectral coefficients of the audio frames to be coded in the current time period to obtain a second candidate codebook;

[0024] determining the coding efficiency of coding using the first candidate codebook as a first coding efficiency, and determining the coding efficiency of coding using the second candidate codebook as a second coding efficiency;

[0025] if the first coding efficiency is greater than the second coding efficiency, determining the first candidate codebook as the first target codebook;

[0026] If the first encoding efficiency is less than the second encoding efficiency, the second candidate encoding table is determined as a first target encoding table.

[0027] Optionally, after the quantized spectral coefficients of the audio frame to be encoded are encoded by using the first target encoding table, the method further includes:

[0028] The first target encoding table is added to the padding bit part and / or the additional data part of the audio frame to be encoded.

[0029] To achieve the above object, an embodiment of the present application further provides an audio decoding method, which includes:

[0030] An audio frame to be decoded is obtained, which is encoded by using the audio encoding method;

[0031] An encoding table corresponding to the audio frame to be decoded is obtained as a second target encoding table;

[0032] The audio frame to be decoded is decoded by using the second target encoding table.

[0033] Optionally, the obtaining of the encoding table corresponding to the audio frame to be decoded as the second target encoding table includes:

[0034] The encoding table corresponding to the audio frame to be decoded is read from the padding bit part and / or the additional data part of the audio frame to be encoded as the second target encoding table.

[0035] Optionally, the obtaining of the encoding table corresponding to the audio frame to be decoded as the second target encoding table includes:

[0036] A third candidate encoding table is received from an encoding end as a third candidate encoding table;

[0037] The third candidate encoding table is updated to obtain a second target encoding table by performing probability distribution statistics on quantized spectral coefficients of the audio frame to be decoded in a current time period.

[0038] To achieve the above object, an embodiment of the present application further provides an audio encoding device, which includes:

[0039] A first determining module is configured to determine a category of an audio frame to be encoded as a target category, wherein the category of the audio frame is divided according to a value range of quantized spectral coefficients;

[0040] A selecting module is configured to select an encoding table corresponding to the target category from encoding tables corresponding to multiple categories generated as a first target encoding table;

[0041] An encoding module is configured to encode quantized spectral coefficients of the audio frame to be encoded by using the first target encoding table.

[0042] Optionally, the category division of the audio frame is based on any one or more of the following: code rate, content, stationarity, and harmonic property; and the apparatus further comprises:

[0043] a generating module configured to generate an encoding table corresponding to each of the categories by performing probability distribution statistics on quantized spectral coefficients of audio frames of different categories.

[0044] Optionally, the generating module is specifically configured to:

[0045] generate M*K*N*P*L encoding tables by performing probability distribution statistics on quantized spectral coefficients of N code rates, M categories of signals, K stationarity, L harmonic property signals, and P ranges.

[0046] Optionally, the selecting module comprises:

[0047] a selecting sub-module configured to select an encoding table corresponding to the target category from the pre-generated encoding tables corresponding to the multiple categories as a first candidate encoding table.

[0048] an updating sub-module configured to update the first candidate encoding table by performing probability distribution statistics on quantized spectral coefficients of audio frames to be encoded in a current time period to obtain a first target encoding table.

[0049] Optionally, the apparatus further comprises:

[0050] a determining module configured to determine whether an interval between the current time and a time when the first candidate encoding table was last updated reaches a preset time length condition after the selecting sub-module selects an encoding table corresponding to the target category from the pre-generated encoding tables corresponding to the multiple categories as the first candidate encoding table; if not, trigger a second determining module, and if so, trigger the updating sub-module.

[0051] the second determining module is configured to determine the selected first candidate encoding table as the first target encoding table.

[0052] Optionally, the selecting module is specifically configured to:

[0053] select an encoding table corresponding to the target category from the pre-generated encoding tables corresponding to the multiple categories as the first candidate encoding table.

[0054] update the first candidate encoding table by performing probability distribution statistics on quantized spectral coefficients of audio frames to be encoded in a current time period to obtain a second candidate encoding table.

[0055] determining an encoding efficiency of encoding by using the first candidate encoding table as a first encoding efficiency; and determining an encoding efficiency of encoding by using the second candidate encoding table as a second encoding efficiency;

[0056] if the first encoding efficiency is greater than the second encoding efficiency, determining the first candidate encoding table as a first target encoding table;

[0057] if the first encoding efficiency is less than the second encoding efficiency, determining the second candidate encoding table as the first target encoding table.

[0058] Optionally, the apparatus further comprises:

[0059] an adding module, configured to add the first target encoding table to a padding bit part and / or an additional data part of the audio frame to be encoded.

[0060] To achieve the above object, an embodiment of the present application further provides an audio decoding apparatus, comprising:

[0061] a first obtaining module, configured to obtain an audio frame to be decoded, wherein the audio frame to be decoded is encoded by using the audio encoding apparatus;

[0062] a second obtaining module, configured to obtain an encoding table corresponding to the audio frame to be decoded as a second target encoding table;

[0063] a decoding module, configured to decode the audio frame to be decoded by using the second target encoding table.

[0064] Optionally, the second obtaining module is specifically configured to:

[0065] read the encoding table corresponding to the audio frame to be decoded from a padding bit part and / or an additional data part of the audio frame to be encoded as the second target encoding table.

[0066] Optionally, the second obtaining module is specifically configured to:

[0067] receive an encoding table sent by an encoding end as a third candidate encoding table;

[0068] update the third candidate encoding table by performing probability distribution statistics on quantized spectral coefficients of the audio frame to be decoded in a current time period to obtain a second target encoding table.

[0069] To achieve the above object, an embodiment of the present application further provides an electronic device, comprising a processor and a memory, wherein:

[0070] the memory is configured to store a computer program;

[0071] A processor is configured to implement any of the above-mentioned audio encoding method or audio decoding method when executing a program stored in a memory.

[0072] To achieve the above object, the embodiment of the present application further provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program is configured to implement any of the above-mentioned audio encoding method or audio decoding method when executed by a processor.

[0073] According to the embodiment of the present application, the category of the audio frame to be encoded is determined as the target category, wherein the category of the audio frame is divided according to the value range of the quantized spectral coefficient; the encoding table corresponding to the target category is selected from the generated encoding tables corresponding to the multiple categories as the first target encoding table; and the quantized spectral coefficient of the audio frame to be encoded is encoded by using the first target encoding table. It can be seen that, in the present solution, different encoding tables are used to encode the audio frames of different categories, and the category of the audio frame is divided according to the value range of the quantized spectral coefficient, that is, the present solution takes into account the different characteristics of different quantized spectral coefficients, and the encoding efficiency is improved.

[0074] Of course, implementing any product or method of the present application does not necessarily require achieving all the advantages mentioned above at the same time. BRIEF DESCRIPTION OF DRAWINGS

[0075] In order to more clearly illustrate the technical solutions of the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor based on these drawings.

[0076] Figure 1 The first flowchart of the audio encoding method provided by an exemplary embodiment of the present application is shown in the figure;

[0077] Figure 2 The flowchart of generating an encoding table provided by an exemplary embodiment of the present application is shown in the figure;

[0078] Figure 3 The structure diagram of the DRA audio frame provided by an exemplary embodiment of the present application is shown in the figure;

[0079] Figure 4 The second flowchart of the audio encoding method provided by an exemplary embodiment of the present application is shown in the figure;

[0080] Figure 5 The first flowchart of the audio decoding method provided by an exemplary embodiment of the present application is shown in the figure;

[0081] Figure 6 A second flowchart of an audio decoding method according to an example embodiment of the present application is provided.

[0082] Figure 7 A third flowchart of an audio decoding method according to an example embodiment of the present application is provided.

[0083] Figure 8 A structure diagram of an audio encoding device according to an example embodiment of the present application is provided.

[0084] Figure 9 A structure diagram of an audio decoding device according to an example embodiment of the present application is provided.

[0085] Figure 10 A structure diagram of an electronic device according to an example embodiment of the present application is provided. DETAILED DESCRIPTION

[0086] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art based on the present application are within the scope of protection of the present application.

[0087] To achieve the above-mentioned purpose, the embodiments of the present application provide an audio encoding and decoding method, device, equipment and storage medium. The method and device can be applied to various electronic equipment, and the specific type of the electronic equipment is not limited. First, the audio encoding method will be described in detail below. The steps in the following method embodiments can be executed in a logical order, and the step label or the order of introducing each step does not limit the execution order of each step.

[0088] Figure 1 A first flowchart of an audio encoding method according to an example embodiment of the present application is provided, which includes:

[0089] S101: Determine the category of the audio frame to be encoded as the target category; wherein the category of the audio frame is divided according to the numerical range of the quantized spectral coefficient.

[0090] In an example embodiment of the present application, the category of the audio frame can be divided only according to the numerical range of the quantized spectral coefficient of the audio frame. Taking Huffman encoding as an example, P ranges can be divided according to the maximum spectral coefficient of each codebook interval, P represents a positive integer, and 9 codebooks for encoding and dividing the steady-state signal in DRA (Digital Rise Audio, digital audio encoding technology) are referred to.

[0091] In other exemplary embodiments of the present application, the classification basis of the audio frame can include any one or more of the following: code rate, content, stationarity, harmonic, etc. The code rate can be a code rate range or a code rate value. The content can be classified into speech, music, mixed, etc. The stationarity can be classified into transient, quasi-transient, stationary, etc. The stationary signal can be further classified into harmonic, such as strong harmonic, general harmonic, weak harmonic, and non-harmonic, etc.

[0092] S102: Selecting the encoding table corresponding to the target category from the generated encoding tables corresponding to the multiple categories as a first target encoding table.

[0093] S103: Encoding the quantized spectral coefficients of the audio frame to be encoded using the first target encoding table.

[0094] For example, if Huffman encoding is used, the encoding table is a code book. If arithmetic encoding is used, the encoding table is a cumulative probability table. The encoding table includes the probability of each byte occurrence. The quantized spectral coefficients of the audio frame can be encoded according to the encoding table.

[0095] In an exemplary embodiment of the present application, the process of generating the encoding tables corresponding to the multiple categories can include: generating the encoding tables corresponding to the multiple categories by performing probability distribution statistics on the quantized spectral coefficients of the audio frames of different categories.

[0096] As described above, the classification basis of the audio frame can include the value range of the quantized spectral coefficients, code rate, content, stationarity, harmonic, etc. In an exemplary embodiment of the present application, M*K*N*P*L encoding tables are generated by performing probability distribution statistics on N code rates, M types of signals, K types of stationarity, L types of harmonic signals, and P ranges of quantized spectral coefficients. M, K, N, P, and L are positive integers.

[0097] For example, as shown in FIG. 2, the encoding tables corresponding to the multiple categories can be generated by the following steps: Figure 2

[0098] S201: First, determine the N code rates supported by the encoding end (encoder or encoding software), such as 24 kbps, 32 kbps, 40 kbps, 48 kbps, etc. or N code rate ranges, such as 20-27.999 kbps, 28-35.999 kbps, 36-43.999 kbps, etc.

[0099] S202: Classify the audio frame into M types of signals, such as speech, music, mixed, etc. using a classification algorithm.

[0100] ​S203, detecting the stability of the input audio frame signal by a signal detection method, and dividing into K kinds of stable signals, such as transient, quasi-transient, steady state and the like.

[0101] S204, if it is a steady state signal, further dividing into L kinds of harmonic signals, such as strong harmonic, general harmonic, weak harmonic and non-harmonic and the like; if it is a non-steady state signal, set L = 1.

[0102] S205, the probability distribution statistics of N kinds of code rate (or code rate range), M kinds of signals, K kinds of stability, L kinds of harmonic signals and P range of quantization spectrum coefficients are obtained N*M*K*L*P probability distribution. If Huffman coding algorithm is used, Huffman code book design is carried out, and M*K*N*P*L code book is generated. If arithmetic coding is used, M*K*N*P*L cumulative probability table is calculated.

[0103] In an exemplary embodiment of the present application, a plurality of categories corresponding to the code table can be generated in advance, and the code table generated in this embodiment can be fixed.

[0104] Alternatively, in another exemplary embodiment of the present application, the generated code table can be updated every certain period of time, and in this embodiment, S102 can include: selecting the code table corresponding to the target category from the plurality of categories corresponding to the code table generated in advance as the first candidate code table; updating the first candidate code table by performing probability distribution statistics on the quantization spectrum coefficients of the audio frame to be coded in the current time period to obtain the first target code table.

[0105] In this embodiment, the code table that matches the characteristics of the quantization spectrum coefficients of the current time period can be generated according to the probability distribution of the quantization spectrum coefficients of the current time period, and the coding efficiency is further improved by using this code table for coding.

[0106] In an exemplary embodiment of the present application, after selecting the code table corresponding to the target category from the plurality of categories corresponding to the code table generated in advance as the first candidate code table, it can be judged whether the interval between the current time and the time when the first candidate code table is updated last time reaches a preset time length condition; if not, the selected first candidate code table is determined as the first target code table; if it reaches, the first candidate code table is updated by performing probability distribution statistics on the quantization spectrum coefficients of the audio frame to be coded in the current time period to obtain the first target code table.

[0107] For example, a new coding table (also referred to as a dynamic coding table) can be generated or the fixed coding table (the first candidate coding table) can be updated every time interval T or every frame number Nf (Nf=T / N*Fs, N is the frame length of the audio frame to be encoded, that is, the number of samples, and Fs is the audio sampling rate). If the dynamic coding table is generated, the dynamic coding table is used for encoding, and if not, the fixed coding table (the first candidate coding table) is used for encoding in this embodiment, or the last generated dynamic coding table can be used for encoding in other exemplary embodiments.

[0108] In another exemplary embodiment of the present application, the coding efficiency of the fixed coding table and the updated coding table can also be compared, and the coding table with higher coding efficiency is selected for encoding. In this embodiment, S102 can include:

[0109] The coding table corresponding to the target category is selected from the plurality of pre-generated coding tables corresponding to categories as the first candidate coding table;

[0110] The first candidate coding table is updated by performing probability distribution statistics on the quantized spectral coefficients of the audio frames to be encoded in the current time period to obtain a second candidate coding table;

[0111] The coding efficiency of encoding using the first candidate coding table is determined as the first coding efficiency, and the coding efficiency of encoding using the second candidate coding table is determined as the second coding efficiency;

[0112] If the first coding efficiency is greater than the second coding efficiency, the first candidate coding table is determined as the first target coding table;

[0113] If the first coding efficiency is less than the second coding efficiency, the second candidate coding table is determined as the first target coding table.

[0114] In this embodiment, the coding table with higher coding efficiency is selected from the comparison of the fixed coding table and the updated coding table for encoding, further improving the coding efficiency.

[0115] In an exemplary embodiment of the present application, the first target coding table can be sent to the decoding end so that the decoding end uses the first target coding table for decoding, for example, the first target coding table can be added to the padding bit part and / or the additional data part of the audio frame to be encoded.

[0116] As mentioned above, a dynamic encoding table can be generated every certain number of frames Nf or every time period T. In this case, the dynamic encoding table can be sent to the decoding end so that the decoding end can use the dynamic encoding table for decoding. Alternatively, a fixed encoding table can be sent to the decoding end. The decoding end updates the received fixed encoding table by performing probability distribution statistics on the quantization spectral coefficients of the audio frames to be decoded within the current time period, obtains the dynamic encoding table, and then uses the dynamic encoding table for decoding. In this way, only the fixed encoding table needs to be transmitted, and the dynamic encoding table does not need to be transmitted in real time, saving transmission resources.

[0117] The transmission methods are described below. For example, when the encoding mode is CBR (Constant Bitrate) mode, the following two transmission methods can be used: Figure 3 The structure of a DRA (Digital Rise Audio) audio frame shown is used as an example for illustration. The structure of audio frames under other encoding methods is similar, such as... Figure 3 As shown, an audio frame includes a synchronization word, frame header information, multiple channels (channel 0, channel 1, channel 2... channel n), padding bits, and additional data. The frame header information includes frame header type, frame length, number of blocks per frame, sampling rate index, number of channels, LFE (Low Frequency Effect), additional data, MS (Mid-Side), JIC (Jointintensity Coding), etc., which will not be described in detail here.

[0118] (1) Transmission method without occupying extra data blocks: Since each frame has a fixed number of bits, while the audio signal content is changing, more or less useless bits are usually filled in each frame in addition to transmitting the encoded audio content. At this time, the useless padding bits at the end of the audio frame can be used to transmit the first target encoding table. This transmission method will not affect the original audio encoding quality.

[0119] (2) Methods that require additional data blocks for transmission: such as Figure 3 The final appended data portion of the encoded frame can be defined as a data block to transmit the first target encoding table. By transmitting the padding bit portion and the appended data portion together, the transmission speed can be accelerated, and the update rate of the first target encoding table can be improved.

[0120] The first transmission mode does not affect the original audio encoding quality, but the transmission speed is usually slow, resulting in slow updating of the first target encoding table, which belongs to slow adaptive entropy encoding. If the probability distribution of the quantized spectral coefficients of the encoded audio signal changes at a rate lower than the updating rate of the current first target encoding table, the adaptive entropy encoding efficiency will not be affected. If the change is faster than the updating rate of the current first target encoding table, the adaptive ability will be affected, and the entropy encoding efficiency will be affected. At this time, the second transmission mode can be selected according to the selection.

[0121] The updating rate of the first target encoding table in the second transmission mode is higher, but since the first target encoding table is transmitted using an additional data block, the number of available bits in the front of a frame will be reduced under the fixed code rate requirement, and thus the encoding quality will be affected to a certain extent.

[0122] For example, if the dynamic encoding table is transmitted, the second transmission mode can be used since the use of the dynamic encoding table can improve the entropy encoding efficiency. The comprehensive result of "the second transmission mode affecting the encoding quality" and "the dynamic encoding table improving the entropy encoding efficiency" still improves the encoding effect. In some cases, the comprehensive result does not improve the encoding effect, and only the fixed encoding table can be transmitted.

[0123] For other encoding modes such as VBR (Variable Bitrate, variable code rate) and ABR (Average Bitrate, average code rate), the second transmission mode can be used to transmit the dynamic encoding table.

[0124] Alternatively, the first target encoding table can be added to the additional data part of the audio frame to be encoded.

[0125] After transmission, the encoding end can use the dynamic encoding table for entropy encoding (referred to as an adaptive entropy encoding mode), or the encoding end can use the fixed encoding table for entropy encoding (referred to as a quasi-adaptive entropy encoding mode), or a hybrid mode can be used, that is, the dynamic encoding table and the fixed encoding table are used for entropy encoding, or the encoding efficiency of the fixed encoding table and the dynamic encoding table can be compared, and the encoding table with higher encoding efficiency is selected for entropy encoding, and a bit flag is used to indicate whether the fixed encoding table or the dynamic encoding table is used for the frame.

[0126] For the case of wireless transmission (e.g. broadcast application): when adaptive entropy coding mode is used in audio coding, if the decoding end does not receive from the first frame of coding, it can not receive the complete dynamic coding table, at this time, it is necessary to switch to quasi-adaptive entropy coding mode in time, so that the decoding end can use the fixed coding table for accurate entropy decoding. In addition, the decoding end can re-receive the dynamic coding table until it is received completely, and enable the dynamic coding table for entropy decoding. Or if there is a bit flag indicating whether to use a fixed coding table or a dynamic coding table, then during entropy decoding, the fixed coding table or the dynamic coding table is selected according to the bit flag for accurate entropy decoding.

[0127] For example, as shown in Figure 4 The encoding can be performed by the following steps:

[0128] S401, initialize the encoding, and set to use the fixed coding table.

[0129] S402, perceptual coding of the audio frame to be encoded, wherein for the quantized spectral coefficients, entropy coding is performed based on the fixed coding table.

[0130] For example, assuming that a frame of audio frame to be encoded is input, audio type analysis is performed on the audio frame, and it is determined that the signal type is m; stationary detection is performed on it, and it is determined that the signal stationarity is k; it is determined whether it is a stationary signal; if so, harmonic analysis is performed on it, and it is determined that the signal harmonic is l; if not, set l = 1; according to the stationarity of the signal, perform multi-resolution MDCT (Modified Discrete Cosine Transform); according to the set code rate (or code rate range) n and the bit allocation based on psychoacoustics, quantize the spectral coefficients; divide the quantized spectral system into intervals, and analyze the maximum quantized spectral coefficient in each interval as p; then call the coding table corresponding to the code rate n, type m, stationarity k, harmonic l, and maximum absolute value p of the spectral coefficient as the first target coding table; use the first target coding table to perform entropy coding on the quantized spectral coefficients of the audio frame to be encoded.

[0131] S403, the probability distribution of the quantized spectral coefficients can be continuously or intermittently counted, for example, S403 includes: S4031, judging whether a certain frame number Nf or time T is reached, if so, executing S4032, generating a new dynamic coding table according to the probability distribution obtained by counting, otherwise returning to S402.

[0132] S404, transmit the dynamic coding table, and there are various transmission modes, for example:

[0133] (a) for CBR mode, the dynamic coding table is transmitted by the above-mentioned first transmission mode or second transmission mode.

[0134] (b) for VBR or ABR mode, the dynamic coding table is transmitted by the above-mentioned (2) transmission mode.

[0135] S405, after the dynamic coding table is transmitted, the encoding end switches to entropy coding using the dynamic coding table (adaptive entropy coding mode); or switches to a hybrid mode, that is, selects to use the fixed coding table or the dynamic coding table for entropy coding; or compares the coding efficiency of the fixed coding table and the dynamic coding table, and selects the coding table with higher coding efficiency for entropy coding, and uses one bit flag to indicate whether the fixed coding table or the dynamic coding table is used for the current frame.

[0136] S406, after a period of time (or a certain number of frames), a one-bit flag is used to indicate whether the current frame is a reset state; if it is a reset state, the entropy coding part of the encoding end needs to be reset to the initial state to ensure that the decoding end starts to start at any time, without obtaining the dynamic coding table, needs to wait until the reset frame is received to start using the fixed coding table for accurate entropy decoding, and returns to S402. If it is not a reset state, continue normal entropy coding.

[0137] In this embodiment, the encoding end and the decoding end can be reset at regular intervals. Before the dynamic coding table is generated after the reset, the encoding end and the decoding end use the existing fixed coding table for entropy coding and entropy decoding, which can make the decoding end synchronized with the encoding end.

[0138] For some large file audio codec application cases (such as local, server or remote storage), the above steps can be used for entropy coding, but due to the particularity of its transmission, the decoding end generally initiates the transmission request, so the decoding can be started from the beginning of the file, and there is no need to reset regularly, that is, the above S406 can be omitted, and S405 returns to S403.

[0139] As described above, in some cases, only the fixed coding table needs to be transmitted, and the dynamic coding table does not need to be transmitted in real time. In these cases, the decoding end performs probability distribution statistics on the quantized spectral coefficients of the audio frames to be decoded (corresponding to the probability distribution statistics of the quantized spectral coefficients of the audio frames to be encoded by the encoding end), obtains the same probability distribution as the encoding end, and generates the same dynamic coding table at the decoding end. The dynamic coding table is used for decoding.

[0140] In these cases, S404 and S406 can be omitted, and S405 no longer waits for the dynamic coding table to be transmitted, and S405 returns to execute S402.

[0141] In the application, the category of the audio frame to be encoded is determined as a target category, wherein the category of the audio frame is divided according to the value range of the quantized spectral coefficient; the encoding table corresponding to the target category is selected from the generated encoding tables corresponding to multiple categories as a first target encoding table; and the quantized spectral coefficient of the audio frame to be encoded is encoded by using the first target encoding table. In the application, different encoding tables are used to encode audio frames of different categories, and the category of the audio frame is divided according to the value range of the quantized spectral coefficient, that is, the application takes into account the different characteristics of different quantized spectral coefficients, and the encoding efficiency is improved.

[0142] In addition, in the above embodiments, the wireless transmission and the large file application are considered, and the entropy encoding schemes suitable for different situations are respectively proposed.

[0143] Corresponding to the above encoding method, an audio decoding method is provided in an embodiment of the application, as shown in the following. Figure 5

[0144] S501: An audio frame to be decoded is obtained. The audio frame to be decoded can be encoded by using the audio encoding method in any of the above embodiments.

[0145] S502: An encoding table corresponding to the audio frame to be decoded is obtained as a second target encoding table.

[0146] In an exemplary embodiment of the application, S502 can include: reading the encoding table corresponding to the audio frame to be decoded from the filled bit part and / or the additional data part of the audio frame to be decoded as the second target encoding table.

[0147] In the above embodiments, multiple transmission modes are introduced, and the decoding end can read the encoding table from the corresponding part of the audio frame to which the encoding table is added. In order to distinguish the description, the target encoding table in the encoding end embodiment is referred to as the first target encoding table, and the target encoding table in the decoding end embodiment is referred to as the second target encoding table, and the first target encoding table and the second target encoding table can refer to the same encoding table.

[0148] In another exemplary embodiment of the application, S502 can include: receiving the encoding table sent by the encoding end as a third candidate encoding table; updating the third candidate encoding table to obtain the second target encoding table by performing probability distribution statistics on the quantized spectral coefficient of the audio frame to be decoded in the current time period.

[0149] ​As mentioned above, in some cases, only the fixed coding table needs to be transmitted, and the dynamic coding table does not need to be transmitted in real time. In these cases, the decoding end performs probability distribution statistics on the quantized spectral coefficients of the audio frame to be decoded (corresponding to the probability distribution statistics performed by the encoding end on the quantized spectral coefficients of the audio frame to be encoded), obtains the same probability distribution as the encoding end, and updates the received fixed coding table (the third candidate coding table) to generate the same dynamic coding table, i.e., the second target coding table.

[0150] S503: Decoding the audio frame to be decoded using the second target coding table.

[0151] For example, corresponding to Figure 4 , as shown in Figure 6 , the decoding can be performed by the following steps:

[0152] S601: Initialize decoding, and set to use the fixed coding table for entropy decoding.

[0153] S602: Input an audio frame to be decoded, and check whether the entropy encoding part of the frame at the encoding end is in a reset state. If not, do not decode the frame, continue to input the next audio frame to be decoded, and continue the step; if it is in the reset state, execute S603.

[0154] S603: Start normal decoding, in which the entropy decoding uses the set coding table.

[0155] S604: Detect whether a dynamic coding table is added at the end of the audio frame to be decoded. If yes, execute S605; if not, input the next audio frame to be decoded, and return to S603.

[0156] S605: Start receiving the dynamic coding table, and determine whether a complete dynamic coding table has been received. If not, input the next audio frame to be decoded, and return to S603; if yes, execute S606.

[0157] S606: Set to use the dynamic coding table for entropy decoding (the decoding mode corresponding to the adaptive entropy encoding mode); or determine to set to use the fixed coding table or the dynamic coding table for entropy decoding according to the current frame entropy encoding type identifier (1-bit flag).

[0158] S607: Input the next audio frame to be decoded, and check whether the entropy encoding part of the frame at the encoding end is in a reset state. If not, return to S603; if yes, execute S608, set to use the fixed coding table for entropy decoding (the decoding mode corresponding to the quasi-adaptive entropy encoding mode), and return to S603.

[0159] In this embodiment, the encoding end and the decoding end can be reset at a time, and before a dynamic coding table is generated after the reset, the encoding end and the decoding end use an existing fixed coding table for entropy encoding and entropy decoding, so that the decoding end can be synchronized with the encoding end.

[0160] As described above, for some large file audio codec cases, due to the particularity of transmission, the decoding end generally initiates a transmission request, so that decoding can be performed from the beginning of the file, and a reset is not needed. In this case, S602, S607 and S608 can be omitted in the decoding step, in S603, an encoded frame is input, and normal decoding is started, in which entropy decoding uses a set coding table, and S606 returns to S603.

[0161] As described above, in some cases, only a fixed coding table needs to be transmitted, and a dynamic coding table does not need to be transmitted in real time. In these cases, the decoding step can be as shown in Figure 7

[0162] S701, initialization decoding, and setting to use a fixed coding table for entropy decoding.

[0163] S702, inputting an encoded frame, and starting normal decoding of the encoded frame, in which entropy decoding uses a set coding table.

[0164] S703, starting probability distribution statistics of quantized spectral coefficients obtained through entropy decoding, and judging whether a certain frame number Nf or time T is reached. If yes, S704 is executed, a new coding table (that is, a dynamic coding table) is generated according to the probability distribution, and S705 is entered, otherwise, S702 is returned.

[0165] S705, setting to use a dynamic coding table for entropy decoding (a decoding mode corresponding to an adaptive entropy coding mode); or judging to set to use a fixed coding table or a dynamic coding table for entropy decoding according to a current frame entropy coding type identifier (1-bit flag), and returning to S702.

[0166] For example, the entropy decoding process can include: obtaining quantized spectral coefficients through entropy decoding, performing inverse quantization, scale factor or step size operation, multi-channel decoding, inverse MDCT, and outputting decoded PCM (Pulse Code Modulation).

[0167] Corresponding to the method embodiment, the embodiment of the application further provides an audio encoding device, as shown in Figure 8 , which comprises:

[0168] A first determination module 801 is configured to determine a category of an audio frame to be encoded as a target category, wherein the category of the audio frame is divided according to a value range of quantized spectral coefficients.​

[0169] The selecting module 802 is configured to select the coding table corresponding to the target category from the generated coding tables corresponding to multiple categories as a first target coding table.

[0170] The encoding module 803 is configured to encode the quantized spectral coefficients of the audio frame to be encoded by using the first target coding table.

[0171] In an exemplary embodiment of the present application, the category division of the audio frame further comprises any one or more of the following: code rate, content, stationarity, and harmonic property; and the device further comprises:

[0172] The generating module (not shown in the figure) is configured to generate the coding tables corresponding to the multiple categories by performing probability distribution statistics on the quantized spectral coefficients of the audio frames of different categories.

[0173] In an exemplary embodiment of the present application, the generating module is specifically configured to:

[0174] The generating module is configured to generate M*K*N*P*L coding tables by performing probability distribution statistics on the quantized spectral coefficients of N code rates, M types of signals, K types of stationarity, L types of harmonic property, and P ranges.

[0175] In an exemplary embodiment of the present application, the selecting module 802 comprises a selecting submodule and an updating submodule (not shown in the figure), wherein:

[0176] The selecting submodule is configured to select the coding table corresponding to the target category from the pre-generated coding tables corresponding to multiple categories as a first candidate coding table.

[0177] The updating submodule is configured to update the first candidate coding table by performing probability distribution statistics on the quantized spectral coefficients of the audio frames to be encoded in the current time period to obtain the first target coding table.

[0178] In an exemplary embodiment of the present application, the device further comprises a judging module and a second determining module (not shown in the figure), wherein:

[0179] The judging module is configured to judge whether the interval between the current time and the time when the first candidate coding table is last updated reaches a preset time length condition after the selecting submodule selects the coding table corresponding to the target category from the pre-generated coding tables corresponding to multiple categories as a first candidate coding table; if not, the second determining module is triggered, and if yes, the updating submodule is triggered.

[0180] The second determining module is configured to determine the selected first candidate coding table as the first target coding table.

[0181] In an example embodiment of the present application, the selection module 802 is specifically configured to:

[0182] select, from the pre-generated encoding tables corresponding to the plurality of categories, an encoding table corresponding to the target category as a first candidate encoding table;

[0183] update the first candidate encoding table to obtain a second candidate encoding table by performing probability distribution statistics on the quantized spectral coefficients of the audio frame to be encoded in the current time period;

[0184] determine an encoding efficiency of encoding using the first candidate encoding table as a first encoding efficiency, and determine an encoding efficiency of encoding using the second candidate encoding table as a second encoding efficiency;

[0185] if the first encoding efficiency is greater than the second encoding efficiency, determine the first candidate encoding table as a first target encoding table;

[0186] if the first encoding efficiency is less than the second encoding efficiency, determine the second candidate encoding table as the first target encoding table.

[0187] In an example embodiment of the present application, the apparatus further comprises:

[0188] an adding module (not shown in the figure) configured to add the first target encoding table to the padding bit part and / or the additional data part of the audio frame to be encoded.

[0189] Corresponding to the method embodiments described above, the present application also provides an audio decoding apparatus, as shown in Figure 9 the apparatus comprises:

[0190] a first obtaining module 901 configured to obtain an audio frame to be decoded, the audio frame to be decoded being encoded by using the audio encoding method described above;

[0191] a second obtaining module 902 configured to obtain an encoding table corresponding to the audio frame to be decoded as a second target encoding table;

[0192] a decoding module 903 configured to decode the audio frame to be decoded by using the second target encoding table.

[0193] In an example embodiment of the present application, the second obtaining module 902 is specifically configured to:

[0194] read, from the padding bit part and / or the additional data part of the audio frame to be encoded, the encoding table corresponding to the audio frame to be decoded as the second target encoding table.

[0195] In an example embodiment of the present application, the second obtaining module 902 is specifically configured to:

[0196] receive the encoding table sent by the encoding end as a third candidate encoding table;

[0197] update the third candidate encoding table by performing probability distribution statistics on the quantized spectral coefficients of the audio frames to be decoded in the current time period, to obtain a second target encoding table.

[0198] The embodiment of the present application also provides an electronic device, such as Figure 10 as shown, comprising a processor 1001 and a memory 1002.

[0199] The memory 1002 is used for storing a computer program.

[0200] The processor 1001 is used for executing the program stored in the memory 1002, so as to realize any one of the above-mentioned audio encoding method or audio decoding method.

[0201] The memory mentioned above can include a random access memory (RAM) and a non-volatile memory (NVM), for example, at least one disk memory. Optionally, the memory can also be at least one storage device located away from the processor.

[0202] The processor mentioned above can be a general processor, including a central processing unit (CPU), a network processor (NP), etc.; can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component.

[0203] In another embodiment provided by the present application, a computer readable storage medium is also provided, and the computer readable storage medium stores a computer program, and the computer program is executed by a processor to realize any one of the above-mentioned audio encoding method or audio decoding method.

[0204] In another embodiment provided by the present application, a computer program product containing instructions is also provided, and when the computer program product is run on a computer, the computer is caused to execute any one of the above-mentioned audio encoding method or audio decoding method.

[0205] In the embodiments described above, all or some of the steps can be implemented by software, hardware or firmware, or any combination thereof. When implemented by software, all or some of the steps can be implemented in the form of one or more computer programs. The computer program can be stored in any computer readable medium, and loaded into the computer system for execution. The computer readable medium includes: a computer storage medium and a computer communication medium. The computer storage medium includes: volatile media (such as random access memory (RAM) and others) and non-volatile media (such as read-only memory (ROM), floppy disks, CD-ROMs, optical disks, hard disks, etc.). The computer communication medium includes: computer networks and other media.

[0206] It should be noted that, in this document, the terms such as first and second are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. In addition, the terms "comprise", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device that includes a list of elements does not only include those elements, but also includes other elements not explicitly listed, or further includes elements inherent in such a process, method, article or device. Without more limitations, the element defined by the statement "comprises a" does not exclude the presence of additional identical elements in the process, method, article or device that includes the element.

[0207] Each of the embodiments in the specification is described in a related manner, and the same or similar parts between each of the embodiments can be referred to each other. Each of the embodiments focuses on the difference from other embodiments. In particular, for the device embodiments, the equipment embodiments, the computer readable storage medium embodiments, and the computer program product embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the part of the method embodiments.

[0208] The above merely describes the preferred embodiments of the present application, but is not used to limit the protection scope of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. An audio encoding method, comprising: determining a category of an audio frame to be encoded as a target category; wherein the category of the audio frame is divided according to a numerical range including quantized spectral coefficients, and the numerical range includes a plurality of ranges; selecting an encoding table corresponding to the target category from a plurality of generated encoding tables, at least part of the encoding tables corresponding to the plurality of ranges respectively, as a first target encoding table; encoding quantized spectral coefficients of the audio frame to be encoded by using the first target encoding table.

2. The method of claim 1, wherein, The category of the audio frame is further divided according to any one or more of the following: code rate, content, stationarity, and harmonicity; The process of generating the plurality of encoding tables corresponding to the categories includes: generating the plurality of encoding tables corresponding to the categories by performing probability distribution statistics on quantized spectral coefficients of audio frames of different categories.

3. The method of claim 2, wherein, The process of generating the plurality of encoding tables corresponding to the categories by performing probability distribution statistics on quantized spectral coefficients of audio frames of different categories includes: generating M*K*N*P*L encoding tables by performing probability distribution statistics on quantized spectral coefficients of N code rates, M types of signals, K types of stationarity, L types of harmonicity, and P ranges.

4. The method of claim 1, wherein, The process of selecting the encoding table corresponding to the target category from the plurality of generated encoding tables as the first target encoding table includes: selecting the encoding table corresponding to the target category from a plurality of pre-generated encoding tables as a first candidate encoding table; updating the first candidate encoding table by performing probability distribution statistics on quantized spectral coefficients of audio frames to be encoded in a current time period to obtain the first target encoding table.

5. The method of claim 4, wherein, The process of selecting the encoding table corresponding to the target category from the plurality of pre-generated encoding tables as the first candidate encoding table further includes: determining whether the interval between the current time and the time of the last update of the first candidate encoding table reaches a preset time length condition; if not, determining the selected first candidate encoding table as the first target encoding table; if so, performing the step of updating the first candidate encoding table by performing probability distribution statistics on quantized spectral coefficients of audio frames to be encoded in a current time period to obtain the first target encoding table.

6. The method of claim 1, wherein, The process of selecting the encoding table corresponding to the target category from the plurality of pre-generated encoding tables as the first candidate encoding table includes: selecting the encoding table corresponding to the target category from a plurality of pre-generated encoding tables as a first candidate encoding table; updating the first candidate encoding table by performing probability distribution statistics on quantized spectral coefficients of audio frames to be encoded in a current time period to obtain a second candidate encoding table; determining an encoding efficiency of encoding by using the first candidate encoding table as a first encoding efficiency, and determining an encoding efficiency of encoding by using the second candidate encoding table as a second encoding efficiency; if the first encoding efficiency is greater than the second encoding efficiency, determining the first candidate encoding table as the first target encoding table; if the first encoding efficiency is less than the second encoding efficiency, determining the second candidate encoding table as the first target encoding table.

7. The method of claim 1, wherein, The encoding of the quantized spectral coefficients of the audio frame to be encoded by using the first target coding table further comprises: adding the first target coding table to the padding bit part and / or the additional data part of the audio frame to be encoded.

8. An audio decoding method, comprising: obtaining an audio frame to be decoded, which is encoded by using the audio encoding method in any one of claims 1-7; obtaining a coding table corresponding to the audio frame to be decoded as a second target coding table; decoding the audio frame to be decoded by using the second target coding table.

9. The method of claim 8, wherein, The obtaining of the coding table corresponding to the audio frame to be decoded as the second target coding table comprises: reading the coding table corresponding to the audio frame to be decoded from the padding bit part and / or the additional data part of the audio frame to be encoded as the second target coding table.

10. The method of claim 8, wherein, The obtaining of the coding table corresponding to the audio frame to be decoded as the second target coding table comprises: receiving the coding table sent by the encoding end as a third candidate coding table; updating the third candidate coding table to obtain the second target coding table by performing probability distribution statistics on the quantized spectral coefficients of the audio frame to be decoded in a current time period.

11. An audio encoding apparatus, comprising: a first determining module configured to determine a category of an audio frame to be encoded as a target category, wherein the category of the audio frame is divided according to a numerical range of quantized spectral coefficients, and the numerical range comprises a plurality of ranges; a selecting module configured to select a coding table corresponding to the target category from a plurality of coding tables corresponding to a plurality of categories as a first target coding table, wherein at least part of the coding tables correspond to the plurality of ranges respectively; an encoding module configured to encode quantized spectral coefficients of the audio frame to be encoded by using the first target coding table.

12. The apparatus of claim 11, wherein, The category of the audio frame is further divided according to any one or more of the following: code rate, content, stationarity, and harmonic property; and the apparatus further comprises: a generating module configured to generate the plurality of coding tables corresponding to the plurality of categories by performing probability distribution statistics on quantized spectral coefficients of audio frames of different categories.

13. The apparatus of claim 12, wherein, The generating module is specifically configured to: generate M*K*N*P*L coding tables by performing probability distribution statistics on quantized spectral coefficients of N code rates, M types of signals, K types of stationarity, L types of harmonic property signals, and P ranges.

14. The apparatus of claim 11, wherein, The selecting module comprises: a selecting submodule configured to select a coding table corresponding to the target category from a plurality of coding tables corresponding to a plurality of categories generated in advance as a first candidate coding table; an updating submodule configured to update the first candidate coding table to obtain the first target coding table by performing probability distribution statistics on quantized spectral coefficients of audio frames to be encoded in a current time period.

15. The apparatus of claim 14, wherein, The apparatus further comprises: a judging module configured to, after the selecting submodule selects a coding table corresponding to the target category from a plurality of coding tables corresponding to a plurality of categories generated in advance as the first candidate coding table, judge whether an interval between a current time and a time when the first candidate coding table is last updated reaches a preset time length condition; if not, triggering a second determining module, and if so, triggering the updating submodule. The second determining module is configured to determine the selected first candidate coding table as the first target coding table.

16. The apparatus of claim 11, wherein, The selection module is specifically configured to: select a coding table corresponding to the target category from the pre-generated coding tables corresponding to multiple categories as the first candidate coding table; update the first candidate coding table to obtain a second candidate coding table by performing probability distribution statistics on the quantized spectral coefficients of the audio frame to be coded in the current time period; determine the coding efficiency of coding by using the first candidate coding table as a first coding efficiency; determine the coding efficiency of coding by using the second candidate coding table as a second coding efficiency; if the first coding efficiency is greater than the second coding efficiency, determine the first candidate coding table as the first target coding table; if the first coding efficiency is less than the second coding efficiency, determine the second candidate coding table as the first target coding table.

17. The apparatus of claim 11, wherein, The apparatus further includes: The adding module is configured to add the first target coding table to the padding bit part and / or the additional data part of the audio frame to be coded.

18. An audio decoding apparatus, comprising: A first obtaining module is configured to obtain an audio frame to be decoded, wherein the audio frame to be decoded is coded by using the audio coding apparatus according to any one of claims 11-17; A second obtaining module is configured to obtain a coding table corresponding to the audio frame to be decoded as a second target coding table; A decoding module is configured to decode the audio frame to be decoded by using the second target coding table.

19. The apparatus of claim 18, wherein, The second obtaining module is specifically configured to: read the coding table corresponding to the audio frame to be decoded from the padding bit part and / or the additional data part of the audio frame to be coded as the second target coding table.

20. The apparatus of claim 18, wherein, The second obtaining module is specifically configured to: receive a coding table sent by an encoding end as a third candidate coding table; update the third candidate coding table to obtain a second target coding table by performing probability distribution statistics on the quantized spectral coefficients of the audio frame to be decoded in the current time period.

21. An electronic device, comprising a processor and a memory, wherein: The memory is configured to store a computer program; The processor is configured to execute the program stored on the memory to implement the method steps according to any one of claims 1-10.

22. A computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the method steps according to any one of claims 1-10.

Images