Audio data rhythm analysis and playing method and device and storage medium

Abstract

The present application relates to the field of audio analysis, and discloses a kind of audio data's tone analysis and play method, equipment and storage medium.The method comprises: receiving audio data;According to preset phoneme recognition algorithm, the phoneme recognition processing is carried out to the audio data, and phoneme sequence is obtained;According to preset tone rhythm data, the phoneme sequence is divided and marked, and tone sequence is obtained;Using preset description library, the tone sequence is described and matched, and the description text corresponding to the audio data is obtained;When the audio data plays, based on the playing position of the audio data, the text field corresponding to the playing position in the description text is shown.In the embodiment of the present application, the technical problem that the current analysis technology lacks tone analysis of audio and cannot meet the demand of people for audio music appreciation is solved.

Description

Technical Field

[0001] This invention relates to the field of audio analysis, and more particularly to a method, device, and storage medium for analyzing and playing audio data based on its rhythm. Background Technology

[0002] Audio analysis refers to the digital processing and analysis of audio signals to extract useful information. It can be applied in many fields, such as speech recognition, music information retrieval, sound event detection, and emotion recognition.

[0003] In speech recognition, audio analysis can convert speech signals into text, enabling computers to understand human language. In music information retrieval, audio analysis can extract song features such as rhythm, melody, and pitch, thereby enabling song search and recommendation functions. In sound event detection, audio analysis can identify various sound events in the environment, such as car horns, dog barks, and human voices. In emotion recognition, audio analysis can infer the speaker's emotional state, such as anger, sadness, and happiness, based on sound features. Although audio analysis has diverse goals and functions, current analysis techniques lack the ability to analyze the rhythm and melody of audio, failing to meet people's needs for appreciating audio music. Therefore, a new technology is needed to address these current technical challenges. Summary of the Invention

[0004] The main objective of this invention is to solve the technical problem that current analysis techniques lack the ability to analyze the rhythm of audio, thus failing to meet people's needs for appreciating audio music.

[0005] The first aspect of this invention provides a method for rhythm analysis and playback of audio data, the method comprising:

[0006] Receive audio data;

[0007] The audio data is processed by a preset phoneme recognition algorithm to obtain a phoneme sequence.

[0008] Based on the preset rhythm data, the phoneme sequence is divided and marked to obtain the rhythm sequence;

[0009] Using a preset description library, the rhythm sequence is subjected to description matching processing to obtain the description text corresponding to the audio data;

[0010] When the audio data is played, the text field corresponding to the playback position in the description text is displayed based on the playback position of the audio data.

[0011] Optionally, in a first implementation of the first aspect of the present invention, the step of performing phoneme recognition processing on the audio data according to a preset phoneme recognition algorithm to obtain a phoneme sequence includes:

[0012] According to the preset frame window, the audio data is decomposed into N frames of audio, where N is a positive integer;

[0013] Vectorize N frames of audio data to obtain N frame vectors;

[0014] Based on the preset set of convolutional matrices, convolution processing is performed on the N frame vectors respectively to obtain N convolutional frame vectors;

[0015] The N convolutional frame vectors are normalized to obtain N normalized frame vectors;

[0016] According to the preset activation function, the N normalized frame vectors are activated to obtain N activation values;

[0017] Generate a phoneme sequence based on N activation values.

[0018] Optionally, in a second implementation of the first aspect of the present invention, the vectorization processing of N frames of audio data to obtain N frame vectors includes:

[0019] Perform sequence convolution on the frame audio data to obtain convolutional frame data;

[0020] The convolutional frame data is averaged at different time points to obtain the frame vector.

[0021] Optionally, in a third implementation of the first aspect of the present invention, the step of performing sequence convolution processing on the frame audio data to obtain convolutional frame data includes:

[0022] Based on a preset first convolution kernel, the frame data of t-2, t-1, t, t+1, and t+2 in the range of the frame audio data T are extracted and convolved to obtain the first convolution sub-audio.

[0023] Based on the preset second convolution kernel, the frame data of t-2, t, and t+2 in the range of the frame audio data T are extracted and convolved to obtain the second convolution sub-audio.

[0024] Based on a preset third convolution kernel, the frame data t in the range of the frame audio data T is extracted and convolved to obtain the third convolution sub-audio.

[0025] The first convolutional sub-audio, the second convolutional audio, and the third convolutional audio are concatenated in parallel to obtain convolutional frame data.

[0026] Optionally, in a fourth implementation of the first aspect of the present invention, the step of performing time-frame averaging on the convolutional frame data to obtain the frame vector includes:

[0027]

[0028] Where H is the average time frame, h t1 h is the audio value of the first convolutional sub-audio at time t. t2 h is the audio value of the second convolutional sub-audio at time t. t3 is the audio value of the third convolutional sub-audio at time t, where T is the end time of the frame audio data;

[0029] A frame vector is generated based on the average time frame.

[0030] Optionally, in a fifth implementation of the first aspect of the present invention, the step of activating the N normalized frame vectors according to a preset activation function to obtain N activation values ​​includes:

[0031] Based on the softmax activation function, the N normalized frame vectors are activated respectively to obtain N activation values.

[0032] Optionally, in a sixth implementation of the first aspect of the present invention, the step of dividing and marking the phoneme sequence according to preset rhythm data to obtain a rhythm sequence includes:

[0033] Read the rhythm sequence from the preset rhythm data;

[0034] Determine whether the phonological sequence has a matching sequence in the phoneme sequence;

[0035] If they exist, the matching sequences in the phoneme sequence are represented.

[0036] If it does not exist, another rhythm sequence of the preset rhythm data is read.

[0037] Optionally, in the seventh implementation of the first aspect of the present invention, the step of using a preset descriptive library to perform descriptive matching processing on the phonological sequence to obtain the descriptive text corresponding to the audio data includes:

[0038] Select the target musical note name from the musical note sequence;

[0039] Match the description field corresponding to the target musical note name in the preset description library, and establish a mapping relationship between the description field and the musical note sequence;

[0040] Based on the musical rhythm order, the description fields corresponding to all musical rhythm names in the musical rhythm sequence are combined to generate the description text corresponding to the audio data.

[0041] A second aspect of the present invention provides an audio data rhythm analysis and playback device, comprising: a memory and at least one processor, wherein the memory stores instructions, and the memory and the at least one processor are interconnected via a circuit; the at least one processor invokes the instructions in the memory to cause the audio data rhythm analysis and playback device to execute the above-described audio data rhythm analysis and playback method.

[0042] A third aspect of the present invention provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the above-described method for analyzing and playing audio data.

[0043] In this embodiment of the invention, by using a phoneme recognition process to identify various beats and instrument types in audio data, the identified phonemes are arranged in chronological order. The rhythm and melody are analyzed to obtain a rhythm order and are labeled with names. In a text description database, relevant text description documents are found, and then the text data corresponding to the audio data is combined based on the text description fields. This solves the technical problem that the analysis technology lacks rhythm analysis of audio and cannot meet people's needs for appreciating audio music. Attached Figure Description

[0044] Figure 1 This is a schematic diagram of an embodiment of the audio data rhythm analysis and playback method in this invention;

[0045] Figure 2 This is a schematic diagram of a specific embodiment of step 102 in the present invention;

[0046] Figure 3 This is a schematic diagram of a specific embodiment of step 103 in the present invention;

[0047] Figure 4 This is a schematic diagram of a specific embodiment of step 104 in the present invention;

[0048] Figure 5 This is a schematic diagram of an embodiment of an audio data rhythm analysis and playback device according to an embodiment of the present invention. Detailed Implementation

[0049] This invention provides a method, device, and storage medium for audio data rhythm analysis and playback.

[0050] The embodiments of the present invention will now be described in more detail with reference to the accompanying drawings. While some embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the present disclosure. It should be understood that the accompanying drawings and embodiments are for illustrative purposes only and are not intended to limit the scope of protection of the present invention.

[0051] In the description of the embodiments disclosed in this invention, the term "comprising" and similar terms should be understood as open-ended inclusion, i.e., "including but not limited to". The term "based on" should be understood as "at least partially based on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The terms "first", "second", etc., may refer to different or the same objects. Other explicit and implicit definitions may also be included below.

[0052] For ease of understanding, the specific process of the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 1 One embodiment of the audio data rhythm analysis and playback method in this invention includes:

[0053] 101. Receive audio data;

[0054] In this embodiment, the audio data is imported into the analysis software to analyze the audio signal of the audio information in the data.

[0055] 102. Based on a preset phoneme recognition algorithm, perform phoneme recognition processing on the audio data to obtain a phoneme sequence;

[0056] In this embodiment, phoneme identification is performed on the audio data, the composition of instruments in each audio segment and the pitch of related instruments are analyzed, and a phoneme pitch sequence of related instruments is generated based on the combination of instruments and pitches, so as to match and divide the rhythm and melody in the subsequent process.

[0057] For details, please refer to Figure 2 , Figure 2 This is a specific embodiment of step 102 in the present invention, and step 102 can be implemented in the following ways:

[0058] 1021. According to the preset frame window, the audio data is subjected to frame decomposition processing to obtain N frames of audio, where N is a positive integer;

[0059] 1022. Vectorize N frames of audio data to obtain N frame vectors;

[0060] 1023. Based on the preset set of convolutional matrices, perform convolution processing on the N frame vectors respectively to obtain N convolutional frame vectors;

[0061] 1024. Normalize the N convolutional frame vectors to obtain N normalized frame vectors;

[0062] 1025. According to the preset activation function, activate the N normalized frame vectors to obtain N activation values;

[0063] 1026. Generate a phoneme sequence based on N activation values.

[0064] In steps 1021-1026, phoneme identification is a decomposition and identification process. Assuming the audio data is video data of length 7000T, it should be noted that T is the length of the frame window. The frame window length T contains audio data from multiple actual captured frames t. The audio combination is ultimately generated from the audio of captured frames t. Therefore, the audio data is first decomposed into 7000 frames of audio using the frame window.

[0065] Vectorization of frame audio can be achieved using convolutional kernels. The current range of the audio wave is used as a feature, and the current values ​​of frame t are combined to form a frame vector of length T. Elements without a current range are set to zero, resulting in 7000 frame vectors. These frame vectors contain the desired audio features, generated through continuous convolutions using convolutional sets such as ResNet, resulting in N convolutional frame vectors.

[0066] The normalization of the N convolutional frame vectors uses the Batch Normalization (BN) method. Here, BN is a fundamental component of neural networks, which aggregates scattered features.

[0067] Let m represent the batch size and represent the number of features, i.e., the number of sample feature values. During training, for each batch of data, the Batch Normalization (BN) process performs the following operations: normalizes the data, then performs a linear transformation on it, using the mean and variance of the calculated individual vectors, then calculates the unbiased estimate, and uses the back-input error to generate N normalized frame vectors.

[0068] Finally, the softmax activation function or the ReLU activation function is used to obtain activation values. Based on the activation values, the specified instrument and tone are found, and a phoneme sequence is generated.

[0069] It should be noted that the activation process performed on the N normalized frame vectors according to the preset activation function to obtain N activation values ​​includes:

[0070] 10251. According to the softmax activation function, activate the N normalized frame vectors respectively to obtain N activation values.

[0071] In step 10251, the softmax activation function is activated by first calculating the element-wise mean of the N normalized frame vectors, and then importing e. x / ∑e x The specific settings in the activation function can be slightly adjusted according to your needs.

[0072] Specifically, step 1022 also includes the following specific implementation methods:

[0073] 10221. Perform sequence convolution on the frame audio data to obtain convolutional frame data;

[0074] 10222. Perform time-frame averaging on the convolutional frame data to obtain the frame vector.

[0075] In steps 10221-10222, convolutions are performed on individual frames of audio data T with a window length of t. These convolutions are then combined and concatenated to form convolutional frame data. This sequence concatenation primarily reduces feature loss during data extraction from the window length of T. The features of the convolutional frames are then averaged, and the resulting average value is set as an element of the frame vector. The frame vector is obtained by averaging all convolutional frames according to their element positions.

[0076] In detail, the convolution method in step 10221 can be performed by following these steps:

[0077] 102211. Based on a preset first convolution kernel, extract frame data from the range of frame audio data T at t-2, t-1, t, t+1, and t+2 and perform convolution processing to obtain the first convolution sub-audio.

[0078] 102212. Based on the preset second convolution kernel, extract the frame data t-2, t, t+2 in the range of the frame audio data T and perform convolution processing to obtain the second convolution sub-audio.

[0079] 102213. Based on the preset third convolution kernel, extract the frame data t in the range of the frame audio data T and perform convolution processing to obtain the third convolution sub-audio.

[0080] 102214. The first convolutional sub-audio, the second convolutional audio, and the third convolutional audio are concatenated in parallel to obtain convolutional frame data.

[0081] In steps 102211-102214, the first convolution kernel is a 5*1 vector. It convolves the 1*5 vector formed by the frame data t-2, t-1, t, t+1, and t+2 in the range of frame audio data T to obtain a 1*5 vector. It then convolves all t frames in the range of T to obtain the first convolution sub-audio of length T.

[0082] The second convolution kernel is a 3x1 vector. It convolves the 1x3 vector formed by the frame data t-2, t, and t+2 within the frame audio data range T. Then, it convolves the 1x3 vector formed by the frame data t-1, t+1, and t+3 within the next frame audio data range T. The resulting two convolutional data are rearranged into a 1x6 vector according to their original order. This vector is then processed within the frame audio data range T to obtain the second convolutional sub-audio of length T.

[0083] Finally, using a 1*1 kernel vector, the frame data t within the range of frame audio data T is convolved to obtain the third convolutional sub-audio of length T.

[0084] This extraction method can preserve the information relationship between the preceding and following frames, and avoid the loss of audio connection between the upper and lower audio in independent convolution. The first, second, and third convolutional sub-audio of length T are concatenated to obtain 3*T convolutional frame data.

[0085] Therefore, the step of performing time-frame averaging on the convolutional frame data to obtain the frame vector includes:

[0086]

[0087] Where H is the average time frame, h t1 h is the audio value of the first convolutional sub-audio at time t. t2 h is the audio value of the second convolutional sub-audio at time t. t3 The audio value at time t is the audio value of the third convolutional sub-audio, where T is the end time of the frame audio data; a frame vector is generated based on the average time frame.

[0088] In 3*T convolutional frame data, the elements in the same row are averaged, and after averaging all the elements, a 1*T vector value of a frame vector is obtained.

[0089] 103. Based on the preset rhythm data, the phoneme sequence is divided and marked to obtain the rhythm sequence;

[0090] In this embodiment, for example, if the canon melody is “DA-Bm-F#-GDGA”, the segments of the “DA-Bm-F#-GDGA” melody matched in the phoneme sequence are all divided and marked as canons. If there are multiple rhythms, then multiple rhythm names are marked to obtain a rhythm sequence.

[0091] For further details, please refer to... Figure 3 , Figure 3 As a specific embodiment of step 103 in this invention, step 103 includes the following specific implementation methods:

[0092] 1031. Read the rhythm sequence of the preset rhythm data;

[0093] 1032. Determine whether the phonological sequence has a matching sequence in the phoneme sequence;

[0094] 1033. If they exist, the matching sequences in the phoneme sequence are represented.

[0095] 1034. If it does not exist, then read another rhythm sequence of the preset rhythm data.

[0096] In steps 1031-1034, the canon melody “DA-Bm-F#-GDGA” is read first. The rhythmic sequence “DA-Bm-F#-GDGA” is used to match the phoneme sequence. The matched sequence is marked. If it does not exist, another rhythmic melody such as the natural major key “CDEFGABC” or the natural minor key “ABCDEFGA” is searched in the rhythmic scale.

[0097] 104. Using a preset description library, perform description matching processing on the phonological sequence to obtain the description text corresponding to the audio data;

[0098] In this embodiment, for example, if the musical sequence has the playback order of "Canon - Ring Scale - Lightning Scale - Canon", and all of them are played using the suona, then the text description first uses the description instrument as the suona, and then combines the text describing "Canon", "Ring Scale", "Lightning Scale", and "Canon" according to the order to obtain the description text.

[0099] For details, please refer to Figure 4 , Figure 4 This is a specific embodiment of step 104 in the present invention, which includes the following specific implementation methods:

[0100] 1041. In the musical sequence, select the target musical name;

[0101] 1042. Match the description field corresponding to the target musical note name in the preset description library, and establish a mapping relationship between the description field and the musical note sequence;

[0102] 1043. Based on the musical rhythm sorting of the musical rhythm sequence, combine the description fields corresponding to all musical rhythm names in the musical rhythm sequence to generate the description text corresponding to the audio data.

[0103] In steps 1041-1043, based on the target musical scale name "Canon-Circular Scale-Lightning Scale-Canon", the description fields corresponding to "Canon", "Circular Scale", "Lightning Scale", and "Canon" are matched from the preset description library, generating a mapping relationship between all musical scale names and description fields. By sorting "Canon-Circular Scale-Lightning Scale-Canon" in the description fields, all description fields are combined to obtain the description text corresponding to the audio data.

[0104] 105. When the audio data is played, the text field corresponding to the playback position in the description text is displayed based on the playback position of the audio data.

[0105] In this embodiment, when playing audio data, based on the playback position of "Canon-Circular Scale-Lightning Scale-Canon", such as the playback position of "Lightning Scale", the descriptive text corresponding to "Lightning Scale" is displayed.

[0106] In this embodiment of the invention, by using a phoneme recognition process to identify various beats and instrument types in audio data, the identified phonemes are arranged in chronological order. The rhythm and melody are analyzed to obtain a rhythm order and are labeled with names. In a text description database, relevant text description documents are found, and then the text data corresponding to the audio data is combined based on the text description fields. This solves the technical problem that the analysis technology lacks rhythm analysis of audio and cannot meet people's needs for appreciating audio music.

[0107] Figure 5 This is a schematic diagram of the structure of an audio data rhythm analysis and playback device 500 provided in an embodiment of the present invention. The audio data rhythm analysis and playback device 500 can vary significantly due to different configurations or performance. It may include one or more central processing units (CPUs) 510 (e.g., one or more processors) and a memory 520, and one or more storage media 530 (e.g., one or more mass storage devices) storing application programs 533 or data 532. The memory 520 and storage media 530 can be temporary or persistent storage. The program stored in the storage media 530 may include one or more modules (not shown in the diagram), each module may include a series of instruction operations on the audio data rhythm analysis and playback device 500. Furthermore, the processor 510 may be configured to communicate with the storage media 530 and execute the series of instruction operations in the storage media 530 on the audio data rhythm analysis and playback device 500.

[0108] The audio data-based rhythm analysis playback device 500 may also include one or more power supplies 540, one or more wired or wireless network interfaces 550, one or more input / output interfaces 560, and / or one or more operating systems 531, such as Windows Server, Mac OS X, Unix, Linux, Free BSD, etc. Those skilled in the art will understand that... Figure 5 The illustrated structure of the audio data rhythm analysis playback device does not constitute a limitation on audio data rhythm analysis playback devices, and may include more or fewer components than illustrated, or combine certain components, or have different component arrangements.

[0109] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when the instructions are executed on a computer, cause the computer to perform the steps of the audio data rhythm analysis and playback method.

[0110] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0111] Furthermore, although the operations are described in a specific order, this should be understood as requiring that such operations be performed in the specific order shown or in sequential order, or requiring that all illustrated operations be performed to achieve the desired result. In certain environments, multitasking and parallel processing may be advantageous. Similarly, although several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of this disclosure. Certain features described in the context of individual embodiments may also be implemented in combination in a single implementation. Conversely, various features described in the context of a single implementation may also be implemented individually or in any suitable sub-combination in multiple implementations.

[0112] Although the subject matter has been described using language specific to structural features and / or methodological logic, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. Rather, the specific features and actions described above are merely illustrative examples of implementing the claims.

Claims

1. A method for analyzing and playing audio data based on its rhythm, characterized in that, Including the following steps: Receive audio data; The audio data is processed by a preset phoneme recognition algorithm to obtain a phoneme sequence. Based on the preset rhythm data, the phoneme sequence is divided and marked to obtain the rhythm sequence; Using a preset description library, the rhythm sequence is subjected to description matching processing to obtain the description text corresponding to the audio data; When the audio data is played, the text field corresponding to the playback position in the description text is displayed based on the playback position of the audio data. The step of performing phoneme recognition processing on the audio data according to a preset phoneme recognition algorithm to obtain a phoneme sequence includes: According to the preset frame window, the audio data is decomposed into N frames of audio, where N is a positive integer; Vectorize N frames of audio data to obtain N frame vectors; Based on the preset set of convolutional matrices, convolution processing is performed on the N frame vectors respectively to obtain N convolutional frame vectors; The N convolutional frame vectors are normalized to obtain N normalized frame vectors; According to the preset activation function, the N normalized frame vectors are activated to obtain N activation values; Generate a phoneme sequence based on N activation values; The vectorization process of N frames of audio data to obtain N frame vectors includes: Perform sequence convolution on the frame audio data to obtain convolutional frame data; The convolutional frame data is averaged at different time points to obtain the frame vector; The step of performing sequence convolution processing on the frame audio data to obtain convolutional frame data includes: Based on a preset first convolution kernel, the frame data of t-2, t-1, t, t+1, and t+2 in the range of the frame audio data T are extracted and convolved to obtain the first convolution sub-audio. Based on the preset second convolution kernel, the frame data of t-2, t, and t+2 in the range of the frame audio data T are extracted and convolved to obtain the second convolution sub-audio. Based on a preset third convolution kernel, the frame data t in the range of the frame audio data T is extracted and convolved to obtain the third convolution sub-audio. The first convolutional sub-audio, the second convolutional audio, and the third convolutional audio are concatenated in parallel to obtain convolutional frame data; The step of using a preset description library to perform description matching processing on the phonological sequence to obtain the description text corresponding to the audio data includes: Select the target musical note name from the musical note sequence; Match the description field corresponding to the target musical note name in the preset description library, and establish a mapping relationship between the description field and the musical note sequence; Based on the musical rhythm order, the description fields corresponding to all musical rhythm names in the musical rhythm sequence are combined to generate the description text corresponding to the audio data.

2. The method for analyzing and playing audio data according to claim 1, characterized in that, The step of performing time-frame averaging on the convolutional frame data to obtain the frame vector includes: Where H is the average time frame, h t1 h is the audio value of the first convolutional sub-audio at time t. t2 h is the audio value of the second convolutional sub-audio at time t. t3 is the audio value of the third convolutional sub-audio at time t, where T is the end time of the frame audio data; A frame vector is generated based on the average time frame.

3. The method for analyzing and playing audio data according to claim 1, characterized in that, The activation process performed on the N normalized frame vectors according to the preset activation function to obtain N activation values ​​includes: Based on the softmax activation function, the N normalized frame vectors are activated respectively to obtain N activation values.

4. The method for analyzing and playing audio data according to claim 1, characterized in that, The step of dividing and marking the phoneme sequence according to the preset rhythm data to obtain the rhythm sequence includes: Read the rhythm sequence from the preset rhythm data; Determine whether the phonological sequence has a matching sequence in the phoneme sequence; If they exist, the matching sequences in the phoneme sequence are represented. If it does not exist, another rhythm sequence of the preset rhythm data is read.

5. An audio data rhythm analysis and playback device, characterized in that, The audio data rhythm analysis and playback device includes: a memory and at least one processor, wherein the memory stores instructions, and the memory and the at least one processor are interconnected via a line; The at least one processor invokes the instructions in the memory to cause the audio data rhythm analysis playback device to execute the audio data rhythm analysis playback method as described in any one of claims 1-4.

6. A computer-readable storage medium storing a computer program thereon, characterized in that, When the computer program is executed by the processor, it implements the method for rhythmic analysis and playback of audio data as described in any one of claims 1-4.

Images