CN112634841A - Guitar music automatic generation method based on voice recognition - Google Patents
Guitar music automatic generation method based on voice recognition Download PDFInfo
- Publication number
- CN112634841A CN112634841A CN202011392002.1A CN202011392002A CN112634841A CN 112634841 A CN112634841 A CN 112634841A CN 202011392002 A CN202011392002 A CN 202011392002A CN 112634841 A CN112634841 A CN 112634841A
- Authority
- CN
- China
- Prior art keywords
- music
- chord
- guitar
- information
- playing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 33
- 230000033764 rhythmic process Effects 0.000 claims abstract description 23
- 238000009432 framing Methods 0.000 claims abstract description 7
- 238000003062 neural network model Methods 0.000 claims description 15
- 238000013528 artificial neural network Methods 0.000 claims description 14
- 238000010408 sweeping Methods 0.000 claims description 10
- 238000012549 training Methods 0.000 claims description 10
- 238000001228 spectrum Methods 0.000 claims description 5
- 239000011295 pitch Substances 0.000 description 33
- 210000005069 ears Anatomy 0.000 description 7
- 238000000354 decomposition reaction Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 5
- 238000006748 scratching Methods 0.000 description 5
- 230000002393 scratching effect Effects 0.000 description 5
- 238000013135 deep learning Methods 0.000 description 4
- 239000000203 mixture Substances 0.000 description 4
- 230000004304 visual acuity Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10H—ELECTROPHONIC MUSICAL INSTRUMENTS; INSTRUMENTS IN WHICH THE TONES ARE GENERATED BY ELECTROMECHANICAL MEANS OR ELECTRONIC GENERATORS, OR IN WHICH THE TONES ARE SYNTHESISED FROM A DATA STORE
- G10H1/00—Details of electrophonic musical instruments
- G10H1/0008—Associated control or indicating means
- G10H1/0025—Automatic or semi-automatic music composition, e.g. producing random music, applying rules from music theory or modifying a musical piece
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/03—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters
- G10L25/24—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters the extracted parameters being the cepstrum
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/27—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the analysis technique
- G10L25/30—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the analysis technique using neural networks
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/90—Pitch determination of speech signals
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10H—ELECTROPHONIC MUSICAL INSTRUMENTS; INSTRUMENTS IN WHICH THE TONES ARE GENERATED BY ELECTROMECHANICAL MEANS OR ELECTRONIC GENERATORS, OR IN WHICH THE TONES ARE SYNTHESISED FROM A DATA STORE
- G10H2210/00—Aspects or methods of musical processing having intrinsic musical character, i.e. involving musical theory or musical parameters or relying on musical knowledge, as applied in electrophonic musical tools or instruments
- G10H2210/031—Musical analysis, i.e. isolation, extraction or identification of musical elements or musical parameters from a raw acoustic signal or from an encoded audio signal
- G10H2210/056—Musical analysis, i.e. isolation, extraction or identification of musical elements or musical parameters from a raw acoustic signal or from an encoded audio signal for extraction or identification of individual instrumental parts, e.g. melody, chords, bass; Identification or separation of instrumental parts by their characteristic voices or timbres
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10H—ELECTROPHONIC MUSICAL INSTRUMENTS; INSTRUMENTS IN WHICH THE TONES ARE GENERATED BY ELECTROMECHANICAL MEANS OR ELECTRONIC GENERATORS, OR IN WHICH THE TONES ARE SYNTHESISED FROM A DATA STORE
- G10H2210/00—Aspects or methods of musical processing having intrinsic musical character, i.e. involving musical theory or musical parameters or relying on musical knowledge, as applied in electrophonic musical tools or instruments
- G10H2210/031—Musical analysis, i.e. isolation, extraction or identification of musical elements or musical parameters from a raw acoustic signal or from an encoded audio signal
- G10H2210/066—Musical analysis, i.e. isolation, extraction or identification of musical elements or musical parameters from a raw acoustic signal or from an encoded audio signal for pitch analysis as part of wider processing for musical purposes, e.g. transcription, musical performance evaluation; Pitch recognition, e.g. in polyphonic sounds; Estimation or use of missing fundamental
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10H—ELECTROPHONIC MUSICAL INSTRUMENTS; INSTRUMENTS IN WHICH THE TONES ARE GENERATED BY ELECTROMECHANICAL MEANS OR ELECTRONIC GENERATORS, OR IN WHICH THE TONES ARE SYNTHESISED FROM A DATA STORE
- G10H2210/00—Aspects or methods of musical processing having intrinsic musical character, i.e. involving musical theory or musical parameters or relying on musical knowledge, as applied in electrophonic musical tools or instruments
- G10H2210/101—Music Composition or musical creation; Tools or processes therefor
- G10H2210/111—Automatic composing, i.e. using predefined musical rules
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/90—Pitch determination of speech signals
- G10L2025/906—Pitch tracking
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Signal Processing (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Theoretical Computer Science (AREA)
- Auxiliary Devices For Music (AREA)
Abstract
The invention discloses a guitar music automatic generation method based on voice recognition, which comprises the following steps of: step 201, performing framing operation on an input guitar playing audio file to be identified, and extracting the Mel cepstrum coefficient characteristics of each frame; step 202, calculating the classification of the Mel cepstrum coefficient characteristics of each frame in the step 201, and obtaining the pitch and playing mode information of each frame of the tested audio according to the output layer label of the model; step 203, analyzing chord, tone and rhythm information of the music based on the pitch information in the music score file in the step 202; and step 204, integrating the pitch, playing mode, chord, tone and rhythm information output in the steps 202 and 203, and outputting a music score file. The method can automatically generate the guitar music, replaces manual music picking work, and effectively improves the work efficiency.
Description
Technical Field
The invention belongs to the technical field of sound, and particularly relates to a music score generation method.
Background
In the field of music score, the current technical research mainly focuses on directions of music retrieval, music performance evaluation, automatic composition and the like; the guitar music is processed by focusing on the picture music, and music score is generated by identifying the picture music, and automatic playing and playing can be performed.
In the actual guitar learning process, people often encounter guitar playing of interest, but no music score in any form can be referred to, and in this case, the music score is determined according to the played sound. In reality, the process is generally called as music score scratching, repeated listening and comparison are needed for music score scratching work, and finally the tone, rhythm and chord configuration and decomposition or chord sweeping details of the music score are determined; the staffs taking music must train for a long time to have sufficient ability to distinguish the ears, and also have to master sufficient music theory and deep knowledge of guitar principle and playing technique, so to speak, it is the high-level skill in guitar learning.
It is difficult for the ordinary guitar fans to do. If the guitar music can be directly generated from the playing sound, the learning difficulty of a beginner can be reduced, convenience can be provided for professionals, and only the music score generated automatically needs to be optimized.
The guitar is a stringed musical instrument, the vibration of strings gives out sound, the vibration frequency of the strings determines the pitch, the vibration frequency of the strings is influenced by the diameters and the vibration lengths of the strings, the different diameters of the strings are different, the different grades on the fingerboard determine the vibration lengths of the strings, and therefore the combination of the strings and the grades determines the pitch of the playing. In addition, the guitar belongs to the standard tuning musical instrument, and the standard tones of one string to six strings are E, B, G, D, A, E in turn, so that different strings and taste combinations may give the same pitch.
The chord is a musical concept and is composed of a group of sounds satisfying a certain interval relationship, wherein the lowest pitch of the sounds is a root and is the basis of the chord, and the intervals between different tones determine the color of the chord, which jointly form a chord. The chord is played mainly by two types of chord decomposition and chord sweeping, the chord decomposition is to pop out the chord in turn to form the sound, and the chord sweeping sweeps a plurality of strings simultaneously to produce a plurality of chord tones.
The playing of the guitar can be divided into accompaniment and solo according to the function, the accompaniment generally only plays the chord, provides harmony for the melody, and the solo plays the melody while playing the chord. Whether the music accompaniment or the solo, the pitch of each note is firstly determined when the music score is played, and if the music score is swept, the chord and the swept string are determined, which all need to be distinguished by human ears.
With the development of artificial intelligence technology, sound processing based on deep learning has a great progress, and the method achieves ideal effects in scenes such as voice recognition, scene classification and the like, and is widely applied. The strong learning ability of the neural network is utilized, the work of identifying the pitch of the note can be completed instead of human ears, all the sounds which can be played by the guitar are used as parameters of training data to train a neural network model, unknown sounds are classified by using the model, the pitch information in playing can be determined, and the first step of scratching the music score is also completed.
After the pitch of the music is determined, the musical instruments are divided into bars, and the chord used by each bar is further determined, usually the root of the chord is played first, and the first beat of the bar is the repeated beat, and then the chord trend of the music is determined by combining the combined sound information in the chord.
Finally, the key and the rhythm of the music are determined, the key is generally determined by the composition of the chord, and the rhythm reflects the weight distribution and can be determined by analyzing the energy distribution of the music.
Disclosure of Invention
The invention aims to improve the prior technical problem, namely the invention aims to provide a guitar spectrum automatic generation method based on voice recognition. The technical scheme of the invention is as follows:
a guitar music automatic generation method based on voice recognition is characterized by comprising a process of picking up music of a guitar playing audio file to be recognized, wherein the process comprises the following steps:
step 201, performing framing operation on an input guitar playing audio file to be identified, and extracting the Mel cepstrum coefficient characteristics of each frame;
step 202, calculating the classification of the Mel cepstrum coefficient characteristics of each frame in the step 201, and obtaining the pitch and playing mode information of each frame of the tested audio according to the output layer label of the model;
step 203, analyzing chord, tone and rhythm information of the music based on the pitch information in the music score file in the step 202;
and step 204, integrating the pitch, the playing mode, the chord, the tone, the rhythm and the information output in the steps 202 and 203, and outputting a music score file.
Further, the step 203 comprises the following sub-steps:
step 2031, determining the positions of all chord root sounds in the music, calculating the time intervals, selecting the interval duration with the most frequent occurrence, preliminarily dividing the music into uniform time segments, comparing the sounds appearing in each time segment with the chord template, calculating the similarity, selecting the chord with the highest similarity as the chord of the time segment, and writing the chord information into the music score file;
step 2032, comparing the chord tables of the guitar according to the chord information obtained in step 2031, comparing the number of the neutralized strings in the music in each tune, selecting the tune with the most number as the tonality of the music, and writing the tonality in the music file;
step 2033, calculating the energy of each chord root, calculating the average value of all the energies as a threshold, selecting the root with the energy larger than the threshold, calculating the time interval between adjacent roots, and selecting the most frequently occurring time interval as the bar of the music; calculating the number of notes of each measure according to the duration of the measure, taking the root as accent, selecting notes with the energy close to that of the root from other notes as accent beats, determining which one of four beats, four beats and three beats or eight beats is selected to obtain rhythm information, and writing the rhythm information into a music score file;
step 2034, calculating the duration of a quarter note according to the bar duration and the rhythm information to obtain the playing speed of the music, and writing the playing speed into the music file.
Further, in step 201, the classification to which the mel-frequency cepstrum coefficient feature of each frame belongs is calculated based on a deep neural network model, and the building of the deep neural network model includes the following steps:
step 101, collecting guitar playing sound information by using a recording device, wherein the sound information comprises the grade string plucking sound of each string and all chord sweeping sounds, and generating an audio file;
102, marking the collected guitar sound information, wherein the marking content is a pitch and a playing mode, the pitch may include one or more than one, and the playing mode is divided into string plucking and string sweeping;
103, performing framing operation on the audio file in the step 101, and extracting the Mel cepstrum coefficient characteristics of each frame;
and step 104, training the deep neural network by using a back propagation algorithm based on the characteristic file obtained in the step 103 and the labeled file obtained in the step 102 to obtain a deep neural network model.
Based on the automatic generation method of the guitar music, for guitar playing without picture music, the guitar music can be automatically generated, manual music picking work is replaced, and the working efficiency can be effectively improved for professionals; the method reduces the difficulty of picking up the music score, even a beginner can independently finish the music score picking, and the guitar learning work is convenient. In addition, the method utilizes a deep learning technology, training data can be continuously added according to the characteristics of the neural network aiming at the weaknesses of the neural network, the resolving power exceeding that of human ears is finally obtained, and the method is finally greatly superior to the traditional method in effect and efficiency by combining the strong computing power of a computer.
Detailed Description
The following description is presented to disclose the invention so as to enable any person skilled in the art to practice the invention. The preferred embodiments in the following description are given by way of example only, and other obvious variations will occur to those skilled in the art.
For ease of understanding, before describing embodiments of the present invention, the features of the guitar sound will be described first.
The guitar is a stringed musical instrument, the vibration of strings gives out sound, the vibration frequency of the strings determines the pitch, the vibration frequency of the strings is influenced by the diameters and the vibration lengths of the strings, the different diameters of the strings are different, the different grades on the fingerboard determine the vibration lengths of the strings, and therefore the combination of the strings and the grades determines the pitch of the playing. In addition, the guitar belongs to the standard tuning musical instrument, and the standard tones of one string to six strings are E, B, G, D, A, E in turn, so that different strings and taste combinations may give the same pitch.
The chord is a musical concept and is composed of a group of sounds satisfying a certain interval relationship, wherein the lowest pitch of the sounds is a root and is the basis of the chord, and the intervals between different tones determine the color of the chord, which jointly form a chord. The chord is played mainly by two types of chord decomposition and chord sweeping, the chord decomposition is to pop out the chord in turn to form the sound, and the chord sweeping sweeps a plurality of strings simultaneously to produce a plurality of chord tones.
The playing of the guitar can be divided into accompaniment and solo according to the function, the accompaniment generally only plays the chord, provides harmony for the melody, and the solo plays the melody while playing the chord. Whether the music accompaniment or the solo, the pitch of each note is firstly determined when the music score is played, and if the music score is swept, the chord and the swept string are determined, which all need to be distinguished by human ears.
The invention utilizes the strong learning ability of the neural network to replace the human ear to finish the work of identifying the pitch of the note, takes all the sounds played by the guitar as the parameters of training data to train the neural network model, and uses the model to classify the unknown sounds, so that the pitch information in playing can be determined, and the first step of scratching the music score is finished. After the pitch of the music is determined, the musical instruments are divided into bars, and the chord used by each bar is further determined, usually the root of the chord is played first, and the first beat of the bar is the repeated beat, and then the chord trend of the music is determined by combining the combined sound information in the chord. Finally, the key and the rhythm of the music are determined, the key is generally determined by the composition of the chord, and the rhythm reflects the weight distribution and can be determined by analyzing the energy distribution of the music.
The guitar spectrum automatic generation method based on the voice recognition provided by the embodiment of the invention comprises two steps which are sequentially executed: and establishing a neural network model, and performing score scratching on the guitar playing audio file to be identified.
The neural network model building method comprises the following steps:
step 101, collecting guitar playing sound information by using a recording device, wherein the sound information comprises the grade string-plucking sound of each string and all chord string-sweeping sounds, and generating an audio file. Here, the guitar playing sound may come from a finger or a pick, and the audio file is preferably saved in 16k16bit PCM format.
And 102, marking the collected sound information of the guitar, wherein the marking content is a pitch and a playing mode, the pitch may contain one or more than one, and the playing mode is divided into string plucking and string sweeping. The output result of this step is a label file in text format.
And 103, performing framing operation on the audio file in the step 101, extracting the Mel cepstrum coefficient characteristics of each frame, wherein the output result of the step is a binary characteristic file.
And step 104, training the deep neural network by using a back propagation algorithm based on the characteristic file obtained in the step 103 and the labeled file obtained in the step 102 to obtain a deep neural network model. The output result of this step is a binary model file.
Establishing the deep neural network model, namely training the neural network model by using guitar playing sound and labels by using a deep learning technology; the characteristics of the vibration of different pitches on the frequency domain are learned by converting the audio frequency of the time domain into the frequency domain and extracting relevant characteristics and utilizing the strong learning capacity of the neural network, so that the human ear can be replaced, and the work of distinguishing different pitches is completed. According to the characteristics of the neural network, training data can be continuously added aiming at the weaknesses of the neural network, the resolving power exceeding that of human ears is finally obtained, and the effect and the efficiency are greatly superior to those of the traditional method by combining the strong computing power of a computer.
The method specifically comprises the following steps of:
step 201, performing framing operation on an input guitar playing audio file to be identified, and extracting the mel cepstrum coefficient characteristics of each frame. The output result of this step is a binary signature file.
Step 202, calculating the classification of the Mel cepstrum coefficient characteristics of each frame in the step 201 by using the deep neural network model generated in the step 104, obtaining the pitch and the playing mode information of each frame of the tested audio according to the output layer label of the model, newly building a binary music score file, and writing the corresponding pitch information;
step 203, analyzing the chord, tone and rhythm information of the music based on the pitch information in the music file in step 202, including the following steps:
step 2031, determining the positions of all chord root sounds in the music, calculating the time intervals, selecting the interval duration with the most frequent occurrence, preliminarily dividing the music into uniform time segments, comparing the sounds appearing in each time segment with the chord template, calculating the similarity, selecting the chord with the highest similarity as the chord of the time segment, and writing the chord information into the music score file;
step 2032, comparing the chord tables of the guitar according to the chord information obtained in step 2031, comparing the number of the neutralized strings in the music in each tune, selecting the tune with the most number as the tonality of the music, and writing the tonality in the music file;
step 2033, calculating the energy of each chord root, calculating the average value of all the energies as a threshold, selecting the root with the energy larger than the threshold, calculating the time interval between adjacent roots, and selecting the most frequently occurring time interval as the bar of the music; calculating the number of notes of each measure according to the duration of the measure, taking the root as accent, selecting notes with the energy close to that of the root from other notes as accent beats, determining which one of four beats, four beats and three beats or eight beats is selected to obtain rhythm information, and writing the rhythm information into a music score file;
step 2034, calculating the duration of a quarter note according to the bar duration and the rhythm information to obtain the playing speed of the music, and writing the playing speed into a music score file;
and 204, integrating the previous pitch, playing mode, chord, tone and rhythm information, and outputting a music score file in a text form.
The automatic guitar music generation method directly generates the guitar music from the playing sound of the guitar, and has the following characteristics and effects:
1. compared with a method for obtaining music scores from picture music, the method can automatically generate the guitar music for playing the guitar without the picture music, replaces the manual music picking work, and can effectively improve the working efficiency for professionals; the method reduces the difficulty of picking up the music score, even a beginner can independently finish the music score picking, and the guitar learning work is convenient.
2. Training a neural network model by using guitar playing sound and labels by utilizing a deep learning technology; the characteristics of the vibration of different pitches on the frequency domain are learned by converting the audio frequency of the time domain into the frequency domain and extracting relevant characteristics and utilizing the strong learning capacity of the neural network, so that the human ear can be replaced, and the work of distinguishing different pitches is completed. According to the characteristics of the neural network, training data can be continuously added aiming at the weaknesses of the neural network, the resolving power exceeding that of human ears is finally obtained, and the effect and the efficiency are greatly superior to those of the traditional method by combining the strong computing power of a computer.
3. After the neural network identifies the pitch information of each frame of audio, the information of chord distribution, tone, rhythm, speed and the like of the music score is obtained through analysis by combining the pitch composition of the chord and the characteristics of guitar playing. These are all done automatically, do not need manual participation, help the guitar fan to do guitar study faster better.
The foregoing shows and describes the general principles, essential features, and advantages of the invention. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, which are given by way of illustration of the principles of the present invention, and that various changes and modifications may be made without departing from the spirit and scope of the invention. Such changes and modifications are intended to be within the scope of the claimed invention. The scope of the invention is defined by the appended claims and equivalents thereof.
Claims (4)
1. A guitar music automatic generation method based on voice recognition is characterized by comprising a process of picking up music of a guitar playing audio file to be recognized, wherein the process comprises the following steps:
step 201, performing framing operation on an input guitar playing audio file to be identified, and extracting the Mel cepstrum coefficient characteristics of each frame;
step 202, calculating the classification of the Mel cepstrum coefficient characteristics of each frame in the step 201, and obtaining the pitch and playing mode information of each frame of the tested audio according to the output layer label of the model;
step 203, analyzing chord, tone and rhythm information of the music based on the pitch information in the music score file in the step 202;
and step 204, integrating the pitch, playing mode, chord, tone and rhythm information output in the steps 202 and 203, and outputting a music score file.
2. A method for automatic generation of guitar spectrum based on voice recognition according to claim 1, characterized by the step 203 comprising the sub-steps of:
step 2031, determining the positions of all chord root sounds in the music, calculating the time intervals, selecting the interval duration with the most frequent occurrence, preliminarily dividing the music into uniform time segments, comparing the sounds appearing in each time segment with the chord template, calculating the similarity, selecting the chord with the highest similarity as the chord of the time segment, and writing the chord information into the music score file;
step 2032, comparing the chord tables of the guitar according to the chord information obtained in step 2031, comparing the number of the neutralized strings in the music in each tune, selecting the tune with the most number as the tonality of the music, and writing the tonality in the music file;
step 2033, calculating the energy of each chord root, calculating the average value of all the energies as a threshold, selecting the root with the energy larger than the threshold, calculating the time interval between adjacent roots, and selecting the most frequently occurring time interval as the bar of the music; calculating the number of notes of each measure according to the duration of the measure, taking the root as accent, selecting notes with the energy close to that of the root from other notes as accent beats, determining which one of four beats, four beats and three beats or eight beats is selected to obtain rhythm information, and writing the rhythm information into a music score file;
step 2034, calculating the duration of a quarter note according to the bar duration and the rhythm information to obtain the playing speed of the music, and writing the playing speed into the music file.
3. The method for automatically generating guitar spectrum based on voice recognition according to claim 1 or 2, wherein in the step 201, the classification to which the mel cepstral coefficient feature of each frame belongs is calculated based on a deep neural network model, and the deep neural network model is established by the following steps:
step 101, collecting guitar playing sound information by using a recording device, wherein the sound information comprises the grade string plucking sound of each string and all chord sweeping sounds, and generating an audio file;
102, marking the collected guitar sound information, wherein the marking content is pitch and playing mode;
103, performing framing operation on the audio file in the step 101, and extracting the Mel cepstrum coefficient characteristics of each frame;
and step 104, training the deep neural network by using a back propagation algorithm based on the characteristic file obtained in the step 103 and the labeled file obtained in the step 102 to obtain a deep neural network model.
4. The method of automatic generation of guitar spectrum based on voice recognition according to claim 3, characterized in that the guitar playing voice information comes from finger or pick playing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011392002.1A CN112634841B (en) | 2020-12-02 | 2020-12-02 | Guitar music automatic generation method based on voice recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011392002.1A CN112634841B (en) | 2020-12-02 | 2020-12-02 | Guitar music automatic generation method based on voice recognition |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112634841A true CN112634841A (en) | 2021-04-09 |
| CN112634841B CN112634841B (en) | 2022-11-29 |
Family
ID=75307435
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011392002.1A Active CN112634841B (en) | 2020-12-02 | 2020-12-02 | Guitar music automatic generation method based on voice recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112634841B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113707112A (en) * | 2021-08-13 | 2021-11-26 | 陕西师范大学 | Recursive jump connection deep learning music automatic generation method based on layer standardization |
| CN113763913A (en) * | 2021-09-16 | 2021-12-07 | 腾讯音乐娱乐科技（深圳）有限公司 | Music score generation method, electronic device and readable storage medium |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040231498A1 (en) * | 2003-02-14 | 2004-11-25 | Tao Li | Music feature extraction using wavelet coefficient histograms |
| CN102723079A (en) * | 2012-06-07 | 2012-10-10 | 天津大学 | Music and chord automatic identification method based on sparse representation |
| JP3185768U (en) * | 2013-05-23 | 2013-09-05 | 二郎 須澤 | Score for shakuhachi |
| CN103714806A (en) * | 2014-01-07 | 2014-04-09 | 天津大学 | Chord recognition method combining SVM with enhanced PCP |
| CN104992712A (en) * | 2015-07-06 | 2015-10-21 | 成都云创新科技有限公司 | Music reorganization-based music score automatic formation method |
| CN106205570A (en) * | 2016-08-10 | 2016-12-07 | 秦桂芳 | A kind of music score recording method |
| US10008188B1 (en) * | 2017-01-31 | 2018-06-26 | Kyocera Document Solutions Inc. | Musical score generator |
| CN110310621A (en) * | 2019-05-16 | 2019-10-08 | 平安科技（深圳）有限公司 | Sing synthetic method, device, equipment and computer readable storage medium |
| CN111898753A (en) * | 2020-08-05 | 2020-11-06 | 字节跳动有限公司 | Music transcription model training method, music transcription method and corresponding device |
-
2020
- 2020-12-02 CN CN202011392002.1A patent/CN112634841B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040231498A1 (en) * | 2003-02-14 | 2004-11-25 | Tao Li | Music feature extraction using wavelet coefficient histograms |
| CN102723079A (en) * | 2012-06-07 | 2012-10-10 | 天津大学 | Music and chord automatic identification method based on sparse representation |
| JP3185768U (en) * | 2013-05-23 | 2013-09-05 | 二郎 須澤 | Score for shakuhachi |
| CN103714806A (en) * | 2014-01-07 | 2014-04-09 | 天津大学 | Chord recognition method combining SVM with enhanced PCP |
| CN104992712A (en) * | 2015-07-06 | 2015-10-21 | 成都云创新科技有限公司 | Music reorganization-based music score automatic formation method |
| CN106205570A (en) * | 2016-08-10 | 2016-12-07 | 秦桂芳 | A kind of music score recording method |
| US10008188B1 (en) * | 2017-01-31 | 2018-06-26 | Kyocera Document Solutions Inc. | Musical score generator |
| CN110310621A (en) * | 2019-05-16 | 2019-10-08 | 平安科技（深圳）有限公司 | Sing synthetic method, device, equipment and computer readable storage medium |
| CN111898753A (en) * | 2020-08-05 | 2020-11-06 | 字节跳动有限公司 | Music transcription model training method, music transcription method and corresponding device |
Non-Patent Citations (1)
| Title |
|---|
| 王峰等: "基于CRFs和MPCP特征的和弦识别研究", 《计算机工程与应用》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113707112A (en) * | 2021-08-13 | 2021-11-26 | 陕西师范大学 | Recursive jump connection deep learning music automatic generation method based on layer standardization |
| CN113707112B (en) * | 2021-08-13 | 2024-05-28 | 陕西师范大学 | Automatic generation method of recursion jump connection deep learning music based on layer standardization |
| CN113763913A (en) * | 2021-09-16 | 2021-12-07 | 腾讯音乐娱乐科技（深圳）有限公司 | Music score generation method, electronic device and readable storage medium |
| WO2023040332A1 (en) * | 2021-09-16 | 2023-03-23 | 腾讯音乐娱乐科技（深圳）有限公司 | Method for generating musical score, electronic device, and readable storage medium |
| CN113763913B (en) * | 2021-09-16 | 2024-06-18 | 腾讯音乐娱乐科技（深圳）有限公司 | Music score generating method, electronic equipment and readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112634841B (en) | 2022-11-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Barbancho et al. | Automatic transcription of guitar chords and fingering from audio | |
| Lee et al. | Acoustic chord transcription and key extraction from audio using key-dependent HMMs trained on synthesized audio | |
| Barbedo et al. | Automatic genre classification of musical signals | |
| Hung et al. | Frame-level instrument recognition by timbre and pitch | |
| CN112382257B (en) | Audio processing method, device, equipment and medium | |
| Lee et al. | A Unified System for Chord Transcription and Key Extraction Using Hidden Markov Models. | |
| WO2013080210A1 (en) | Method for extracting representative segments from music | |
| CN112634841B (en) | Guitar music automatic generation method based on voice recognition | |
| JP2010054802A (en) | Unit rhythm extraction method from musical acoustic signal, musical piece structure estimation method using this method, and replacing method of percussion instrument pattern in musical acoustic signal | |
| CN113192471B (en) | Musical main melody track recognition method based on neural network | |
| Chordia | Segmentation and Recognition of Tabla Strokes. | |
| Paulus | Signal processing methods for drum transcription and music structure analysis | |
| EP2342708B1 (en) | Method for analyzing a digital music audio signal | |
| Ramirez et al. | Automatic performer identification in commercial monophonic jazz performances | |
| Kum et al. | Pseudo-label transfer from frame-level to note-level in a teacher-student framework for singing transcription from polyphonic music | |
| Lerch | Software-based extraction of objective parameters from music performances | |
| CN110134823B (en) | MIDI music genre classification method based on normalized note display Markov model | |
| Setragno et al. | Feature-based characterization of violin timbre | |
| Jensen et al. | Binary decision tree classification of musical sounds | |
| Nichols et al. | Automatically discovering talented musicians with acoustic analysis of youtube videos | |
| MA et al. | Four-way classification of tabla strokes with models adapted from Automatic Drum Transcription | |
| Zhao et al. | Transfer learning for violinist identification | |
| Camurri et al. | An experiment on analysis and synthesis of musical expressivity | |
| Gogineni et al. | Mridangam artist identification from taniavartanam audio | |
| Zhao et al. | Violinist identification using note-level timbre feature distributions |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |