CN105227763B - A kind of instrumental audio real time method for segmenting realized on Intelligent mobile equipment - Google Patents
A kind of instrumental audio real time method for segmenting realized on Intelligent mobile equipment Download PDFInfo
- Publication number
- CN105227763B CN105227763B CN201510549631.3A CN201510549631A CN105227763B CN 105227763 B CN105227763 B CN 105227763B CN 201510549631 A CN201510549631 A CN 201510549631A CN 105227763 B CN105227763 B CN 105227763B
- Authority
- CN
- China
- Prior art keywords
- audio
- frame
- real time
- signal
- segmenting
- 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.)
- Expired - Fee Related
Links
Abstract
The invention discloses a kind of instrumental audio real time method for segmenting realized on Intelligent mobile equipment, this method can realize the simultaneously operating of recording and segmentation on the mobile apparatus.In order to reach real-time effect, the present invention determines the presence of true trip point after Quick Pretreatment has been carried out by the short-time energy for calculating signal in fixed length window and the mutation for finding average energy value so that is treated as possibility in collection.For the audio signal of instruments sound, the present invention can be while user's shoegaze, the audio signal collected is split, and quickly shown, carrying out the deep level of processing such as audio identification, the semantic structuring information of acquisition, progress audio retrieval for user prepares.
Description
Technical field
The present invention relates to, more particularly to a kind of instrumental audio real time method for segmenting realized on Intelligent mobile equipment.
Background technology
Audio segmentation is the steps necessary for carrying out audio identification, obtaining semantic structuring information, carrying out audio retrieval.Tradition
Audio frequency splitting method be just to start to split audio after whole audio-frequency informations is obtained, that is, be not carried out recording and point
The simultaneously operating cut.Conventional audio dividing method is calculated at PC ends, because PC does not possess portability, is limited in intelligence
It is some on mobile communication equipment to need to carry out using and promoting for audio segmentation application.
The content of the invention
The defects of the technical problem to be solved in the present invention is to be directed in the prior art, there is provided one kind is in Intelligent mobile equipment
The instrumental audio real time method for segmenting of upper realization.
The technical solution adopted for the present invention to solve the technical problems is：A kind of musical instrument realized on Intelligent mobile equipment
Audio real time method for segmenting, comprises the following steps：
1) voice data is gathered：Obtain the voice data of musical instrument in real time by the sound pick-up outfit on mobile communication equipment；
2) the Jing Yin elimination in front end：Eliminate the quiet data of beginning in the voice data of collection；
3) voice data after elimination quiet data is handled in real time, specifically included：
3.1) preemphasis digital filtering：Preemphasis processing sum is carried out to the audio signal by the Jing Yin Processing for removing in front end
Word filters；
3.2) framing：Continuous framing is seamlessly transitted using alternately framing method；Assuming that the length wlen of a frame,
Step-length is inc, then the repeating signal length of adjacent two frame is wlen-inc；
3.3) short-time energy is calculated：To each frame signal after framing, short-time energy is calculated：Wherein join
Number n is length of window, and x (m) is signal sequence.
3.4) background sound eliminates：If frame signal energy value is less than 0.1 times of audio signal ceiling capacity, the frame signal is
Background sound, background sound is deleted；
Wherein, the audio signal ceiling capacity of audio signal ceiling capacity at that time, and audio signal ceiling capacity is real-time
Renewal；
3.5) cut-point is inquired about：To the average energy value per adjacent 3 frameIt is compared, if there is larger gap, then recognizes
Catastrophe point to be present, catastrophe point is cut-point, that is, is met：Wherein, Th is the threshold value of adjacent 3 frame before and after judging
Coefficient；
4) the instrumental audio Real-time segmentation that the position according to where cut-point is realized.
By such scheme, original audio signal front end amplitude is set in the step 2) less than 10-4Signal be Jing Yin
By such scheme, in the step 3.1), preemphasis processing is carried out to sample according to 6dB/oct ratio；Numeral
Filtering uses below equation：Between y (n)=x (n)-a*x (n-1), wherein parameter a values are 0.9~1, before x (n) is filtering
Sampled signal, y (n) are filtered signal.
By such scheme, in the step 3.2), framing strategy is as follows：Frame length L=m × 256 are set, wherein m is 1~16
In arbitrary integer, frame walk length takes 0.3~0.7 times of frame length L.
By such scheme, in the step 3.5), between Th spans are 3~7.
The beneficial effect comprise that：The inventive method is directed to the instrumental audio Real-time segmentation of mobile terminal, is adapted to move
Moved end is handled, and can reach the advantages of speed is fast and consumption resource is few, can realize and the cutting operation in recording is carried out to audio-source, non-
Often it is applied to audio Real-time segmentation class application.
Brief description of the drawings
Below in conjunction with drawings and Examples, the invention will be further described, in accompanying drawing：
Fig. 1 is the method flow diagram of the embodiment of the present invention；
Fig. 2 is the flow chart that short-time energy is calculated in the embodiment of the present invention；
Fig. 3 is the flow chart that cut-point is inquired about in the embodiment of the present invention.
Embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, with reference to embodiments, to the present invention
It is further elaborated.It should be appreciated that specific embodiment described herein is not used to limit only to explain the present invention
The fixed present invention.
As shown in figure 1, a kind of instrumental audio real time method for segmenting realized on Intelligent mobile equipment, including following step
Suddenly：
1) voice data is gathered：The voice data that musical instrument plays is obtained by the sound pick-up outfit on mobile communication equipment in real time；
Sample frequency is 44.1KHz or 22.1KHz, and audio signal files form is mp3.
2) the Jing Yin elimination in front end：Eliminate the quiet data of beginning in the voice data of collection；In sampling process, by
A period of time is needed in recording preparation, causing the front end of sampled data to have does not have musical sounding in some time, that is, the standard played
In the standby stage, the background sound that the amplitude that this section gathers is similar to 0 is exactly quiet data.Obviously this part dumb data should be eliminated,
To remove redundancy as calculating speed can be improved.The system thinks that original audio signal front end amplitude is less than 10-4Signal be quiet
Sound, under the required user environment of the present invention, such quiet data may be considered completely ineffective interference data, should be all
Deleted.
3) voice data after elimination quiet data is handled in real time, specifically included：
3.1) preemphasis digital filtering：To carrying out digital filtering by the audio signal of the Jing Yin Processing for removing in front end；Due to
Audio signal has decay in transmitting procedure, therefore carries out preemphasis processing to sample according to 6dB/oct ratio so that defeated
Going out level mutually will not produce relatively large deviation, while can also eliminate low-frequency disturbance.Formula is：Y (n)=x (n)-a*x (n-1), ginseng
Number between number a values 0.9~1.
3.2) framing：Audio sample attribute changes more gently, within the time of this smooth variation in the short period of time
The audio frequency characteristics extracted keep basicly stable.Therefore, the present invention is carried out smooth using alternately framing method to continuous framing
Transition, to keep its continuity.Frame length L=m × 256 are set, wherein m is the arbitrary integer in 1~16, and frame walk length takes frame length
0.3~0.7 times of L, i.e., 70%~30% duplicate data per two adjacent frames all be present, the leakage of cut-point can be reduced with this
Inspection rate and false drop rate.
3.3) short-time energy is calculated：To each frame signal after framing, short-time energy is calculated：Parameter n
For length of window.
3.4) background sound eliminates：The energy of background sound is small compared with playing music, in a stable environment, it is believed that works as energy value
It is background sound during less than 0.1 times of audio signal ceiling capacity, this part background sound is deleted, speed is calculated to improve
Degree.
3.5) cut-point is inquired about：To the average energy value per adjacent 3 frameIt is compared, if there is larger gap, then recognizes
Catastrophe point (cut-point) to be present, that is, meet：Wherein, Th be judge before and after adjacent 3 frame threshold coefficient, one
As take between 3~7.
4) the instrumental audio Real-time segmentation that the position according to where cut-point is realized.
The present invention can use Java language to carry out specific procedure design, be broadly divided into two modules of recording and segmentation.In order to
Ensure to complete the real-time processing of audio signal on mobile communication equipment, two background threads are opened in design process, respectively
For recording and splitting.
1) the notice thread by the use of InitNotifications () function as initialization recording, when the buffering area of recording is expired
Data are stored in allData objects afterwards, identify for the pretreatment of subsequent audio data, segmentation and finally data storage.
2) delSilence () function is established, to the deletion to front end quiet data, that is, original audio front end is deleted and shakes
Width is less than 10-4Signal.
3) filter (short [] data, int z) function is established, to be filtered to the signal after Jing Yin deletion.Filter letter
Number is filter (), parameter value 0.97 or 0.98, and this filter function does convolution algorithm with the signal after Jing Yin deletion, as a result
It is stored in preProcess () function.
4) during shoegaze, after a note is pressed, the amplitude of audio is gradually decayed.To prevent high frequency
Sound interval is fallen, and audio signal is carried out into alternately sub-frame processing, processing function is frame (int m, double c), and wherein m is frame
Long interception coefficient, c are that frame moves coefficient, and between usual frame shifting coefficient is 0.3~0.7, framing result is stored in function
In frameSignal () function.
5) short-time energy is calculated to the result frameSignal () after framing, short-time energy is frameSignal ()
In every frame element quadratic sum, i.e.,As a result transmitted by function shortEnergy ().Idiographic flow such as Fig. 2
It is shown.
6) in order to accelerate the calculating speed of lookup cut-point, need to delete outlier before entering cut-point and searching.By
After calculating short-time energy, amplification can be produced to noise caused by some outliers, therefore, it is necessary to reject these outliers.
Compared with the music played, the energy of noise will be far smaller than playing music, and therefore, this method passes through many experiments, it is believed that when
Energy value is noise when being less than 0.1 times of audio signal ceiling capacity, deletes this partial noise.
7) after smooth and denoising, audio variation tendency is clearly.This method is by judging being averaged per adjacent 3 frame
Value is with the presence or absence of identification foundation of the large change as cut-point, it is believed that if the change of consecutive frame is clearly, then it is assumed that have
One new note occurs.Function findSegment (double [] shortEnergy, the int BgTh) realizations of this process,
Wherein BgTh is the threshold coefficient of background sound, generally takes 0.1.If the energy datum of a certain frame is less than the BgTh of ceiling capacity
Times, then it is assumed that cut-point is not present in the frame, without carrying out cut-point inquiry.The result for meeting cut-point inquiry is to split
Point, the position occurred is marked with pos [].
Idiographic flow is as shown in Figure 3.
8) the instrumental audio Real-time segmentation that the position according to where cut-point can quickly be realized, pos [] set is exactly point
Location sets where cutpoint.
It should be appreciated that for those of ordinary skills, can according to the above description be improved or converted,
And all these modifications and variations should all belong to the protection domain of appended claims of the present invention.
Claims (4)
1. a kind of instrumental audio real time method for segmenting realized on Intelligent mobile equipment, it is characterised in that comprise the following steps：
1) voice data is gathered：Obtain the voice data of musical instrument in real time by the sound pick-up outfit on mobile communication equipment；
2) the Jing Yin elimination in front end：Eliminate the quiet data of beginning in the voice data of collection；The quiet data is original
Front end amplitude is less than 10 in voice data-4Data；
3) voice data after elimination quiet data is handled in real time, specifically included：
3.1) preemphasis digital filtering：Preemphasis processing is carried out to the audio signal by the Jing Yin Processing for removing in front end and numeral is filtered
Ripple；
3.2) framing：Continuous framing is seamlessly transitted using alternately framing method；
3.3) short-time energy is calculated：To each frame signal after framing, short-time energy is calculated：Wherein parameter n
For length of window, x (m) is signal sequence；
3.4) background sound eliminates：If frame signal energy value is less than 0.1 times of audio signal ceiling capacity, the frame signal is background
Sound, background sound is deleted；
3.5) inquiry cut-point：Average energy value H per adjacent 3 frame is compared, if there is larger gap, then it is assumed that exist
Catastrophe point, catastrophe point are cut-point, that is, are met：H1＞ H2* Th, wherein, Th is the threshold coefficient of adjacent 3 frame before and after judging；
4) the instrumental audio Real-time segmentation that the position according to where cut-point is realized.
2. instrumental audio real time method for segmenting according to claim 1, it is characterised in that in the step 3.1), according to
6dB/oct ratio carries out preemphasis processing to sample；Digital filtering uses below equation：Y (n)=x (n)-a*x (n-1), its
Between middle parameter a values are 0.9~1, x (n) is the sampled signal before filtering, and y (n) is filtered signal.
3. instrumental audio real time method for segmenting according to claim 1, it is characterised in that in the step 3.2), framing
Strategy is as follows：Frame length L=m × 256 are set, wherein m is the arbitrary integer in 1~16, and frame walk length takes the 0.3~0.7 of frame length L
Times.
4. instrumental audio real time method for segmenting according to claim 1, it is characterised in that in the step 3.5), Th takes
Between value scope is 3~7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510549631.3A CN105227763B (en) | 2015-08-31 | 2015-08-31 | A kind of instrumental audio real time method for segmenting realized on Intelligent mobile equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510549631.3A CN105227763B (en) | 2015-08-31 | 2015-08-31 | A kind of instrumental audio real time method for segmenting realized on Intelligent mobile equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105227763A CN105227763A (en) | 2016-01-06 |
| CN105227763B true CN105227763B (en) | 2018-03-20 |
Family
ID=54996428
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510549631.3A Expired - Fee Related CN105227763B (en) | 2015-08-31 | 2015-08-31 | A kind of instrumental audio real time method for segmenting realized on Intelligent mobile equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105227763B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107241617A (en) * | 2016-03-29 | 2017-10-10 | 北京新媒传信科技有限公司 | The recognition methods of video file and device |
| CN106297824B (en) * | 2016-09-30 | 2017-08-01 | 西安交通大学 | A kind of audio frequency splitting method based on layering reliability variation tendency |
| CN110418245B (en) * | 2018-04-28 | 2021-03-19 | 深圳市冠旭电子股份有限公司 | Method and device for reducing reaction delay of Bluetooth sound box and terminal equipment |
| CN109934100B (en) * | 2019-01-24 | 2021-07-13 | 西安交通大学 | Amplitude-frequency time-varying process signal segmentation method based on sliding window |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101299344A (en) * | 2008-05-23 | 2008-11-05 | 嘉兴闻泰通讯科技有限公司 | Method for automatically sectioning MP3 sound recording document in mobile phone |
| CN101494049A (en) * | 2009-03-11 | 2009-07-29 | 北京邮电大学 | Method for extracting audio characteristic parameter of audio monitoring system |
| CN104347096A (en) * | 2013-08-09 | 2015-02-11 | 上海证大喜马拉雅网络科技有限公司 | Recording system and method integrating audio cutting, continuous recording and combination |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| SG150415A1 (en) * | 2007-09-05 | 2009-03-30 | Creative Tech Ltd | A method for incorporating a soundtrack into an edited video-with-audio recording and an audio tag |
| US8785760B2 (en) * | 2009-06-01 | 2014-07-22 | Music Mastermind, Inc. | System and method for applying a chain of effects to a musical composition |
-
2015
- 2015-08-31 CN CN201510549631.3A patent/CN105227763B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101299344A (en) * | 2008-05-23 | 2008-11-05 | 嘉兴闻泰通讯科技有限公司 | Method for automatically sectioning MP3 sound recording document in mobile phone |
| CN101494049A (en) * | 2009-03-11 | 2009-07-29 | 北京邮电大学 | Method for extracting audio characteristic parameter of audio monitoring system |
| CN104347096A (en) * | 2013-08-09 | 2015-02-11 | 上海证大喜马拉雅网络科技有限公司 | Recording system and method integrating audio cutting, continuous recording and combination |
Non-Patent Citations (2)
| Title |
|---|
| 《一种快速的音频分割与分类技术》;梁伟等;《第八届全国人机语音通讯学术会议论文集》;20051001;第80-84页 * |
| 《基于分层熵检测的音频分割算法》;王志明等;《科学技术与工程》;20090901;第9卷(第17期);第5012-5016页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105227763A (en) | 2016-01-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105227763B (en) | A kind of instrumental audio real time method for segmenting realized on Intelligent mobile equipment | |
| CN104681038B (en) | Audio signal quality detection method and device | |
| CN101366078A (en) | Neural network classifier for separating audio sources from a monophonic audio signal | |
| CN101221762A (en) | MP3 compression field audio partitioning method | |
| CN107622773B (en) | Audio feature extraction method and device and electronic equipment | |
| CN107274911A (en) | A kind of similarity analysis method based on sound characteristic | |
| JP2005227782A (en) | Apparatus and method for detecting voiced sound and unvoiced sound | |
| CN105975568A (en) | Audio processing method and apparatus | |
| CN109087634A (en) | A kind of sound quality setting method based on audio classification | |
| CN111312290B (en) | Audio data tone quality detection method and device | |
| CN101499085B (en) | Method and apparatus for fast extracting key frame | |
| CN110767248B (en) | Anti-modulation interference audio fingerprint extraction method | |
| CN103440870A (en) | Method and device for voice frequency noise reduction | |
| CN106550268A (en) | Method for processing video frequency and video process apparatus | |
| JP3607450B2 (en) | Audio information classification device | |
| CN110989020B (en) | Filtering method and system for noise interference of audio magnetotelluric data | |
| CN107247574A (en) | A kind of audio output control method and device | |
| CN104240697A (en) | Audio data feature extraction method and device | |
| CN113990297A (en) | ENF-based audio tampering identification method | |
| CN111639225B (en) | Audio information detection method, device and storage medium | |
| Cai et al. | A modified pitch detection method based on wavelet transform | |
| Maulana | The utilization of Gaussian filter method on voice record frequency noise | |
| CN113593603A (en) | Audio category determination method and device, storage medium and electronic device | |
| CN113450816A (en) | Voice active noise reduction method based on deep neural network | |
| CN106548784A (en) | A kind of evaluation methodology of speech data and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180320 Termination date: 20200831 |