CN1138253C - Methods for identifying the characteristics of sound sources - Google Patents
Methods for identifying the characteristics of sound sources Download PDFInfo
- Publication number
- CN1138253C CN1138253C CNB001168630A CN00116863A CN1138253C CN 1138253 C CN1138253 C CN 1138253C CN B001168630 A CNB001168630 A CN B001168630A CN 00116863 A CN00116863 A CN 00116863A CN 1138253 C CN1138253 C CN 1138253C
- Authority
- CN
- China
- Prior art keywords
- matrix
- sound
- row
- singular value
- value
- 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
- 238000000034 method Methods 0.000 title claims abstract description 16
- 230000005855 radiation Effects 0.000 claims abstract description 19
- 238000001228 spectrum Methods 0.000 claims abstract description 17
- 239000011159 matrix material Substances 0.000 claims description 61
- 238000004458 analytical method Methods 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000000354 decomposition reaction Methods 0.000 abstract description 2
- 238000003491 array Methods 0.000 abstract 2
- 230000000694 effects Effects 0.000 description 4
- 230000003595 spectral effect Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000005259 measurement Methods 0.000 description 3
- 230000007246 mechanism Effects 0.000 description 3
- 230000015572 biosynthetic process Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 241001342895 Chorus Species 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- HAORKNGNJCEJBX-UHFFFAOYSA-N cyprodinil Chemical compound N=1C(C)=CC(C2CC2)=NC=1NC1=CC=CC=C1 HAORKNGNJCEJBX-UHFFFAOYSA-N 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012886 linear function Methods 0.000 description 1
- 230000035800 maturation Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000011664 signaling Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
Images
Landscapes
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The present invention relates to a method for identifying sound source characteristics. The method comprises the steps that on an occasion when a plurality of sound sources simultaneously generate sound radiation, the number of main sound sources is determined, and sound conducting apparatuses are arranged; sound pressure signals are converted into cross spectrum arrays of an analyzed frequency range; the cross spectrum arrays are processed in the mode of singular value decomposition to obtain a singular value array and a right characteristic vector array; the positions of singular values of the singular value array are adjusted according to position maximum values; diagonal elements at the same positions of the singular value array are connected into a curve, and thus, each noise characteristic is obtained.
Description
Technical field
What the present invention relates to is a kind of method of identifying sound source characteristic, belongs to noise class field in the Speciality of Physics.
Technical background
The Noise Control problem more and more is subject to people's attention, in the Research And Engineering of vehicle noise, Control of Mechanical Noise is used, seek the main parts size that produces noise, analyze sound source characteristic and travel path, can correct judgement mechanism of noise generation be the prerequisite that find targeted innovative approach.Exist simultaneously for a plurality of noise sources, and the situation of radiated noise, at first to judge the number of main sound source and their position, this existing several different methods is suggested, wherein, with singular value decomposition method maturation, explicit physical meaning, effect is better especially.Find out the quantity of noise source, and finding after the position or sounding parts of noise source, also needing to obtain the characteristic of the noise of each noise source institute radiation.This therefrom obtains the sound of everyone solo as 3 people's chorus sound is handled.But, up to the present, how the characteristic of identifying sound source does not still have effective ways, majority method only simply points out to adopt manual method to connect the frequency characteristic that spectral line comes identifying sound source, this often is only applicable to the simplest sound source, complicated a little any situation is just powerless, even simple situation also can be owing to judging not to connecting wrong spectral line.By prior art documents and analyze, Jap.P.: the spy opens flat 11-83613, name is called: source of sound characteristic means of identification and device thereof, found an identifying sound source frequency characteristic effective way, for free found field or near the test condition of free found field, this method has effect preferably, goes for some mechanical noise sound source characteristic identifications.But aspect practical application, this method is for the more serious test condition of echo reverberation ratio, and as compartment or other closed chambers etc., the result of sound source identification is just not ideal enough.
Summary of the invention
The objective of the invention is to overcome deficiency of the prior art, a kind of method of identifying sound source characteristic is provided.The acoustical signal that the present invention can directly record from the scene obtains the frequency characteristic of sound source, thereby can understand each noise source mechanism of production, finds the most effective noise-reduction method.
Technical scheme of the present invention is as follows: the method for identifying sound source characteristic is divided into following three steps:
(1) produces the sound radiation occasion simultaneously in a plurality of sound sources, at first determine the quantity of main sound source, near these noise sources, arrange the microphone identical then with noise source quantity, these microphones are measured sound pressure signal simultaneously, and with these sound pressure signals by business software or the Calculation Method of programming voluntarily, obtain the cross-spectrum matrix of institute's analysis frequency scope, the cross-spectrum matrix to each frequency carries out svd then, obtains singular value matrix and right eigenvectors matrix;
(2) the peaked position of each column element mould in the right eigenvectors matrix of searching, respectively be listed as earlier peaked size, then, seek each row maximal value position in proper order by it, promptly find out the maximum value position of those row of row maximal value maximum earlier, other element that this maximal value is expert at is changed to 0 again, and the like, seek all the other each row maximum value positions, until the peaked position of definite every row of right eigenvectors matrix, adjust the position of singular value in the singular value matrix according to peaked position, these positions;
(3) value with the same position diagonal element of adjusted singular value matrix in each frequency connects into curve, can obtain the frequency characteristic of each noise source radiation sound.
According to the value of right eigenvectors matrix V, construct a replacement matrix P, rearrange the diagonal element of singular value matrix Λ, the diagonal element of the singular value matrix after the replacement has reflected the frequency characteristic of sound source, concrete grammar is as follows:
(1) to input signal cross-spectrum matrix S
nMake svd, obtain singular value matrix Λ and right eigenvectors matrix V,
(2) whole elements of n being tieed up square formation P are initialized as 0,
(3) maximal value of each column element mould in the searching V matrix, determine their positions in V, respectively be listed as earlier peaked size, seek each row maximum value position in proper order by it then, promptly find out the maximum value position of those row of row maximal value maximum earlier, other element that this maximal value is expert at is changed to 0 again, and the like, seek all the other each row maximum value positions
(4) element corresponding to the every row maximum value position of V among the P is changed to 1, so can obtains the matrix P that resets,
(5) replacement singular value matrix Λ, i.e. Λ
z=P* Λ * P
T,
Singular value matrix Λ after the replacement
zStill be diagonal matrix, but Λ in whole spectral line
zThe diagonal element at same position place represent the tolerance of same noncoherent signal energy, with each frequency Λ
zIn the value of diagonal element at same position place connect into curve, can obtain the frequency characteristic of each incoherent sound source radiation sound.
The present invention has substantive distinguishing features and marked improvement, when often having a plurality of sound sources, vehicle, naval vessel and other complicated machineries exist simultaneously, when equipment operation, these sound sources are radiated noise simultaneously, therefore generally can not directly measure the radiation sound of each sound source, the present invention can be under general in-site measurement condition, with the frequency characteristic of more common instrumental analysis main sound source, thereby grasp the mechanism of production of each sound source, provide foundation for taking the most effective noise reduction measure.The present invention analyzes the frequency characteristic of each sound source, can also come the running status of checkout equipment by acoustical signal, tracing trouble.
Description of drawings
Below in conjunction with accompanying drawing the present invention is further described:
Fig. 1 sound source analysis and signal measurement relation synoptic diagram
Fig. 2 sound source characteristic is debated and is known the simulation synoptic diagram
Three sound source radiation acoustic frequencies of Fig. 3 signal schematic representation
Three sound source radiation sound of Fig. 4 singular value frequency curve chart
Three low voice speaking postpone singular values of sound source radiation of Fig. 5 curve map
Embodiment
The present invention mainly is divided into following three steps:
1, produces the sound radiation occasion simultaneously in a plurality of sound sources, at first determine the quantity of main sound source, near these noise sources, arrange the microphone identical then with noise source quantity, these microphones are measured sound pressure signal simultaneously, and with these sound pressure signals by business software or the Calculation Method of programming voluntarily, obtain the cross-spectrum matrix of institute's analysis frequency scope, the cross-spectrum matrix to each frequency carries out svd then, obtains singular value matrix and right eigenvectors matrix;
2, seek the peaked position of each column element mould in the right eigenvectors matrix, respectively be listed as earlier peaked size, then, seek each row maximal value position in proper order by it, promptly find out the maximum value position of those row of row maximal value maximum earlier, other element that this maximal value is expert at is changed to 0 again, and the like, seek all the other each row maximum value positions, until the peaked position of definite every row of right eigenvectors matrix, adjust the position of singular value in the singular value matrix according to peaked position, these positions;
3, the value with the same position diagonal element of adjusted singular value matrix in each frequency connects into curve, can obtain the frequency characteristic of each noise source radiation sound.
As shown in Figure 1, the relation of sound source and measuring-signal in the noise analysis, wherein, sound-source signal vector U is by the radiation sound su of n separate physical sound sources
i(i=1,2 ..., n) to form, they all are immeasurablel usually; N the input signal x that sensor records
i(i=1,2 ..., n) constituted input vector.The signal x that common each microphone records
iIn comprised a plurality of or whole sound source u
iRadiation sound, the transport function between sound-source signal and the input is G, the normally full battle array of G, promptly input is the linear function of sound-source signal:
X＝G
HU (1)
G
HThe conjugate transpose of representing matrix G.The cross-spectrum S of sound-source signal then
UUCross-spectrum S with input
XXBetween the pass be:
S
XX＝G
HS
UUG (2)
To input cross-spectrum S
XXMake svd:
S
XX＝UΛV
H (3)
Owing to all be separate with n the sound source that finds someway, so the cross-spectrum matrix S of input
XXIt is invertible matrix.V in the formula (3) is S
XXRight eigenvectors matrix, and U is S
XXThe left eigenvector matrix, U, V
HBe orthogonal matrix; Singular value matrix Λ is a diagonal matrix, its diagonal element λ
iSize be the tolerance of corresponding sound-source signal energy, but can not reflect the frequency characteristic of sound-source signal.
The present invention constructs a replacement matrix P according to the value of right eigenvectors matrix V, rearranges the diagonal element of singular value matrix Λ, and the diagonal element of the singular value matrix after the replacement has reflected the frequency characteristic of sound source, and concrete grammar is as follows:
(1) to input signal cross-spectrum matrix S
XXMake svd, obtain singular value matrix Λ and right eigenvectors matrix V,
(2) whole elements of n being tieed up square formation P are initialized as 0,
(3) maximal value of each column element mould in the searching V matrix, determine their positions in V, respectively be listed as earlier peaked size, seek each row maximum value position in proper order by it then, promptly find out the maximum value position of those row of row maximal value maximum earlier, other element that this maximal value is expert at is changed to 0 again, and the like, seek all the other each row maximum value positions
(4) element corresponding to the every row maximum value position of V among the P is changed to 1, so can obtains the matrix P that resets,
(5) replacement singular value matrix Λ, i.e. Λ
z=P* Λ * P
T,
Singular value matrix Λ after the replacement
zStill be diagonal matrix, but Λ in whole spectral line
zThe diagonal element at same position place represent the tolerance of same noncoherent signal energy, with each frequency Λ
zIn the value of diagonal element at same position place connect into curve, can obtain the frequency characteristic of each incoherent sound source radiation sound.
Below further introducing embodiment, as shown in Figure 2, if there are three sound sources to produce sound radiation simultaneously, is respectively su
1, su
2And su
3, and these radiation sound can not record separately, so the frequency characteristic of radiation sound is (as the auto-power spectrum u among Fig. 3
1, u
2And u
3) also can't learn.Near these sound sources, arrange three microphone C
1, C
2, C
3, microphone is the closer to sound source, and the effect of identification is just good more usually.The sound pressure signal that microphone records is delivered to dedicated analysis device or PC computing machine behind amplifier F, computing machine or analyzer can calculate the cross-spectrum matrix of institute's analysis frequency scope (as: 0-5000Hz) according to the sound pressure signal that microphone records, the cross-spectrum matrix is carried out svd, and the frequency curve of three singular values is seen the λ among Fig. 4
1, λ
2And λ
3, obviously singular value is arranged by size, can not embody the characteristic of each sound source; The order and the position of greatest member in each column vector in the right eigenvectors matrix that obtains during again according to svd, the order of replacement singular value is linked to be curve λ with the singular value after resetting
Z1, λ
Z2And λ
Z3(see figure 5) can be grasped the characteristic of each sound source radiation sound.Comparison diagram 3 and Fig. 5 can see that the singular value spectrogram after the replacement has reflected the frequency characteristic of sound source well.Even it is serious or to measure noise higher to contain weak signal or coupling in measurement, identification effect of the present invention is all fairly good.
Claims (2)
1, a kind of method of identifying sound source characteristic, it is characterized in that producing the sound radiation occasion simultaneously in a plurality of sound sources, at first determine the quantity of main sound source, near these noise sources, arrange the microphone identical then with noise source quantity, these microphones are measured sound pressure signal simultaneously, and with these sound pressure signals by business software or the Calculation Method of programming voluntarily, obtain the cross-spectrum matrix of institute's analysis frequency scope, cross-spectrum matrix to each frequency carries out svd then, obtain singular value matrix and right eigenvectors matrix, promptly find out the maximum value position of those row of row maximal value maximum earlier, until the peaked position of definite every row of right eigenvectors matrix, adjust the position of singular value in the singular value matrix according to peaked position, these positions, the value of the same position diagonal element of adjusted singular value matrix in each frequency is connected into curve, can obtain the frequency characteristic of each noise source radiation sound.
2, the method of this identifying sound source characteristic according to claim 1, its feature also is to seek the peaked position of each column element mould in the right eigenvectors matrix, respectively be listed as earlier peaked size, then, seek each row maximal value position in proper order by it, promptly find out the maximum value position of those row of row maximal value maximum earlier, other element that this maximal value is expert at is changed to 0 again, and the like, seek all the other each row maximum value positions, until the peaked position of definite every row of right eigenvectors matrix, adjust the position of singular value in the singular value matrix according to peaked position, these positions.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB001168630A CN1138253C (en) | 2000-06-29 | 2000-06-29 | Methods for identifying the characteristics of sound sources |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB001168630A CN1138253C (en) | 2000-06-29 | 2000-06-29 | Methods for identifying the characteristics of sound sources |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1290923A CN1290923A (en) | 2001-04-11 |
| CN1138253C true CN1138253C (en) | 2004-02-11 |
Family
ID=4586262
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB001168630A Expired - Fee Related CN1138253C (en) | 2000-06-29 | 2000-06-29 | Methods for identifying the characteristics of sound sources |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1138253C (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7424119B2 (en) * | 2003-08-29 | 2008-09-09 | Audio-Technica, U.S., Inc. | Voice matching system for audio transducers |
| CN1302461C (en) * | 2004-04-13 | 2007-02-28 | 中国科学院声学研究所 | Noise inhibiting method in vertical array receiving signal cross spectrum arra yevaluation in sea |
| JP4872871B2 (en) * | 2007-09-27 | 2012-02-08 | ソニー株式会社 | Sound source direction detecting device, sound source direction detecting method, and sound source direction detecting camera |
| CN101936818B (en) * | 2010-08-27 | 2012-09-05 | 上海交通大学 | Diagnostic system of non-contact type rotary mechanical failure |
| CN105698918B (en) * | 2014-11-24 | 2019-01-22 | 广州汽车集团股份有限公司 | A method and device for visually comparing vibration and noise colormaps |
| CN109874084A (en) * | 2019-02-20 | 2019-06-11 | 北京安声浩朗科技有限公司 | Acoustic constituents extracting method and system |
-
2000
- 2000-06-29 CN CNB001168630A patent/CN1138253C/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| CN1290923A (en) | 2001-04-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR100493172B1 (en) | Microphone array structure, method and apparatus for beamforming with constant directivity and method and apparatus for estimating direction of arrival, employing the same | |
| US6542857B1 (en) | System and method for characterizing synthesizing and/or canceling out acoustic signals from inanimate sound sources | |
| Michelsen et al. | Plants as transmission channels for insect vibrational songs | |
| CN1664610B (en) | Method for beamforming using a microphone array | |
| US6725108B1 (en) | System and method for interpretation and visualization of acoustic spectra, particularly to discover the pitch and timbre of musical sounds | |
| CN109448389B (en) | Intelligent detection method for automobile whistling | |
| CN1138253C (en) | Methods for identifying the characteristics of sound sources | |
| CN112487703B (en) | Underdetermined broadband signal DOA estimation method based on sparse Bayes in unknown noise field | |
| CN109243429A (en) | A kind of pronunciation modeling method and device | |
| CN110515034A (en) | An acoustic signal azimuth measurement system and method | |
| CN110010110B (en) | Method for simulating timbre of loudspeaker box | |
| CN116567515B (en) | Microphone array calibration method | |
| CN112540346A (en) | Sound source positioning method based on signal-to-noise ratio weight optimization updating | |
| CN1202408C (en) | Process for identifying non-stable acoustical source characteristic applying principal component analyzing technique | |
| CN118298842A (en) | Audio separation method and device based on memory and calculation integrated chip and electronic equipment | |
| CN110632605A (en) | Wide-tolerance large-aperture towed linear array time domain single-beam processing method | |
| Wang et al. | A novel transmissibility matrix method for identifying partially coherent noise sources on vehicles | |
| CN1875292A (en) | Resolving directional information in sonar arrays | |
| CN2909198Y (en) | Sound field measuring instrument using transmission method | |
| CN206269919U (en) | Violin entirety acoustic vibration and the measuring system of firmware vibration | |
| CN114167355B (en) | An underwater DOA estimation method in autocorrelation domain based on sparse nested linear arrays | |
| Wun et al. | Evaluation of weighted principal-component analysis matching for wavetable synthesis | |
| KR102707335B1 (en) | Method and apparatus for estimating blind reverberation time using attentive pooling-based weighted sum of spectral decay rates | |
| CN118194145A (en) | Acoustic signal classification method based on multi-resolution analysis and feature separation | |
| Dubnov et al. | Clustering Musical Sounds Using Polyspectral Distance Measures |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C06 | Publication | ||
| PB01 | Publication | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C19 | Lapse of patent right due to non-payment of the annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |