CN105371945B - Remove the method and device of noise in hall impulse response measurement acoustical signal - Google Patents
Remove the method and device of noise in hall impulse response measurement acoustical signal Download PDFInfo
- Publication number
- CN105371945B CN105371945B CN201510650988.0A CN201510650988A CN105371945B CN 105371945 B CN105371945 B CN 105371945B CN 201510650988 A CN201510650988 A CN 201510650988A CN 105371945 B CN105371945 B CN 105371945B
- Authority
- CN
- China
- Prior art keywords
- signal
- time
- frequency
- noise
- impulse response
- 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.)
- Active
Links
- 230000004044 response Effects 0.000 title claims abstract description 129
- 238000005259 measurement Methods 0.000 title claims abstract description 52
- 238000000034 method Methods 0.000 title claims abstract description 38
- 238000004458 analytical method Methods 0.000 claims abstract description 39
- 238000004364 calculation method Methods 0.000 claims abstract description 7
- 238000006243 chemical reaction Methods 0.000 claims description 44
- 230000009466 transformation Effects 0.000 claims description 27
- 230000006870 function Effects 0.000 claims description 21
- 238000009826 distribution Methods 0.000 claims description 13
- 230000008569 process Effects 0.000 claims description 10
- 239000002131 composite material Substances 0.000 claims description 6
- 230000007274 generation of a signal involved in cell-cell signaling Effects 0.000 claims description 5
- 238000009499 grossing Methods 0.000 claims description 4
- 230000008054 signal transmission Effects 0.000 claims description 4
- 210000001367 artery Anatomy 0.000 claims description 2
- 230000009514 concussion Effects 0.000 claims description 2
- 230000007423 decrease Effects 0.000 claims description 2
- 238000001914 filtration Methods 0.000 claims description 2
- 238000004080 punching Methods 0.000 claims description 2
- 238000005316 response function Methods 0.000 claims description 2
- 238000001228 spectrum Methods 0.000 claims description 2
- 210000003462 vein Anatomy 0.000 claims description 2
- 230000014509 gene expression Effects 0.000 claims 1
- 238000012360 testing method Methods 0.000 description 8
- 238000005516 engineering process Methods 0.000 description 5
- 238000010276 construction Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 230000015572 biosynthetic process Effects 0.000 description 2
- 230000003247 decreasing effect Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000003786 synthesis reaction Methods 0.000 description 2
- 238000010521 absorption reaction Methods 0.000 description 1
- 230000033228 biological regulation Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000009792 diffusion process Methods 0.000 description 1
- 210000005069 ears Anatomy 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000010892 electric spark Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 239000004615 ingredient Substances 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Landscapes
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The present invention relates to a kind of method and devices of noise in removal hall impulse response measurement acoustical signal, comprising: 1) signal emission module issues weighting frequency sweep or multifrequency sine acoustical signal into the hall；2) sound pressure signal of the signal acquisition module acquisition hall；3) signal time-frequency denoising module removes the different time Frequency domain noise being mixed into signal and with time-frequency domain nonstationary noise by time frequency analysis, threshold denoising and useful signal reconstruct, and the time frequency signal after denoising is returned into time domain, it is sent to impulse response computing module；4) sound pressure signal after denoising is done relevant calculation with frequency sweep or multifrequency sine digital signal by impulse response computing module, obtains the impulse response of the hall；5) the noise acoustical power denoising module that is averaged calculates steady-state noise acoustical power in impulse response first, is then deducted in Energy impulse response；6) display and memory module show calculated result.Compared with prior art, invention significantly improves the precision of live sound quality measurement and noiseproof features.
Description
Technical field
The present invention relates to the measuring techniques of field of acoustics, are rung more particularly, to one kind based on time frequency analysis removal hall pulse
The method and device of noise in acoustical signal should be measured.
Background technique
Impulse response is to reflect the important physical amount of auditorium, and the reverberation time can be directly calculated from impulse response
(RT60), early stage decay time (EDT), the clarity factor (D50), the perspectivity factor (C50), the loudness factor (G), ears are related
The auditoriums parameters such as coefficient (IACC).Hall impulse response acoustical signal is accurately measured to check and accept auditorium acoustical design, sound quality
Development and engineering practice with environmental noise abatement technology are of great significance.Live impulse response measurement is often subject to language
Sound, traffic and construction equipment noise jamming especially have the scene of spectators or the in-site measurement of construction period, and noise is to test
It interferes almost inevitable.Measurement for acoustical parameters such as reverberation time, the test of specification regulation will reach 35dB in signal-to-noise ratio
In the case where measurement it is just effective, however other than laboratory and the late into the night quiet environment, in-site measurement is extremely difficult to so high
Signal-to-noise ratio.Usually it can lead to test invalidation in the case where signal-to-noise ratio is less than 25dB, therefore anti-noise technology is for impulse response
In-site measurement is highly important.
Presently relevant technology is widely used in impulse response measurement technology.Method, needle are interrupted relative to steady-state noise
The signal-to-noise ratio of measurement can be effectively improved to the steady-state noise method, but this technology is for nonstationary noise, impulsive noise etc.
Denoising effect and bad, and ambient noise such as voice, traffic and construction noise are largely nonstationary noise, impulsive noise.
Therefore development for nonstationary noise, impulsive noise denoising method for improving the precision of live impulse response measurement and reliable
Property is very necessary.
Summary of the invention
It is an object of the present invention to overcome the above-mentioned drawbacks of the prior art and provide one kind accurately and fast, certainly
The method and device of noise in the high removal hall impulse response measurement acoustical signal of dynamicization degree, within the scope of 100~8000Hz,
Signal-to-noise ratio needed for measurement impulse response 10dB be can be decreased to, the precision and reliability of in-site measurement substantially increased.
The purpose of the present invention can be achieved through the following technical solutions:
A method of noise in removal hall impulse response measurement acoustical signal, which comprises the following steps:
1) signal emission module is generated by sequentially connected digital signal and weighted units, digital-to-analogue conversion card, power are put
Big device and loudspeaker issue weighting frequency sweep acoustical signal or multifrequency sine acoustical signal into the hall, shown in acoustical signal in the hall
Noise has marked difference in time-frequency domain distribution characteristics；
2) signal acquisition module is connect by sequentially connected microphone, preamplifier, analog-to-digital conversion card and digital signal
Unit is received, the sound pressure signal of the hall is acquired, and sends it to signal time-frequency denoising module；
3) signal time-frequency denoising module is mixed into signal by time frequency analysis, threshold denoising and the removal of useful signal reconstructing method
Interior different time Frequency domain noise, and with the nonstationary noise component in time-frequency domain, and when the time frequency signal after denoising is returned to
Domain is sent to impulse response computing module；
4) impulse response computing module is sound pressure signal and frequency sweep after denoising or the digital signal of multifrequency sine acoustical signal
Do relevant calculation, obtain the impulse response of the hall, and send responses to noise be averaged acoustical power denoising module；
5) the noise acoustical power denoising module that is averaged calculates steady-state noise acoustical power in impulse response first, then in energy
It is deducted in impulse response；
6) as the result is shown and memory module to be shown in the sound pressure signal curve measured in the hall, acoustic pressure pulse response bent
Line, Energy impulse response curve, and store its digital signal.
The digital signal generates and weighted units include digital signal generation unit and weighted units, the number
The calculating process that signal generation unit generates signal is as follows, by taking Linear chirp as an example:
Wherein, A is signal amplitude；T is signal total duration；T is time variable；f1It is low cutoff frequency；f2It is high cutoff frequency
Rate；
The calculating process that the signal weighting unit is weighted signal is as follows:
Wherein, T is signal total duration；T is time variable；e0It (t) is Linear chirp；ω is angular frequency.
The digital signal generate and weighted units to limited Linear chirp in time domain and frequency domain in time domain
Starting and ending section has carried out Sine Modulated, and the amplitude of signal is made smoothly to decay to zero in starting point and end, modulation
Principle is to guarantee that modulated signal is not decayed substantially in the frequency range self-energy of measurement request, avoids the frequency of filtering
Spectrum leakage, reduces the concussion of frequency domain amplitude, keeps filtered signal energy computation more acurrate, the pulse for obtaining autocorrelation calculation
It is more sharp.
The signal time-frequency denoising module includes time-frequency conversion submodule, threshold denoising submodule, useful signal reconstruct
Submodule and time-frequency inverse transformation submodule, in which:
11) time-frequency conversion submodule, for the acoustic pressure time-domain signal of acquisition to be transformed to time-frequency domain signal, it is therefore an objective to make to mix
Enter the different time Frequency domain noise in signal on time-frequency domain with Signal separator, transform method is obtained using VARIANT FILTER OF DISCRETE GABOR TRANSFORM
Time-frequency distributions:
Wherein, W=exp (j2 π/N)；It is the amplitude of time frequency analysis coefficient；It is the acoustic pressure time domain letter of acquisition
Number；Δ M is time-domain analysis step-length；Δ N is frequency-domain analysis step-length；It is the analytic function of Gabor transformation；
12) threshold denoising submodule is made an uproar interior noise for effectively removing different time frequency domain, specifically: according to signal with make an uproar
The time-frequency domain distribution characteristics otherness of sound handles the time-frequency conversion coefficient being calculated, and determines the denoising threshold in time-frequency domain
Value, the signal higher than threshold value retain, and lower than the signal zero setting of threshold value, constitute new time-frequency conversion coefficient sets, form useful signal；
13) useful signal reconstructs submodule, and the time-frequency of several useful signals for will obtain in different periods measurement becomes
The amplitude for changing coefficient is compared, and same time frequency point selects the smallest coefficient of energy magnitude, the time-frequency conversion coefficient reconstructed
Group, wherein the smallest coefficient of energy magnitude is the smallest signal of noise pollution degree, which can effectively remove in same time-frequency domain
The unstable state component of noise；
14) time-frequency inverse transformation submodule will for carrying out time-frequency inverse transformation to the time-frequency conversion coefficient sets obtained after reconstruct
Time-frequency impulse response signal transforms in time domain, is obtained using Discrete Gabor Expansion:
Wherein, W=exp (j2 π/N)；It is the acoustic pressure time-domain signal after inverse transformation；It is the time frequency analysis of reconstruct
The amplitude of coefficient；It is the composite function of Gabor transformation.
The impulse response computing module is to utilize the digital linear swept-frequency signal e modulated in signal emission module0(t)
With sound pressure signal y (t) after the denoising of received time frequency analysis make relevant calculation, obtain noisy acoustic pressure of the measuring point about room
Impulse response p (t), wherein y (t) is time-domain signalContaining part simultaneously with frequency noise n (t):
y(t)*e0 -1(t)=e0(t)*p(t)*e0 -1(t)+n(t)*e0 -1(t)=p (t)+n (t) * e0 -1(t)
P (t) still with noise n (t) * e0 -1(t) it mixes, influences measurement accuracy, wherein n (t) * e0 -1(t) basic
Show as the steady-state noise in statistical significance.
The noise be averaged acoustical power denoising module working principle it is as follows:
It calculates noise using impulse response latter stage noise signal in the highest flight to be averaged acoustical power, then by impulse response
Energy impulse response is converted to, and deducts noise within Energy impulse response all the period of time and is averaged acoustical power, detailed process is as follows:
Energy impulse response is integrated using time sliding window exponential smoothing, obtains the Energy impulse response sound of smoothing
Can attenuation curve, and deduct noise in window in entirely integral time domain and be averaged acoustical power, the steady-state component in removal noise, it is assumed that
Impulse response p (t) in the hall is made of signal and steady-state noise:
P (t)=p (0) e-σt+N(t)
P (0) is the initial value of impulse response function；N (t) is steady-state noise, i.e. noise n (t) * e0 -1(t)；
Latter stage signal prevailing for noise can ask its average acoustical power to obtain:
Wherein t1At the time of capable of being flooded by ambient noise for signal sound, T t1Noise total duration after moment, under normal conditions T
Length enable toRelatively stable estimated value is obtained, length is T as a result,0Sound in sliding window can indicate are as follows:
After the average acoustical power for removing noise, then Energy impulse response sound energy attenuation curve is obtained are as follows:
A kind of device removing noise in hall impulse response measurement acoustical signal, which is characterized in that emit mould including signal
Block, signal acquisition module, signal time-frequency denoising module, impulse response computing module, the noise acoustical power that be averaged denoise module and knot
Fruit shows and memory module；The signal emission module is responsible for issuing weighting frequency sweep acoustical signal into the hall；The signal
Acquisition module, signal time-frequency denoising module, impulse response computing module, noise be averaged acoustical power denoise module, as the result is shown with
Memory module is sequentially connected, and the impulse response computing module is also connected with signal emission module.
The signal emission module includes sequentially connected digital signal generation and weighted units, digital-to-analogue conversion card, function
Rate amplifier and loudspeaker, for completing the task of transmitting weighting frequency sweep or multifrequency sine acoustical signal into the hall, the number
The starting and ending section of the generation of word signal and weighted units to limited Linear chirp in time domain and frequency domain in time domain is distinguished
Sine Modulated has been carried out, the amplitude of signal is made smoothly to decay to zero in starting point and end；
The signal acquisition module includes sequentially connected microphone, preamplifier, analog-to-digital conversion card and number letter
Acquisition signal to complete the task that sound pressure signal acquires in the hall, and is sent to signal time-frequency denoising module by number receiving unit.
The signal time-frequency denoising module includes time-frequency conversion submodule, threshold denoising submodule, useful signal reconstruct
Submodule and time-frequency inverse transformation submodule, in which:
Time-frequency conversion submodule, for the acoustic pressure time-domain signal of acquisition to be transformed to time-frequency domain signal；
Threshold denoising submodule, for according to the time-frequency domain distribution characteristics otherness of signal and noise to be calculated when
Frequency transformation coefficient is handled, and determines the noise-removed threshold value in time-frequency domain, and the signal higher than threshold value retains, and the signal lower than threshold value is set
Zero, new time-frequency conversion coefficient sets are constituted, useful signal is formed；
Useful signal reconstructs submodule, the time-frequency conversion system of several useful signals for will obtain in different periods measurement
Several amplitudes are compared, the same time frequency point selection the smallest coefficient of energy magnitude, the time-frequency conversion coefficient sets reconstructed,
The middle the smallest coefficient of energy magnitude is the smallest signal of noise pollution degree；
Time-frequency inverse transformation submodule, for carrying out time-frequency inverse transformation to the time-frequency conversion coefficient sets obtained after reconstruct.
The noise acoustical power denoising module that is averaged is calculated using impulse response latter stage noise signal in the highest flight
Noise is averaged acoustical power, impulse response is then converted to Energy impulse response, and deduct within Energy impulse response all the period of time
Average noise acoustical power.
Selection signal transmission channel and acquisition channel；Input transmitting signal duration, acquisition signal duration, sample frequency, letter
Number amplitude, the initial frequency of swept-frequency signal and cutoff frequency.
When time frequency analysis select window function type, window function width, time analysis step-length, frequency analysis step-length, by pair
Analysis and composite function is calculated in even window function calculating instrument case.
It is described as the result is shown and memory module includes sound pressure curve display unit before showing denoising collected, denoising
Sound pressure curve display unit after preceding impulse response display unit, denoising, the impulse response display unit after denoising；And it goes
The impulse response storage unit before sound pressure signal storage unit, denoising before making an uproar, the sound pressure signal storage unit after denoising, denoising
Impulse response storage unit afterwards；
Wherein, acquisition sound pressure curve display unit display acquisition sound pressure curve figure, the impulse response are shown
Unit shows acoustic pressure and Energy impulse response curve graph；The sound pressure signal storage unit is stored in collected sound in the hall
Press signal, the impulse response storage unit storage acoustic pressure pulse response.
Compared with prior art, the invention has the following advantages that
1, sound field is motivated using the frequency sweep of weighting or multifrequency sine acoustical signal, by receiving signal and Digital Sweep or multifrequency
Sinusoidal digital signal relevant calculation obtains impulse response；Frequency sweep or multifrequency sine acoustical signal and most of noise signal are in time-frequency domain
Upper characteristic distributions are different, by time frequency analysis and threshold denoising, are easy to distinguish and eliminate major part with the distribution of signal time-frequency domain not
Same noise signal, obtains useful signal.
2, the useful signal of several periods is compared, the minimum time-frequency coefficients of identical time frequency point selection amplitude, i.e.,
By the acoustical signal that noise pollution degree is minimum, the time-frequency coefficients group reconstructed, i.e., noisy least time frequency signal (institute's Noise
Major part is steady-state noise)；Because the acoustical signal of different time frequency points is constant, and noise signal is to become, and selects amplitude most
Low time-frequency coefficients can remove and signal is with the unstable state component of the noise in time-frequency domain, and this method removes nonstationary noise
It makes an uproar significant effect.
3, time-domain signal is obtained using reconstruction signal, the impulse response containing steady-state noise is obtained by relevant calculation, then
The average acoustical power of noise is calculated using the prevailing signal segment of latter stage impulse response noise, and in all the period of time energy pulse
The average acoustical power that noise is deducted in response, removes the steady-state noise components in Energy impulse response.
By the above-mentioned synthesis denoising to noisy acoustical signal, voice, traffic, equipment and construction etc. can be effectively removed
All kinds of stable states, unstable state and impulsive noise can will measure necessary signal-to-noise ratio by 35dB and be down to 10dB, significantly improve impulse response
The precision and reliability of in-site measurement, reduce the difficulty of in-site measurement, and survey crew is not necessarily to obtain 35dB signal-to-noise ratio and adjust
With powerful acoustical signal emitter (power of acoustical signal emitter can be reduced by 100 times), also the late into the night need not be waited until again
Carry out measurement.
4, the advantages of carrying out sine weighting to Linear chirp is that (1) obtained impulse response is more sharp, is more connect
Nearly ideal impulse response, is remarkably improved the precision of hall acoustic pressure pulse response measurement；(2) electroacoustics system (sound source) is improved to ring
The ride comfort of process is answered, if signal is substantially increased or terminates suddenly, electroacoustics system can be made to generate significant non-linear distortion, sternly
Ghost image rings measurement accuracy.(3) anti-distortion ability is strong, before non-linear harmonic wave ingredient being arranged in ping, is easy to reject, from
And improve the signal-to-noise ratio of impulse response.
Similar impulse response measurement system largely uses unmodulated swept-frequency signal both at home and abroad in the market, frequently results in
Electroacoustics system generates significant non-linear distortion, seriously affects measurement accuracy.
5, the characteristics of being directed to different noises, uses the synthesis Denoising Algorithm based on time frequency analysis, in 100~8000Hz range
It is interior, signal-to-noise ratio needed for measurement impulse response can be decreased to 10dB, substantially increase the validity and reliability of in-site measurement,
The requirement to acoustical signal emission system power is reduced, realizes the portability of measuring system.
Similar impulse response measurement system is most of without using complete acoustical signal denoising system both at home and abroad in the market,
It must be measured under conditions of signal-to-noise ratio is greater than 35dB.During this defect limit value significantly measuring system at the scene measures
Using.
6, there is data transmitting, acquisition, storage, processing analysis integration, avoid artificial treatment in traditional measurement method
Error and mistake caused by data significantly improve the precision and efficiency of measurement.Signal transmitting, based Denoising and arteries and veins of the present invention
The overall process of punching as the result is shown only needs 2 minutes time.It is mixed that 1/3OTC within the scope of 100~8000Hz is calculated using impulse response
Ringing the auditoriums parameters such as time only needs the shorter time.
The equipment for interrupting the method measurement reverberation time using steady-state noise both at home and abroad in the market measures 100~8000Hz range
The interior 1/3OTC reverberation time needs 30-45 minutes, and is not capable of measuring other acoustical parameters.
It is time-consuming and laborious in the market both at home and abroad using the equipment of nature pulse (stamp balloon, electric spark etc.) measurement impulse response,
And signal-to-noise ratio wretched insufficiency, especially low frequency part.Signal-to-noise ratio deficiency is also often resulted in laboratory, measures asking for failure
Topic.
7, process and result visualization are denoised, tester can decline bent by observation impulse response and each sound control
The form of line can tentatively judge the validity of test and denoising result, and discovery wave chopping, signal-to-noise ratio is insufficient, sound field diffusion level is poor
The problem of leading to false test result, in time adjustment test and denoising scheme, obtain correct impulse response test result.Original creation
Display interface parameter setting, crucial denoising and measurement result are subjected to image conversion show, make operator can be according to image reality
When judge denoising and measurement result it is whether correct, overcome both at home and abroad at present measuring system to signal carry out secret operation, cause
Tester has no way of determining the defect of measurement result reliability.
Detailed description of the invention
Fig. 1 is the structural schematic diagram of apparatus of the present invention；
Fig. 2 is the structural schematic diagram that signal time-frequency of the present invention denoises module；
Specific embodiment
The present invention is described in detail with specific embodiment below in conjunction with the accompanying drawings.
Embodiment
As shown in Figure 1, the device based on noise in time frequency analysis removal hall impulse response measurement acoustical signal, including signal
Transmitting module 1, signal acquisition module 2, signal time-frequency denoising module 3, impulse response computing module 4, the noise acoustical power that is averaged are gone
Module of making an uproar 5, as the result is shown and memory module 6；The signal emission module 1 is responsible for issuing the sound letter such as weighting frequency sweep into the hall
Number；The signal acquisition module 2, signal time-frequency denoising module 3, impulse response computing module 4, the noise acoustical power that be averaged denoise
Module 5 is sequentially connected with memory module 6 as the result is shown, and impulse response computing module is also connected with signal emission module；
The signal emission module includes sequentially connected digital signal generation and weighted units, digital-to-analogue conversion card, function
Rate amplifier and loudspeaker, to complete to emit the task of the acoustical signals such as frequency sweep or multifrequency sine into the hall.Wherein, the number
The starting and ending section of the generation of word signal and weighted units to limited Linear chirp in time domain and frequency domain in time domain carries out
Sine Modulated, makes the amplitude of signal smoothly decay to zero in starting point and end.
The signal acquisition module includes sequentially connected microphone, preamplifier, analog-to-digital conversion card and number letter
Acquisition signal to complete the task that sound pressure signal acquires in the hall, and is sent to signal time-frequency denoising module by number receiving unit.
The signal time-frequency denoising module includes time-frequency conversion submodule 31, threshold denoising submodule 32, useful signal
Reconstruct submodule 33 and time-frequency inverse transformation submodule 34.
11) the time-frequency conversion submodule 31 described in: the acoustic pressure time-domain signal of acquisition is transformed to time-frequency domain signal by the module,
Purpose is to make the noise being mixed into signal on time-frequency domain with Signal separator, and transform method is obtained using VARIANT FILTER OF DISCRETE GABOR TRANSFORM
To time-frequency distributions:
Wherein, W=exp (j2 π/N)；It is the amplitude of time frequency analysis coefficient；It is the acoustic pressure time domain letter of acquisition
Number；Δ M is time-domain analysis step-length；Δ N is frequency-domain analysis step-length；It is the analytic function of Gabor transformation.
12) threshold denoising submodule 32: according to the time-frequency domain distribution characteristics otherness of signal and noise to being calculated
Time-frequency conversion coefficient is handled, and determines the noise-removed threshold value in time-frequency domain, and the signal higher than threshold value retains, lower than the signal of threshold value
Zero setting constitutes new time-frequency conversion coefficient sets, forms useful signal.The module can effectively remove while making an uproar in different frequency domains
Sound.
13) useful signal reconstructs submodule 33: by several effective letters measuring in different periods, obtaining according to the above method
Number the amplitude of time-frequency conversion coefficient be compared, same time frequency point selects the smallest coefficient of energy magnitude (i.e. noise pollution journey
Spend the smallest signal), the time-frequency conversion coefficient sets reconstructed.The module part that can be removed is simultaneously the same as the noise signal in frequency domain
In unstable state component, for nonstationary noise denoising effect it is preferable；
14) time-frequency inverse transformation submodule 34: carrying out time-frequency inverse transformation to the time-frequency conversion coefficient sets that obtain after reconstruct, by when
Frequency impulse response signal transforms in time domain.It is obtained using Discrete Gabor Expansion:
Wherein, W=exp (j2 π/N)；It is the acquisition acoustic pressure time-domain signal after inverse transformation；cm,nIt is reconstruct time frequency analysis
The amplitude of coefficient；It is the composite function of Gabor transformation.
The noise acoustical power denoising module that is averaged is using the signal meter of impulse response latter stage noise in the highest flight
It calculates noise to be averaged acoustical power, impulse response is then converted into Energy impulse response, and in Energy impulse response all the period of time internal buckle
Except average noise acoustical power (steady-state component i.e. in noise).
Based on time frequency analysis removal hall impulse response measurement acoustical signal in noise device, selection signal transmission channel and
Acquisition channel；Input transmitting signal duration, acquisition signal duration, sample frequency, signal amplitude, swept-frequency signal initial frequency and
Cutoff frequency etc..
Window function type is selected when time-frequency based Denoising, window function width, time analysis step-length, frequency analysis step-length, is led to
It crosses antithesis window function calculating instrument case and analysis and composite function is calculated.
It is described as the result is shown and memory module includes sound pressure curve display unit before showing denoising collected, denoising
Sound pressure curve display unit after preceding impulse response display unit, denoising, the impulse response display unit after denoising；And it goes
The impulse response storage unit before sound pressure signal storage unit, denoising before making an uproar, the sound pressure signal storage unit after denoising, denoising
Impulse response storage unit afterwards.
Wherein, acquisition sound pressure curve display unit display acquisition sound pressure curve figure, the impulse response are shown
Unit shows acoustic pressure and Energy impulse response curve graph；The sound pressure signal storage unit is stored in collected sound in the hall
Press signal, the impulse response storage unit storage acoustic pressure pulse response.
Measuring process of the invention are as follows:
1) parameter setting: selection signal transmission channel and acquisition channel；A length of 10s, acquisition signal when input transmitting signal
Shi Changwei 20s, sample frequency are 44100Hz (should be greater than at least 2 times of highest cutoff frequency), signal amplitude 1, swept-frequency signal
Initial frequency is 20Hz and cutoff frequency is 20kHz, impulse response be truncated initial time be 9s and deadline be 20s, pulse
Arrival time criterion is 10dB, environment temperature is 25 DEG C, reverberation building volume is 268m3；
2) experiment is carried out in acoustic investigation institute of Tongji University reverberation chamber, by adjust the indoor absorption of reverberation come
The different reverberation time is obtained, wherein the reverberation time of 1000Hz is respectively 1.5s, 2.4s and 6.0s.When measuring RMR room reverb
Between when, within the scope of measurement frequency, ambient noise sound pressure level on microphone position should be at least lower than the sound pressure level that sound source generates
35dB。
3) experimental data shows useful signal after time frequency analysis, and the size of transformation coefficient has with frequency obviously at any time
Variation tendency, and the transformation coefficient of noise is evenly distributed on time frequency analysis figure and relatively small.The two is on time-frequency domain
There were significant differences for distribution.Useful signal and noise section can be obviously identified from the time frequency analysis figure of signals and associated noises.As a result,
We can artificially select according to the time frequency analysis index variation trend of useful signal and retain the time-frequency conversion of useful signal
Coefficient, and by the time frequency analysis coefficient zero setting of noise section, to realize signal denoising.
4) experiment also respectively processing analyze signal-to-noise ratio under the conditions of 1.5s, 2.4s and 6.0s reverberation time be respectively 25dB,
Signals and associated noises when 20dB, 15dB, 10dB and 5dB.Now by taking third-octave 1000Hz frequency band as an example, provide the different reverberation time,
The reverberation time value and its opposite mistake being calculated under the conditions of different signal-to-noise ratio using ideal signal, signals and associated noises and denoised signal
Difference, (note: the calculating of reverberation time utilizes impulse response reverse integral method to obtain attenuation curve, and calculates as shown in table 1-3
T20):
Table 1
Table 2
Table 3
It is above-mentioned the results showed that either under the conditions of short reverberation, middle reverberation or long reverberation, utilize the present invention to provide
Acoustical signal Denoising Algorithm denoising is carried out to signal, be attained by preferable denoising effect, illustrate that this method is practical, even if
It is (small 30dB more defined than specification) also available more believable test data, error under conditions of signal-to-noise ratio is 5dB
Within controlled range.
Claims (13)
1. a kind of method of noise in removal hall impulse response measurement acoustical signal, which comprises the following steps:
1) signal emission module is generated and weighted units, digital-to-analogue conversion card, power amplifier by sequentially connected digital signal
Weighting frequency sweep acoustical signal or multifrequency sine acoustical signal, the noise in the acoustical signal and the hall are issued into the hall with loudspeaker
There is difference in time-frequency domain distribution characteristics；
2) signal acquisition module passes through sequentially connected microphone, preamplifier, analog-to-digital conversion card and digital signal reception list
Member, acquires the sound pressure signal of the hall, and sends it to signal time-frequency denoising module；
3) signal time-frequency denoising module is mixed into signal by time frequency analysis, threshold denoising and the removal of useful signal reconstructing method
Different time Frequency domain noise, and time domain, hair are returned to the nonstationary noise component in time-frequency domain, and by the time frequency signal after denoising
Give impulse response computing module；
4) digital signal of sound pressure signal and frequency sweep or multifrequency sine acoustical signal after denoising is done phase by impulse response computing module
Close calculate, obtain the impulse response of the hall, and send responses to noise be averaged acoustical power denoise module；
5) the noise acoustical power denoising module that is averaged calculates steady-state noise acoustical power in impulse response first, then in energy pulse
It is deducted in response；
6) as the result is shown and memory module is shown in the sound pressure signal curve measured in the hall, acoustic pressure pulse response curve, energy
Pulse respond is measured, and stores its digital signal.
2. the method for noise, feature exist in a kind of removal hall impulse response measurement acoustical signal according to claim 1
In the digital signal generates and weighted units include digital signal generation unit and weighted units, the digital signal
The calculating process that generation unit generates signal is as follows, Linear chirp are as follows:
Wherein, A is signal amplitude；T is signal total duration；T is time variable；f1It is low cutoff frequency；f2It is higher cutoff frequency；
The calculating process that the weighted units are weighted signal is as follows:
Wherein, T is signal total duration；T is time variable；e0It (t) is Linear chirp；ω is angular frequency.
3. the method for noise, feature exist in a kind of removal hall impulse response measurement acoustical signal according to claim 2
In, the digital signal generate and weighted units to limited Linear chirp in time domain and frequency domain in the starting of time domain and
Ending segment has carried out Sine Modulated, and the amplitude of signal is made smoothly to decay to zero in starting point and end, and the principle of modulation is
Guarantee that modulated signal is not decayed substantially in the frequency range self-energy of measurement request, the frequency spectrum of filtering is avoided to let out
Dew, reduces the concussion of frequency domain amplitude.
4. the method for noise, feature exist in a kind of removal hall impulse response measurement acoustical signal according to claim 1
In the signal time-frequency denoising module includes time-frequency conversion submodule, threshold denoising submodule, useful signal reconstruct submodule
And time-frequency inverse transformation submodule, in which:
11) time-frequency conversion submodule, for the acoustic pressure time-domain signal of acquisition to be transformed to time-frequency domain signal, it is therefore an objective to make to be mixed into letter
For different time Frequency domain noise in number with Signal separator on time-frequency domain, transform method is to obtain time-frequency using VARIANT FILTER OF DISCRETE GABOR TRANSFORM
Distribution:
Wherein, W=exp (j2 π/N)；It is the amplitude of time frequency analysis coefficient；It is the acoustic pressure time-domain signal of acquisition；ΔM
It is time-domain analysis step-length；Δ N is frequency-domain analysis step-length；It is the analytic function of Gabor transformation；
12) threshold denoising submodule is made an uproar interior noise for effectively removing different time frequency domain, specifically: according to signal and noise
Time-frequency domain distribution characteristics otherness handles the time-frequency conversion coefficient being calculated, and determines the noise-removed threshold value in time-frequency domain,
Signal higher than threshold value retains, and lower than the signal zero setting of threshold value, constitutes new time-frequency conversion coefficient sets, forms useful signal；
13) useful signal reconstructs submodule, the time-frequency conversion system of several useful signals for will obtain in different periods measurement
Several amplitudes are compared, the same time frequency point selection the smallest coefficient of energy magnitude, the time-frequency conversion coefficient sets reconstructed,
The middle the smallest coefficient of energy magnitude is the smallest signal of noise pollution degree, which can effectively remove noise in same time-frequency domain
Unstable state component；
14) time-frequency inverse transformation submodule, for carrying out time-frequency inverse transformation to the time-frequency conversion coefficient sets obtained after reconstruct, by time-frequency
Impulse response signal transforms in time domain, is obtained using Discrete Gabor Expansion:
Wherein, W=exp (j2 π/N)；It is the acoustic pressure time-domain signal after inverse transformation；cm,nIt is the time frequency analysis coefficient of reconstruct
Amplitude；It is the composite function of Gabor transformation.
5. the method for noise, feature exist in a kind of removal hall impulse response measurement acoustical signal according to claim 4
In the impulse response computing module is to utilize the digital linear swept-frequency signal e modulated in signal emission module0(t) and institute
Sound pressure signal y (t) after received time frequency analysis denoising makees relevant calculation, and the noisy acoustic pressure pulse for obtaining measuring point about room is rung
P (t) is answered, wherein y (t) is the acoustic pressure time-domain signal after inverse transformationContaining part simultaneously with frequency noise n (t):
y(t)*e0 -1(t)=e0(t)*p(t)*e0 -1(t)+n(t)*e0 -1(t)=p (t)+n (t) * e0 -1(t)
P (t) still with noise n (t) * e0 -1(t) it mixes, influences measurement accuracy, wherein n (t) * e0 -1(t) basic expressions
For the steady-state noise in statistical significance.
6. the method for noise, feature exist in a kind of removal hall impulse response measurement acoustical signal according to claim 4
In, the noise be averaged acoustical power denoising module working principle it is as follows:
Noise is calculated using impulse response latter stage noise signal in the highest flight to be averaged acoustical power, then converts impulse response
It for Energy impulse response, and deducts within Energy impulse response all the period of time noise and is averaged acoustical power, detailed process is as follows:
Energy impulse response is integrated using time sliding window exponential smoothing, the Energy impulse response sound for obtaining smoothing can decline
Subtract curve, and deducts in entirely integral time domain noise in window and be averaged acoustical power, the steady-state component in removal noise, in the hall
Impulse response p (t) is made of signal and steady-state noise:
P (t)=p (0) e-σt+N(t)
P (0) is the initial value of impulse response function；N (t) is steady-state noise, i.e. noise n (t) * e0 -1(t)；
Latter stage signal prevailing for noise can ask its average acoustical power to obtain:
Wherein t1At the time of capable of being flooded by ambient noise for signal sound, T t1The length of noise total duration after moment, T enables toRelatively stable estimated value is obtained, length is T as a result,0Sound in sliding window can indicate are as follows:
After the average acoustical power for removing noise, then Energy impulse response sound energy attenuation curve is obtained are as follows:
7. a kind of device for implementing the method for noise in removal hall impulse response measurement acoustical signal described in claim 1,
It is characterized in that, including signal emission module, signal acquisition module, signal time-frequency denoise module, impulse response computing module, noise
Average acoustical power denoising module and as the result is shown and memory module；The signal emission module is responsible for issuing weighting into the hall
Frequency sweep acoustical signal；The signal acquisition module, signal time-frequency denoising module, impulse response computing module, noise are averaged sound function
Rate denoising module is sequentially connected with memory module as the result is shown, the impulse response computing module also with signal emission module
It is connected.
8. device according to claim 7, which is characterized in that the signal emission module includes sequentially connected number
Signal generates and weighted units, digital-to-analogue conversion card, power amplifier and loudspeaker, for completing the transmitting weighting frequency sweep into the hall
Or the task of multifrequency sine acoustical signal, the digital signal generates and weighted units are linearly swept to limited in time domain and frequency domain
Frequency signal has carried out Sine Modulated in the starting and ending section of time domain respectively, keeps the amplitude of signal smooth in starting point and end
Ground decays to zero；
The signal acquisition module includes that sequentially connected microphone, preamplifier, analog-to-digital conversion card and digital signal connect
Unit is received, to complete the task that sound pressure signal acquires in the hall, and acquisition signal is sent to signal time-frequency denoising module.
9. device according to claim 7, which is characterized in that the signal time-frequency denoising module includes time-frequency conversion
Module, threshold denoising submodule, useful signal reconstruct submodule and time-frequency inverse transformation submodule, in which:
Time-frequency conversion submodule, for the acoustic pressure time-domain signal of acquisition to be transformed to time-frequency domain signal；
Threshold denoising submodule, for being become according to the time-frequency domain distribution characteristics otherness of signal and noise to the time-frequency being calculated
It changing coefficient to be handled, determines the noise-removed threshold value in time-frequency domain, the signal higher than threshold value retains, lower than the signal zero setting of threshold value,
New time-frequency conversion coefficient sets are constituted, useful signal is formed；
Useful signal reconstructs submodule, the time-frequency conversion coefficient of several useful signals for will obtain in different periods measurement
Amplitude is compared, and same time frequency point selects the smallest coefficient of energy magnitude, the time-frequency conversion coefficient sets reconstructed, wherein can
The amount the smallest coefficient of amplitude is the smallest signal of noise pollution degree；
Time-frequency inverse transformation submodule, for carrying out time-frequency inverse transformation to the time-frequency conversion coefficient sets obtained after reconstruct.
10. device according to claim 7, which is characterized in that the noise is averaged acoustical power denoising module using arteries and veins
Punching response latter stage noise signal in the highest flight calculates noise and is averaged acoustical power, and impulse response is then converted to energy pulse
Response, and average noise acoustical power is deducted within Energy impulse response all the period of time.
11. device according to claim 7, which is characterized in that selection signal transmission channel and acquisition channel；Input transmitting
Signal duration, acquisition signal duration, sample frequency, signal amplitude, the initial frequency of swept-frequency signal and cutoff frequency.
12. device according to claim 7, which is characterized in that select window function type, window function wide when time frequency analysis
Analysis and composite function is calculated by antithesis window function calculating instrument case in degree, time analysis step-length, frequency analysis step-length.
13. device according to claim 7, which is characterized in that it is described as the result is shown and memory module includes display institute
The impulse response display unit before sound pressure curve display unit, denoising, the sound pressure curve after denoising before the denoising of acquisition are shown
Impulse response display unit after unit, denoising；And the sound pressure signal storage unit before denoising, the impulse response before denoising are deposited
The impulse response storage unit after sound pressure signal storage unit, denoising after storage unit, denoising；
Wherein, the sound pressure curve display unit display acquisition sound pressure curve figure before the display denoising collected, it is described
Impulse response display unit shows acoustic pressure and Energy impulse response curve graph；The sound pressure signal storage unit is stored in the hall
Interior collected sound pressure signal, the impulse response storage unit storage acoustic pressure pulse response.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510650988.0A CN105371945B (en) | 2015-10-10 | 2015-10-10 | Remove the method and device of noise in hall impulse response measurement acoustical signal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510650988.0A CN105371945B (en) | 2015-10-10 | 2015-10-10 | Remove the method and device of noise in hall impulse response measurement acoustical signal |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105371945A CN105371945A (en) | 2016-03-02 |
| CN105371945B true CN105371945B (en) | 2019-01-25 |
Family
ID=55374347
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510650988.0A Active CN105371945B (en) | 2015-10-10 | 2015-10-10 | Remove the method and device of noise in hall impulse response measurement acoustical signal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105371945B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107785025B (en) * | 2016-08-25 | 2021-06-22 | 上海英波声学工程技术股份有限公司 | Noise removal method and device based on repeated measurement of room impulse response |
| CN107941324A (en) * | 2017-10-08 | 2018-04-20 | 天津大学 | A kind of method of estimation of consumer level inertia sensing unit environments noise |
| CN108010536B (en) * | 2017-12-05 | 2020-07-14 | 深圳市声扬科技有限公司 | Echo cancellation method, device, system and storage medium |
| CN109246573B (en) * | 2018-10-08 | 2020-10-27 | 北京铸声场传媒科技有限公司 | Method and device for measuring frequency response characteristic of audio system |
| CN110445556B (en) * | 2019-07-30 | 2022-03-22 | 成都安杰联科技有限公司 | Device and method for detecting and positioning ETC (electronic toll Collection) interference unit |
| CN113219252B (en) * | 2021-05-14 | 2022-03-01 | 浙江大学 | Method for non-invasive online detection of converter impedance |
| CN114783463B (en) * | 2022-03-08 | 2024-04-26 | 同济大学 | Hall objective tone quality parameter measurement method and system based on impulse response noise reduction |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103050127A (en) * | 2012-11-19 | 2013-04-17 | 上海英波声学工程技术有限公司 | Hall objective audio quality parameter measuring system and method |
| CN103675097A (en) * | 2013-11-25 | 2014-03-26 | 同济大学 | Measuring system for material normal direction incidence acoustical absorption coefficient and acoustic impedance |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH06194217A (en) * | 1992-12-24 | 1994-07-15 | Sekisui House Ltd | Residence acoustic performance measuring device on site |
| JPH07334177A (en) * | 1994-06-08 | 1995-12-22 | Matsushita Electric Ind Co Ltd | Sound field analyzing device |
-
2015
- 2015-10-10 CN CN201510650988.0A patent/CN105371945B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103050127A (en) * | 2012-11-19 | 2013-04-17 | 上海英波声学工程技术有限公司 | Hall objective audio quality parameter measuring system and method |
| CN103675097A (en) * | 2013-11-25 | 2014-03-26 | 同济大学 | Measuring system for material normal direction incidence acoustical absorption coefficient and acoustic impedance |
Non-Patent Citations (1)
| Title |
|---|
| 基于神经网络和遗传算法的厅堂音质最优化研究;陈婧等;《声学技术》;20120831;第31卷(第4期);第27-30页 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105371945A (en) | 2016-03-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105371945B (en) | Remove the method and device of noise in hall impulse response measurement acoustical signal | |
| CN106768266B (en) | A kind of measurement method of workshop reverberation time that rejecting background noise energy | |
| CN101194535B (en) | Method for correcting electroacoustic converter acoustic paramenter and device accomplishing the method | |
| CN101426169B (en) | Time-domain tracking filter quickly detecting acoustic response parameter of sounding body and system | |
| CN105792072B (en) | Sound effect processing method and device and terminal | |
| WO2022012195A1 (en) | Audio signal processing method and related apparatus | |
| CN106031197B (en) | Acoustic treatment equipment, Disposal of Acoustics and Acoustic treatment program | |
| CN109831733A (en) | Test method, device, equipment and the storage medium of audio broadcast performance | |
| Antweiler et al. | Perfect-sweep NLMS for time-variant acoustic system identification | |
| CN107785025B (en) | Noise removal method and device based on repeated measurement of room impulse response | |
| RU2411595C2 (en) | Improved intelligibility of speech in mobile communication device by control of vibrator operation depending on background noise | |
| CN104729677B (en) | A kind of time-domain digital weighted method of nonstationary noise signal | |
| US20120143553A1 (en) | Method for Determining an Averaged Frequency-Dependent Transmission Function for a Disturbed Linear Time-Invariant System, Evaluation Device and Computer Program Product | |
| Müller | Measuring transfer-functions and impulse responses | |
| JP5077847B2 (en) | Reverberation time estimation apparatus and reverberation time estimation method | |
| Hahn et al. | Continuous measurement of impulse responses on a circle using a uniformly moving microphone | |
| JP2019032242A (en) | Attenuation time analytic method, device, and program | |
| Policardi | MLS and Sine-Sweep measurements | |
| Olsson et al. | Incremental use of FFT as a solution for low BT-product reverberation time measurements | |
| CN112992182B (en) | Vehicle wind noise level testing system and testing method thereof | |
| CN104034970A (en) | Sound level measurement method based on computer sound card | |
| JP2932996B2 (en) | Harmonic pitch detector | |
| Garai et al. | Optimizing the exponential sine sweep (ESS) signal for in situ measurements on noise barriers | |
| Chan | Swept sine chirps for measuring impulse response | |
| JPH04295727A (en) | Impulse-response measuring method |
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 | ||
| CP03 | Change of name, title or address | ||
| CP03 | Change of name, title or address |
Address after: 200092 Siping Road 1239, Shanghai, Yangpu District Patentee after: TONGJI University Country or region after: China Patentee after: SHANGHAI YINGBO ACOUSTIC ENGINEERING TECHNOLOGY CO.,LTD. Address before: 200092 Siping Road 1239, Shanghai, Yangpu District Patentee before: TONGJI University Country or region before: China Patentee before: SHANGHAI IMPULSE ACOUSTIC ENGINEERING TECHNOLOGY CO.,LTD. |