CN104729677B - A kind of time-domain digital weighted method of nonstationary noise signal - Google Patents
A kind of time-domain digital weighted method of nonstationary noise signal Download PDFInfo
- Publication number
- CN104729677B CN104729677B CN201510067451.1A CN201510067451A CN104729677B CN 104729677 B CN104729677 B CN 104729677B CN 201510067451 A CN201510067451 A CN 201510067451A CN 104729677 B CN104729677 B CN 104729677B
- Authority
- CN
- China
- Prior art keywords
- mrow
- msub
- weighted
- mtd
- time
- 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
Abstract
The present invention relates to a kind of time-domain digital weighted method of nonstationary noise signal, it comprises the following steps：1) data acquisition is carried out to nonstationary noise signal, obtains noisy digit signal p (τ)；2) weighted amplitude curve A (f) is built according to the octave sound level correction value of noise weighting network；3) mirror transformation processing is carried out to weighted amplitude curve A (f), obtains the aperiodic real even function A'(f of frequency domain weighting network)：4) inverse Fourier Tranform is carried out, the impulse response function A'(t of weighting network is obtained)；5) window function W is chosenβ(t) to A'(t) carry out adding window truncation；6) adding window weighted wavelet function A' is obtainedwβ(t)；7) row interpolation resampling is entered to adding window weighted wavelet function, makes the time interval of raw noise data signal of its time interval with gathering consistent；8) correlation ratio is carried out to conversion to the nonstationary noise data signal p (τ) and adding window weighted wavelet function that collect, obtains time domain weighted fluctuation signal pw(t)；9) the weighted overall sound pressure level curve L of non-stationary signal is obtained by sound level transformation calculationsw(t)。
Description
Technical field
The present invention relates to a kind of noise signal analysis processing method, become rapidly with the time especially with regard to a kind of frequency content
The time-domain digital weighted method of the nonstationary noise signal of change.
Background technology
Human ear is larger for the dynamic range of perception of sound, the loudness size that auditory perception is arrived and acoustic pressure than logarithm value into
Certain proportionate relationship, thus reflected in acoustics with the concept of sound pressure level human ear to loudness from perception by feature.In addition, by
The particularity constructed in human ear, human ear has different sensitivity to the noise of different frequency, in order to obtain the subjective sensation with people
The evaluation criterion being consistent, makes the decibel value that measurement is obtained have certain correlation, international standard with the subjective loudness experienced
Change tissue and apply different frequency band weighteds processing to sound according to the characteristics of equal loudness contour, conventional weighted has A, B, C tri- at present
Kind.To in the processing of voice signal, the calculating of sound pressure level can be according to theoretical formula method, but the weighted sound pressure level of acoustical signal
But there is more problem in processing.Traditional weighted sound pressure level processing method is to use analog filter, by hardware to different frequencies
The sound pressure signal of band carries out the decay of different amplitudes, and the shortcoming of this method is the difficulty of parameter tuning of analog circuit, and is made
It can all cause the change of device parameters with changes in environmental conditions, the heating of device, aging etc..
With the development of computer and signal processing technology, digital method start to be used for the analysis of noise signal with
Processing.For stationary signal, fast fourier transform can be for carrying out weighted calculating to acoustical signal, and obtains preferably essence
Degree.And for non-stationary signal, people carry out weighted processing to acoustical signal frequently with Fourier Tranform in short-term, but it is due to Fu Li
Leaf transformation is the transform method of linear permanent resolution ratio, during processing non-stationary time varying signal, it may appear that the problem of spectrum leakage, and then
Weighted result is set to produce larger error.Small wave converting method is a kind of transform method of non-linear multiresolution, handles non-stationary
With very big advantage during time varying signal.But existing wavelet transformation weighted method is that known wavelet function is changed
Make, processing is filtered respectively according to weighted subband is wide, sound pressure level amendment is then carried out respectively, finally to the weighted sound of each subband
Arbitrarily downgrade and sum up, so as to obtain total weighted sound pressure level, such as：Wang Guangsen etc., the weighted sound level based on orthogonal wavelet transformation
Measurement, Chinese journal of scientific instrument, 2004.This computational methods very complicated calculates (quantity of subband), it is necessary to carry out repeatedly filtering,
It is computationally intensive, while the problem of also there is subband spectrum energy aliasing and bring calculation error.In addition, this method can not be obtained
The weighted time domain fluctuation signal of sound.
The content of the invention
Spectrum leakage can not only be effectively solved the purpose of the present invention is to propose to one kind, reduce weighted calculation error, and
The time-domain digital weighted processing method of the nonstationary noise signal of computational efficiency can be effectively improved.
To achieve the above object, the present invention takes following technical scheme：A kind of time-domain digital meter of nonstationary noise signal
Power method, it comprises the following steps：
1) using frequency acquisition as fsData acquisition is carried out to nonstationary noise signal, nonstationary noise data signal p is obtained
(τ)；
2) weighted amplitude curve A (f) is built according to the octave sound level correction value of noise weighting network：
Where it is assumed that the octave sound level correction value of noise weighting network is Δ Lw[fc(i)], fc(i) it is OcClass octave
In i-th of subband centre frequency, i=1,2 ... .N；
3) mirror transformation processing is carried out to weighted amplitude curve A (f), obtains the aperiodic real even function of frequency domain weighting network
A'(f)：
4) to aperiodic real even function A'(f) inverse Fourier Tranform is carried out, obtain the impulse response function A' of weighting network
(t)：
Wherein, t is time series,It is inverse fourier operator, i.e.,
5) to impulse response function A'(t) carry out adding window truncation, the window function W of selectionβ(t) it is：
Wherein, β is window enhancing rate, TwIt is window width；
6) adding window weighted wavelet function, i.e. simultaneous formula (3) are obtained and formula (4) obtains adding window weighted wavelet function A'wβ(t)
For：
A'wβ(t)=Wβ(t)A'(t) (5)
7) row interpolation resampling is entered to adding window weighted wavelet function, the adding window weighted wavelet function A " after resamplingwβ
(tre) be：
A″wβ(tre)=Interp [A'wβ(t),tre] (6)
Wherein, Interp represents interpolating function；treResampling time series is represented, its time is at intervals of 1/fs；
8) the nonstationary noise data signal p (τ) and the adding window weighted wavelet function after resampling that collect are entered
Row correlation ratio obtains time domain weighted fluctuation signal p to conversionw(t) it is：
In formula, τ is comparison time sequence；
9) time domain weighted fluctuation signal is obtained to the weighted overall sound pressure level curve of non-stationary signal by sound level transformation calculations
Lw(t) it is：
The step 4) in, when carrying out inverse Fourier Tranform, discretization is carried out to frequency f and time t respectively, wherein, frequency
Rate f is discrete thinner, and impulse response function is more accurate, and time t discrete steps are smaller, and the waveform for describing impulse response function is accurate.
In actual applications, f is discrete refines as far as possible, and time t discrete steps are chosen as needed, and selection principle is small
In step 7) in interpolation resampling time interval.
The present invention is due to taking above technical scheme, and it has advantages below：1st, the present invention is first according to noise weighting network
The octave sound level correction value of network, which builds weighted amplitude curve and mirror transformation, inverse fourier are carried out to weighted amplitude curve, to be become
Change, adding window is blocked, adding window weighted wavelet function and resampling processing, then by the nonstationary noise data signal collected with
Adding window weighted wavelet function after resampling carries out correlation ratio pair and sound level transformation calculations obtain the weighted of non-stationary signal
Overall sound pressure level curve, compared with prior art, the present invention do not need hardware filter, directly to nonstationary noise signal at
Reason, therefore the reliability of weighted is effectively increased, do not influenceed by environmental factor and device aging, while weighted can be reduced
The cost of device.2nd, the present invention to non-stationary time-varying noise signal using Fourier Tranform in short-term it is possible to prevente effectively from carry out weighted
When, by the calculation error that frequency leakage is brought, substantially increase the precision of nonstationary noise signal weighted.3rd, the present invention can
To directly obtain the wavelet function of weighting network, filtering is calculated so as to carry out a correlation ratio, you can obtain the sound after weighted
Buckle line, enormously simplify calculating, improve weighted computational efficiency.4th, the present invention can obtain the fluctuation letter of the time domain after weighted
Number, it is easy to go to analyze the main noise source of sounding body from the angle closer to people's auditory perception, is other more professional make an uproar
Sound analysis technology provides the weighted noise time-domain signal on basis.It present invention can be widely used to the weighted processing of various noise signals
During.
Brief description of the drawings
Fig. 1 is the inventive method principle schematic；
Fig. 2 is the handling process schematic diagram of the specific embodiment of the invention；
Fig. 3 is the A weighted overall sound pressure level curve synoptic diagrams obtained using Fourier Tranform；
Fig. 4 is the A weighted overall sound pressure level curve synoptic diagrams obtained using the present invention.
Embodiment
Come to carry out the present invention detailed description below in conjunction with accompanying drawing.It should be appreciated, however, that accompanying drawing has been provided only more
Understand the present invention well, they should not be interpreted as limitation of the present invention.
As shown in Figure 1 and Figure 2, the time-domain digital weighted method of nonstationary noise signal of the invention, comprises the following steps：
1) with the frequency acquisition f of settingsNonstationary noise signal is gathered, nonstationary noise data signal p (τ) is obtained；
2) weighted amplitude curve A (f) is built according to the octave sound level correction value of noise weighting network, wherein, if noise
The octave sound level correction value of weighting network is Δ Lw[fc(i)], fc(i) it is OcThe center frequency of i-th of subband in class octave
Rate, i=1,2 ... .N, build formula as follows：
3) mirror transformation processing is carried out to weighted amplitude curve A (f), obtains the aperiodic real even function of frequency domain weighting network
A'(f), its expression formula is as follows：
4) to the aperiodic real even function A'(f of frequency domain weighting network) inverse Fourier Tranform is carried out, obtain weighting network
Impulse response function A'(t), as shown in formula (3)：
Wherein, t is time series, A'(t) be weighting network impulse response function,It is inverse fourier operator,
I.e.
5) due to impulse response function A'(t) it is unlimited in time, it is impossible to weighted calculating is participated in, thus to A'(t)
Adding window truncation is carried out, the window function of selection is as follows：
Wherein, Wβ(t) it is window function, β is window enhancing rate, TwIt is window width.
6) calculating, which obtains adding window weighted wavelet function, i.e. simultaneous formula (3) and formula (4), can obtain adding window weighted small echo letter
Number, is shown below：
A'wβ(t)=Wβ(t)A'(t) (5)
Wherein, A'wβ(t) it is adding window weighted wavelet function.
7) row interpolation resampling is entered to adding window weighted wavelet function, believes its time interval and the raw noise numeral of collection
Number time interval it is consistent, be shown below：
A″wβ(tre)=Interp [A'wβ(t),tre] (6)
Wherein, A "wβ(tre) it is adding window weighted wavelet function after resampling, Interp represents interpolating function, treRepresent
Resampling time series, its time is at intervals of 1/fs。
8) the nonstationary noise data signal p (τ) and the adding window weighted wavelet function after resampling that collect are entered
Row correlation ratio obtains time domain weighted fluctuation signal, is shown below to conversion：
Wherein, pw(t) it is time domain weighted fluctuation signal, τ is comparison time sequence.
9) time domain weighted fluctuation signal is obtained to the weighted overall sound pressure level curve of non-stationary signal by sound level transformation calculations：
Wherein, Lw(t) be non-stationary signal weighted overall sound pressure level curve.
In a preferred embodiment, step 4) in when carrying out inverse Fourier Tranform, frequency f and time t are entered respectively
Row discretization, discrete thinner of frequency f, impulse response function is more accurate, and time t discrete steps is smaller, careful can more retouch
Paint the waveform of impulse response function.In actual applications, f is discrete refines as far as possible, and time t discrete steps can be according to need
Chosen, principle is less than resampling time interval.
The time-domain digital weighted method of the nonstationary noise signal of the present invention is carried out below by specific embodiment detailed
Explanation：
As shown in Fig. 2 because A-weighted sound level can preferably reflect subjective sensation of the people to noise, in noise testing,
A-weighted sound level is often used as the leading indicator of noise rating.It is thus specific in the embodiment of the present invention to use A weighted third-octaves
Network is calculated, and detailed process is：
1) with frequency acquisition fsTailpipe radiated noise of a certain car under accelerating mode is acquired, obtained non-
F in stationary noise data signal p (τ), the present embodimentsFor 16384Hz；
2) according to the third-octave sound level correction value structure A weighted amplitude curve A (f) (as shown in table 1) of A weighting networks：
Wherein, the correction value in 20~20000Hz of human ear audible frequency range is chosen to build A weighted amplitude curves A
(f) numerical value of the frequency-portions beyond being gone beyond the scope in curve A (f), is set to 0 into processing.
The third-octave sound level correction value of table 1A weighting networks
Note：Table 1 is the international standard (IEC standard) of A weighted amendments
3) mirror transformation processing is carried out to A weighted amplitude curve A (f), obtains the aperiodic real even letter of frequency domain A weighting networks
Number A'(f)：
4) to the aperiodic real even function A'(f of A frequency domain weighting networks) inverse Fourier Tranform is carried out, obtain weighting network
Impulse response function A'(t)：
When carrying out numerical computations, discretization is carried out to frequency f and time t respectively.Wherein, frequency f is discrete in the present embodiment
Interval delta f be 0.5Hz；And time interval discrete time t is 1/65536s.
5) to A'(t) adding window truncation is carried out, beta window can be very good to suppress the distortion of gibbs ripple, thus this reality
Apply selection beta window function in example：
Wherein, Wβ(t) it is beta window function, β is window enhancing rate, and its value is β=0.54/0.46, TwIt is the window of beta window
It is wide.
6) simultaneous formula (3) and formula (4) can obtain the adding window weighted wavelet function of A weighteds：
A'wβ(t)=Wβ(t)A'(t) (5)
Wherein, A'wβ(t) be A weighteds adding window weighted wavelet function.
7) cubic spline interpolation resampling is carried out to the adding window weighted wavelet function of A weighteds, makes its time interval and collection
Raw noise data signal time interval it is consistent：
A″wβ(tre)=Interp [A'wβ(t),tre,spline] (6)
8) the nonstationary noise data signal p (τ) and the A weighted adding window weighted small echo letters after resampling to collecting
Number A "wβ(tre) correlation ratio is carried out to conversion, obtain the time domain weighted fluctuation signal of A weighteds：
9) by the time domain weighted fluctuation signal p of A weightedsw(t) non-stationary tailpipe spoke is obtained by sound level transformation calculations
Penetrate the A weighted overall sound pressure level curves L of noise signalw(t), specifically calculate as shown in formula (8)：
The A weighted overall sound pressure level curves L that calculating can be obtained in actual usew(t) it is analyzed, for evaluating
The quality of different acoustic wave filter structure acoustic attenuation performances, and judge whether the soundproof effect of silencer meets standard requirement.In addition, point
Analysis calculates obtained A weighted overall sound pressure level curve tendencies, can also principium identification tailpipe noise whether can better meet
The subjective feeling of human ear.
In summary, the A weighted overall sound pressure level curves (as shown in Figure 3) and the use present invention obtained using Fourier Tranform
The A weighted overall sound pressure level curve comparison figures (as shown in Figure 4) that the time-domain digital weighted method of proposition is obtained, can from figure
Go out, for nonstationary noise signal, the fluctuation of higher magnitude is occurred in that using the weighted sound pressure level curve of Fourier Tranform, it is former
Weighted error caused by being spectrum leakage；And use the weighted curve that time-domain digital weighted method proposed by the present invention is obtained then
It is relatively smooth, there is not the fluctuation of higher magnitude, effectively increase computational accuracy.
The various embodiments described above are merely to illustrate the present invention, and wherein each implementation steps of method etc. are all to be varied from
, every equivalents carried out on the basis of technical solution of the present invention and improvement should not exclude the protection in the present invention
Outside scope.
Claims (3)
1. a kind of time-domain digital weighted method of nonstationary noise signal, it comprises the following steps：
1) using frequency acquisition as fsData acquisition is carried out to nonstationary noise signal, nonstationary noise data signal p (τ) is obtained；
2) weighted amplitude curve A (f) is built according to the octave sound level correction value of noise weighting network：
<mrow>
<mi>A</mi>
<mrow>
<mo>(</mo>
<mi>f</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<msup>
<mn>10</mn>
<mrow>
<mo>{</mo>
<msub>
<mi>&Delta;L</mi>
<mi>w</mi>
</msub>
<mo>[</mo>
<msub>
<mi>f</mi>
<mi>c</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
<mo>]</mo>
<mo>/</mo>
<mn>20</mn>
<mo>}</mo>
</mrow>
</msup>
<mi>f</mi>
<mo>&Element;</mo>
<mo>[</mo>
<msup>
<mn>2</mn>
<mrow>
<mo>-</mo>
<msub>
<mi>O</mi>
<mi>c</mi>
</msub>
<mo>/</mo>
<mn>2</mn>
</mrow>
</msup>
<msub>
<mi>f</mi>
<mi>c</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
<mo>,</mo>
<msup>
<mn>2</mn>
<mrow>
<msub>
<mi>O</mi>
<mi>c</mi>
</msub>
<mo>/</mo>
<mn>2</mn>
</mrow>
</msup>
<msub>
<mi>f</mi>
<mi>c</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
<mo>]</mo>
<mo>;</mo>
<mi>i</mi>
<mo>=</mo>
<mn>1,2</mn>
<mo>,</mo>
<mo>.</mo>
<mo>.</mo>
<mo>.</mo>
<mi>N</mi>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</mrow>
Where it is assumed that the octave sound level correction value of noise weighting network is Δ Lw[fc(i)], fc(i) it is OcI-th in class octave
The centre frequency of individual subband, i=1,2 ... .N；
3) mirror transformation processing is carried out to weighted amplitude curve A (f), obtains the aperiodic real even function A' of frequency domain weighting network
(f)：
<mrow>
<msup>
<mi>A</mi>
<mo>&prime;</mo>
</msup>
<mrow>
<mo>(</mo>
<mi>f</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfenced open='{' close=''>
<mtable>
<mtr>
<mtd>
<mi>A</mi>
<mrow>
<mo>(</mo>
<mi>f</mi>
<mo>)</mo>
</mrow>
</mtd>
<mtd>
<mi>f</mi>
<mo>&Element;</mo>
<mo>[</mo>
<mn>0</mn>
<mo>,</mo>
<msup>
<mn>2</mn>
<mrow>
<msub>
<mi>O</mi>
<mi>c</mi>
</msub>
<mo>/</mo>
<mn>2</mn>
</mrow>
</msup>
<msub>
<mi>f</mi>
<mi>c</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>N</mi>
<mo>)</mo>
</mrow>
<mo>]</mo>
</mtd>
</mtr>
<mtr>
<mtd>
<mi>A</mi>
<mrow>
<mo>(</mo>
<mo>-</mo>
<mi>f</mi>
<mo>)</mo>
</mrow>
</mtd>
<mtd>
<mi>f</mi>
<mo>&Element;</mo>
<mo>[</mo>
<mo>-</mo>
<msup>
<mn>2</mn>
<mrow>
<msub>
<mi>O</mi>
<mi>c</mi>
</msub>
<mo>/</mo>
<mn>2</mn>
</mrow>
</msup>
<msub>
<mi>f</mi>
<mi>c</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>N</mi>
<mo>)</mo>
</mrow>
<mo>,</mo>
<mn>0</mn>
<mo>)</mo>
</mtd>
</mtr>
<mtr>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<mi>else</mi>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>2</mn>
<mo>)</mo>
</mrow>
</mrow>
4) to aperiodic real even function A'(f) inverse Fourier Tranform is carried out, obtain the impulse response function A'(t of weighting network)：
Wherein, t is time series,It is inverse fourier operator, i.e.,
5) to impulse response function A'(t) carry out adding window truncation, the window function W of selectionβ(t) it is：
<mrow>
<msub>
<mi>W</mi>
<mi>&beta;</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfenced open='{' close='-'>
<mtable>
<mtr>
<mtd>
<mfrac>
<mrow>
<mi>&beta;</mi>
<mo>-</mo>
<mi>cos</mi>
<mo>[</mo>
<mn>2</mn>
<mi>&pi;</mi>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>+</mo>
<msub>
<mi>T</mi>
<mi>w</mi>
</msub>
<mo>/</mo>
<mn>2</mn>
<mo>)</mo>
</mrow>
<mo>/</mo>
<msub>
<mi>T</mi>
<mi>w</mi>
</msub>
<mo>]</mo>
</mrow>
<mrow>
<mn>1</mn>
<mo>+</mo>
<mi>&beta;</mi>
</mrow>
</mfrac>
</mtd>
<mtd>
<mi>t</mi>
<mo>&Element;</mo>
<mo>[</mo>
<mo>-</mo>
<msub>
<mi>T</mi>
<mi>w</mi>
</msub>
<mo>/</mo>
<mn>2</mn>
<mo>,</mo>
<msub>
<mi>T</mi>
<mi>w</mi>
</msub>
<mo>/</mo>
<mn>2</mn>
<mo>]</mo>
</mtd>
</mtr>
<mtr>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<mi>else</mi>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>4</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, β is window enhancing rate, TwIt is window width；
6) adding window weighted wavelet function, i.e. simultaneous formula (3) are obtained and formula (4) obtains adding window weighted wavelet function A'wβ(t) it is：
A′wβ(t)=Wβ(t)A′(t) (5)
7) row interpolation resampling is entered to adding window weighted wavelet function, the adding window weighted wavelet function A " after resamplingwβ(tre) be：
A″wβ(tre)=Interp [A 'wβ(t), tre] (6)
Wherein, Interp represents interpolating function；treResampling time series is represented, its time is at intervals of 1/fs；
8) phase is carried out to the nonstationary noise data signal p (τ) and the adding window weighted wavelet function after resampling that collect
Close and compare conversion, obtain time domain weighted fluctuation signal pw(t) it is：
<mrow>
<msub>
<mi>p</mi>
<mi>w</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<msubsup>
<mo>&Integral;</mo>
<mrow>
<mi>t</mi>
<mo>-</mo>
<msub>
<mi>T</mi>
<mi>w</mi>
</msub>
<mo>/</mo>
<mn>2</mn>
</mrow>
<mrow>
<mi>t</mi>
<mo>+</mo>
<msub>
<mi>T</mi>
<mi>w</mi>
</msub>
<mo>/</mo>
<mn>2</mn>
</mrow>
</msubsup>
<mi>p</mi>
<mrow>
<mo>(</mo>
<mi>&tau;</mi>
<mo>)</mo>
</mrow>
<msubsup>
<mi>A</mi>
<mi>w&beta;</mi>
<mrow>
<mo>&prime;</mo>
<mo>&prime;</mo>
</mrow>
</msubsup>
<mrow>
<mo>(</mo>
<mi>&tau;</mi>
<mo>-</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
<mi>d&tau;&tau;</mi>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>7</mn>
<mo>)</mo>
</mrow>
</mrow>
In formula, τ is comparison time sequence；
9) time domain weighted fluctuation signal is obtained to the weighted overall sound pressure level curve L of non-stationary signal by sound level transformation calculationsw(t)
For：
<mrow>
<msub>
<mi>L</mi>
<mi>w</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<msqrt>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>+</mo>
<msub>
<mi>T</mi>
<mi>w</mi>
</msub>
<mo>)</mo>
</mrow>
<msubsup>
<mo>&Integral;</mo>
<mrow>
<mi>t</mi>
<mo>-</mo>
<msub>
<mi>T</mi>
<mi>w</mi>
</msub>
<mo>/</mo>
<mn>2</mn>
</mrow>
<mrow>
<mi>t</mi>
<mo>+</mo>
<msub>
<mi>T</mi>
<mi>w</mi>
</msub>
<mo>/</mo>
<mn>2</mn>
</mrow>
</msubsup>
<msub>
<mi>p</mi>
<mi>w</mi>
</msub>
<msup>
<mrow>
<mo>(</mo>
<mi>&tau;</mi>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
<mi>d&tau;</mi>
</msqrt>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>8</mn>
<mo>)</mo>
</mrow>
<mo>.</mo>
</mrow>
2. a kind of time-domain digital weighted method of nonstationary noise signal as claimed in claim 1, it is characterised in that：The step
It is rapid 4) in, when carrying out inverse Fourier Tranform, discretization is carried out to frequency f and time t respectively, wherein, frequency f is discrete thinner, punching
Sharp receptance function is more accurate, and time t discrete steps are smaller, and the waveform for describing impulse response function is accurate.
3. a kind of time-domain digital weighted method of nonstationary noise signal as claimed in claim 2, it is characterised in that：In reality
In, f is discrete to be refined as far as possible, and time t discrete steps choose as needed, and selection principle is less than step 7) in interpolation
The time interval of resampling.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510067451.1A CN104729677B (en) | 2014-12-31 | 2015-02-09 | A kind of time-domain digital weighted method of nonstationary noise signal |
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2014108416890 | 2014-12-31 | ||
| CN201410841689 | 2014-12-31 | ||
| CN201510067451.1A CN104729677B (en) | 2014-12-31 | 2015-02-09 | A kind of time-domain digital weighted method of nonstationary noise signal |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104729677A CN104729677A (en) | 2015-06-24 |
| CN104729677B true CN104729677B (en) | 2017-10-03 |
Family
ID=53453784
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510067451.1A Active CN104729677B (en) | 2014-12-31 | 2015-02-09 | A kind of time-domain digital weighted method of nonstationary noise signal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104729677B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107238502B (en) * | 2017-05-25 | 2020-09-08 | 西南交通大学 | Method for evaluating transmission abnormal sound of vehicle hydraulic shock absorber |
| CN107727345B (en) * | 2017-09-19 | 2019-08-16 | 合肥通用机械研究院有限公司 | A kind of Non-stationary vibration signal generation method for accelerated test |
| CN107832261B (en) * | 2017-11-02 | 2020-09-29 | 华东交通大学 | Wavelet transform-based non-stationary exhaust noise signal order quantitative extraction method |
| CN110006524A (en) * | 2018-12-30 | 2019-07-12 | 瑞声科技(新加坡)有限公司 | A sound level test macro and test method applied to linear motor |
| CN113421539B (en) * | 2021-07-19 | 2023-10-10 | 北京安声浩朗科技有限公司 | Active noise reduction method and device, electronic equipment and computer readable storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20130056402A (en) * | 2011-11-22 | 2013-05-30 | 현대자동차주식회사 | Method and apparatus for measuring sandnoise for vehicle |
| CN103983345A (en) * | 2014-03-28 | 2014-08-13 | 东北石油大学 | Single-frequency tone acoustic signal automatic monitoring method based on human auditory characteristics |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH0671357B2 (en) * | 1984-03-08 | 1994-09-07 | 富士通テン株式会社 | Hearing correction device |
| JP5946039B2 (en) * | 2012-08-24 | 2016-07-05 | 株式会社大林組 | Noise monitoring system |
-
2015
- 2015-02-09 CN CN201510067451.1A patent/CN104729677B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20130056402A (en) * | 2011-11-22 | 2013-05-30 | 현대자동차주식회사 | Method and apparatus for measuring sandnoise for vehicle |
| CN103983345A (en) * | 2014-03-28 | 2014-08-13 | 东北石油大学 | Single-frequency tone acoustic signal automatic monitoring method based on human auditory characteristics |
Non-Patent Citations (7)
| Title |
|---|
| 《内燃机噪声源识别的声模态及A计权小波改进算法的研究》;杨金才;《中国优秀博士学位论文全文数据库工程科技Ⅱ辑》;20061215(第12期);C039-7 * |
| 《噪声1/3倍频程计权声级算法》;张登攀;《河南理工大学学报》;20131231;第32卷(第6期);第709-713页 * |
| 《基于Hanning窗插值FFT的频率计权频谱分析方法》;凌菁;《中国仪器仪表》;20140930(第9期);第72-76页 * |
| 《基于正交小波变换的计权声级测量》;汪光森;《仪器仪表学报》;20040430;第25卷(第2期);第198-202页 * |
| 《声级计的时间计权特性》;陈剑林;《检定与校准》;20071231(第7期);第44-48页 * |
| 《声级计的频率计权特性》;陈剑林;《检定与校正》;20081231(第6期);第47-50页 * |
| 《用A计权连续小波变换识别内燃机噪声源》;杨金才;《浙江大学学报》;20060731;第40卷(第7期);第1174-1179页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104729677A (en) | 2015-06-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104729677B (en) | A kind of time-domain digital weighted method of nonstationary noise signal | |
| CN102881289B (en) | Hearing perception characteristic-based objective voice quality evaluation method | |
| KR101110141B1 (en) | Cyclic signal processing method, cyclic signal conversion method, cyclic signal processing device, and cyclic signal analysis method | |
| US20210193149A1 (en) | Method, apparatus and device for voiceprint recognition, and medium | |
| CN104485114B (en) | A kind of method of the voice quality objective evaluation based on auditory perception property | |
| JPS63259696A (en) | Voice pre-processing method and apparatus | |
| CN101426169B (en) | Time-domain tracking filter quickly detecting acoustic response parameter of sounding body and system | |
| CN106255027B (en) | A kind of the sound quality Small Enclosure appraisal procedure and system of non-linear audio system | |
| CN105371945B (en) | Remove the method and device of noise in hall impulse response measurement acoustical signal | |
| JP5992427B2 (en) | Method and apparatus for estimating a pattern related to pitch and / or fundamental frequency in a signal | |
| CN102750955B (en) | Vocoder based on residual signal spectrum reconfiguration | |
| CN103944535B (en) | A kind of method of all phase DFT filter group utilizing Frequency Response to configure and device thereof | |
| CN104050147A (en) | Method and system for converting time domain signals into frequency domain signals | |
| CN103557925B (en) | Underwater target gammatone discrete wavelet coefficient auditory feature extraction method | |
| US20020062211A1 (en) | Easily tunable auditory-based speech signal feature extraction method and apparatus for use in automatic speech recognition | |
| CN103559893B (en) | One is target gammachirp cepstrum coefficient aural signature extracting method under water | |
| JPH03266899A (en) | Device and method for suppressing noise | |
| JP5077847B2 (en) | Reverberation time estimation apparatus and reverberation time estimation method | |
| CN115715413A (en) | Method, device and system for detecting and extracting spatial identifiable sub-band audio source | |
| EP2877820B1 (en) | Method of extracting zero crossing data from full spectrum signals | |
| CN104034970B (en) | A kind of sound level measurement method based on computer sound card | |
| CN112992182B (en) | Vehicle wind noise level testing system and testing method thereof | |
| Xu et al. | Analysis of moving source characteristics using polynomial chirplet transform | |
| Kuan et al. | A robust BFCC feature extraction for ASR system. | |
| Davis | Octave and fractional octave band digital filtering based on the proposed ANSI standard |
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 |