CN115954012B - Periodic transient interference event detection method - Google Patents
Periodic transient interference event detection method Download PDFInfo
- Publication number
- CN115954012B CN115954012B CN202310198975.9A CN202310198975A CN115954012B CN 115954012 B CN115954012 B CN 115954012B CN 202310198975 A CN202310198975 A CN 202310198975A CN 115954012 B CN115954012 B CN 115954012B
- Authority
- CN
- China
- Prior art keywords
- transient
- transient interference
- interference
- frame
- noise
- 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
- 230000001052 transient effect Effects 0.000 title claims abstract description 169
- 230000000737 periodic effect Effects 0.000 title claims abstract description 43
- 238000001514 detection method Methods 0.000 title claims description 11
- 238000001228 spectrum Methods 0.000 claims abstract description 46
- 238000000034 method Methods 0.000 claims abstract description 16
- 238000009499 grossing Methods 0.000 claims description 15
- 230000014509 gene expression Effects 0.000 claims description 9
- 101001120757 Streptococcus pyogenes serotype M49 (strain NZ131) Oleate hydratase Proteins 0.000 claims description 6
- 229940083712 aldosterone antagonist Drugs 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000005070 sampling Methods 0.000 claims description 3
- 230000002708 enhancing effect Effects 0.000 abstract description 4
- 239000007789 gas Substances 0.000 description 3
- 239000000779 smoke Substances 0.000 description 2
- 239000003086 colorant Substances 0.000 description 1
- 238000010411 cooking Methods 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 238000004134 energy conservation Methods 0.000 description 1
- 239000003517 fume Substances 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Landscapes
- Noise Elimination (AREA)
Abstract
S1, estimating a non-transient signal power spectrum and a conditional speech existence probability according to a time-frequency domain signal received by a microphone: s2, estimating a transient noise power spectrum and an optimal gain spectrum; s3, judging whether the transient interference has periodicity according to the conditional existence probability of the transient interference: s31, calculating an average value of an optimal gain function in a specified frequency band according to an optimal gain spectrum of the transient interference noise power spectrum, S32, judging whether the current frame contains transient interference according to the average value of the optimal gain function, and S33, judging whether a periodic transient interference event occurs. The invention utilizes the conditional existence probability of the transient interference obtained in the transient interference enhancing process to calculate the conditional probability of the frame transient interference, and carries out periodic transient interference event judgment on the average value of the conditional probability of the frame transient interference calculated by continuous frames, thereby improving the judgment accuracy and providing a technical basis for controlling voice equipment by utilizing the transient interference.
Description
Technical Field
The invention belongs to the technical field of digital signal processing, and particularly relates to a periodic transient interference event detection method.
Background
In the field of speech signal processing, transient noise is generally regarded as an interference signal, and transient noise suppression is widely used to improve the speech quality, i.e. to improve speech intelligibility and intelligibility. However, transient noise can also be a useful information in situations where a voice call is not required. For example: the range hood is started by utilizing the periodic transient interference formed by the ignition of the gas stove, so that people are prevented from forgetting to start the range hood, and the kitchen fume pollution is reduced. In addition, under the condition that the refrigerator door is forgotten to be closed or is not closed tightly, the refrigerator can give out prompt tones, and the refrigerator door can be automatically closed by utilizing the periodic transient sound, so that energy conservation and emission reduction are realized. Therefore, a method for detecting periodic transient interference events is needed to determine the operating state of the device.
Disclosure of Invention
The invention discloses a method for detecting a periodic transient interference event in order to detect the working state of voice equipment by utilizing transient noise.
The invention discloses a method for detecting a periodic transient interference event, which is characterized by comprising the following steps:
s1, setting a microphone receiving signal as
y(n)=x(n)+t(n)
Where n is the time scale, x (n) is the speech signal, and t (n) is the transient noise.
Time-frequency domain signals received by microphone through short-time Fourier transform
Y (k, l) =x (k, l) +t (k, l), where k is the frequency scale and l is the time domain frame index;
estimating a non-transient signal power spectrum lambda d (k, l) and conditional speech existence probability p (k, l);
s2, estimating transient noise power spectrum lambda t (k, l) and an optimal gain spectrum G (k, l)
The transient noise power spectrum is expressed as
λ t (k,l)=|G(k,l)*Y(k,l)| 2
Where G (k, l) represents the optimal gain spectrum for estimating the transient interference noise power spectrum, and its calculation formula is
G in min Is the preset spectrum gain without transient noise, G H1 Is gain in the presence of transient noise expressed as
Where e is a natural constant, ζ (k, l) is a priori signal-to-noise ratio of transient noise, αt represents a weight factor, αt e [0,1], γ (k, l) is a posterior signal-to-noise ratio, v (k, l) represents a relationship parameter between the a priori signal-to-noise ratio and the posterior signal-to-noise ratio, and the calculation formulas are respectively:
wherein Y (k, l) is the time-frequency domain signal received by the microphone in the S1 step, lambda d (k, l) is a non-transient signal power spectrum;
s3, judging whether the transient interference has periodicity according to the conditional existence probability of the transient interference:
s31, firstly, calculating the average value of the optimal gain function in the appointed frequency band according to the optimal gain frequency spectrum G (k, l) of the transient interference noise power spectrum, namely
Middle bin start To specify the band start of a band, bin end Is the end of the band.
Calculating the average value of the optimal gain function of continuous m frames taking the current frame as the end point
m is the set average optimal frame number;
s32, if
Greater than a set transient interference probability threshold G Th Judging the current frame as transient interference, otherwise, considering that the current frame does not contain transient interference noise;
when the current frame is determined to contain transient interference noise, carrying out periodic transient interference event detection, entering a step S33, otherwise repeating the step S32;
s33, judging whether a periodic transient interference event occurs or not;
s331, constructing an event judgment circulating window for storing a difference value T between the current transient interference frame number and the last transient interference frame number d The expression is
T d = T current - T last
T in current For the current transient interference frame number, T last The number of frames is the number of frames of the last transient interference;
constructing event judgment circulation window to circularly store all T between current transient interference frame number and last transient interference frame number d The frame, the window length of the circulation window is L, for each current frame in the circulation window, calculate the difference of the current first frame;
T d (l)= T current - T last
l=1,2…L
s332, setting the period time of the periodic transient interference event as T 0 Further converting the cycle time into a cycle frame number T
T=T 0 /（len/F s ）
Where len is the length of each frame of signal and Fs is the signal sampling frequency;
setting a minimum threshold T for judging periodic transient interference event min And a maximum threshold value T max And satisfy T min <T<T max ；T min ，T max Respectively representing a threshold minimum value and a threshold maximum value, wherein the unit is the number of frames;
s333, judging the current difference value stored in the circulation window, if the difference value is more than or equal to the minimum threshold value and less than or equal to the maximum threshold value, performing cycle counting,
if T min ≤T d (l) ≤T max Count value Count is incremented by 1;
if the period count of the L frames in the circulation window is greater than the specified period transient interference event threshold T judge And judging that a periodic transient interference event occurs.
Preferably, the step S331 further includes an inter-adjacent interference determination, where an interference threshold δt is set, when Td is greater than δt, it is considered that there is no inter-adjacent interference, and if it is considered that there is inter-adjacent interference, the subsequent step is not performed and the step S331 is repeated.
Preferably, the minimum threshold tmin=t-TR and the maximum threshold tmax=t+tr of the periodic transient interference event are determined, and TR is the set tolerant frame number.
Preferably, the method for detecting a periodic transient interference event according to claim 1, wherein in the step S1, the non-transient signal power spectrum is estimated by using an MCRA algorithm
Alpha in the formula d Is the first smoothing parameter, and alpha d ∈[0,1]，
Representing a time-varying smoothing parameter, p (k, l) is the conditional speech presence probability, expressed as
p(k,l)=α p *p(k,l-1)+( 1-α p )*I(k,l)
Alpha in the formula p Is the second smoothing parameter, and alpha p ∈[0,1]I (k, l) is an identification function, when I (k, l) =1, the kth bin of the first frame is marked as containing speech signals, otherwise not, wherein δ is an empirical threshold, S r (k, l) is the power spectrum ratio, and the expression is
S r （k,l）= S（k,l）/ S min (k, l) where S (k, l) is the time-domain smoothed power of the noisy speech power spectrum, S min (k, l) is the minimum power value searched, and the expressions are in turn
S（k,l）=α S *S（k,l-1）+(1-α S ) *| Y（k,l）| 2
S min （k,l）=min{ S min （k,l-1）, S（k,l）}
Alpha in the formula S Is a third smoothing parameter, and alpha S ∈[0,1]。
The invention utilizes the conditional existence probability of the transient interference obtained in the transient interference enhancing process to calculate the conditional probability of the frame transient interference, and carries out periodic transient interference event judgment on the average value of the conditional probability of the frame transient interference calculated by continuous frames, thereby improving the judgment accuracy and providing a technical basis for controlling voice equipment by utilizing the transient interference.
Drawings
FIG. 1 is a schematic flow chart of a method for detecting a periodic transient interference event according to an embodiment of the present invention;
FIG. 2 is a flowchart illustrating a periodic transient disturbance event determination according to the present invention;
fig. 3 is a graph showing a comparison of a microphone received signal and an enhanced transient noise spectrum in accordance with an embodiment of the present invention.
Description of the embodiments
The following describes the present invention in further detail.
The following detailed description of the invention is, therefore, not to be taken in a limiting sense, and is set forth in the appended drawings.
A flow of one embodiment of the present invention is shown in fig. 1.
S1, estimating a non-transient signal power spectrum lambda according to an MCRA algorithm d (k, l) conditional speech existence probability p (k, l)
Setting the microphone to receive signals
y(n)=x(n)+t(n)
Where n is the time scale, x (n) is the speech signal, and t (n) is the transient noise.
Time-frequency domain signals received by microphone through short-time Fourier transform
Y(k,l)=X(k,l)+T(k,l)
Where k is the frequency scale and l is the time domain frame index.
The non-transient signal power spectrum estimated by using MCRA algorithm is
Alpha in the formula d Is the first smoothing parameter, and alpha d ∈[0,1]，
Representing a time-varying smoothing parameter, p (k, l) is the conditional speech presence probability, expressed as
p(k,l)=α p *p(k,l-1)+( 1-α p )*I(k,l)
Alpha in the formula p Is the second smoothing parameter, and alpha p ∈[0,1]I (k, l) is an identification function, when I (k, l) =1, the kth bin of the first frame is marked as containing speech signals, otherwise not, wherein δ is an empirical threshold, S r (k, l) is the power spectrum ratio, and the expression is
S r （k,l）= S（k,l）/ S min (k, l) where S (k, l) is the time-domain smoothed power of the noisy speech power spectrum, S min (k, l) is the minimum power value searched, and the expressions are in turn
S（k,l）=α S *S（k,l-1）+(1-α S ) *| Y（k,l）| 2
S min （k,l）=min{ S min （k,l-1）, S（k,l）}
Where αS is a third smoothing parameter and αS ε [0,1], the smaller αS the faster the PSD of the captured speech or background noise changes.
In the step, according to the characteristics that the transient interference of the voice and the background noise have different change rates, compared with the slower voice and the pseudo-static background noise, the transient interference is changed rapidly along with time, and the like, the transient interference power spectrum is estimated and obtained by adopting an MCRA algorithm, the optimization can be carried out by adjusting the gentle factor of the MCRA algorithm, and the power spectrum density change of the voice or the background noise is captured more rapidly.
S2, as shown in FIG. 2, estimating transient noise power spectrum lambda based on OM-LSA algorithm t (k, l) and an optimal gain spectrum G (k, l)
The transient noise power spectrum is expressed as
λ t (k,l)=|G(k,l)*Y(k,l)| 2
Where G (k, l) represents the optimal gain spectrum for estimating the transient interference noise power spectrum, and its calculation formula is
G in min Is the spectral gain in the absence of transient noise, G min G is a preset value H1 Is gain in the presence of transient noise expressed as
Where e is a natural constant, ζ (k, l) is a priori signal-to-noise ratio of transient noise, αt represents a weight factor, αt e [0,1], γ (k, l) is a posterior signal-to-noise ratio, v (k, l) represents a relationship parameter between the a priori signal-to-noise ratio and the posterior signal-to-noise ratio, and the calculation formulas are respectively:
wherein Y (k, l) is the time-frequency domain signal received by the microphone in the S1 step, lambda d (k, l) is the non-transient signal power spectrum.
The method can enhance transient interference and inhibit non-transient interference voice signals and background noise by using an OM-LSA algorithm.
S3, judging whether the transient interference has periodicity according to the conditional existence probability of the transient interference:
one specific flow chart for periodic transient disturbance event determination is shown in FIG. 2.
S31, firstly, calculating the average value of the optimal gain function in the appointed frequency band according to the optimal gain spectrum of the transient interference noise power spectrum, namely
Middle bin start To specify the band start of a band, bin end The specified frequency band is the frequency band in which it is necessary to judge whether the transient interference event occurs or not in the present invention.
An average value of the optimal gain function of m consecutive frames ending with the current frame is calculated, m being the set average optimal frame number, for example, m=3,
s32, if
Greater than a set transient interference probability threshold G Th Judging the current frame as transient interference, otherwise, considering that the current frame does not contain transient interference noise;
when the current frame is determined to contain transient interference noise, carrying out periodic transient interference event detection, entering a step S33, otherwise repeating the step S32;
s33, when the current frame is determined to contain transient interference noise, periodic transient interference event detection is carried out, and the specific steps are as follows:
s331, constructing an event judgment circulating window for storing a difference value T between the current transient interference frame number and the last transient interference frame number d The expression is
T d = T current - T last
T in current For the current transient interference frame number, T last The number of frames is the number of frames of the last transient interference;
event judgment circulating window circulating storage timeAll T between the previous transient interference frame number and the last transient interference frame number d For each current frame in the circulation window, calculating the difference value of the current first frame
T d (l)= T current - T last
l=1,2…L
To avoid inter-adjacent interference, an interference threshold delta may be set T When T d Greater than delta T If no adjacent interference exists, proceeding with the subsequent steps, otherwise, considering that adjacent interference exists, T d (l) =0, do not proceed with the subsequent steps and proceed with this step S331;
s332, setting the period time of the periodic transient interference event as T 0 Further converting the cycle time into a cycle frame number T
T=T 0 /（len/F s ）
Where len is the length of the signal per frame and Fs is the signal sampling frequency.
Setting a minimum threshold T for judging the periodic transient interference event according to the value of the period frame number T min And a maximum threshold value T max The specific setting mode is that the period frame number T is respectively increased and decreased by a tolerance frame number TR to obtain a minimum threshold value T min And a maximum threshold value T max ；
Setting a minimum threshold value T according to T min And a maximum threshold value T max The purpose is to expand the value of the period frame number T into a range of values, thereby improving the accuracy of the period counting in step S333, and thus T min ，T max The value of (2) cannot be too large, and is set to an empirical value in the embodiment, for example, T can be taken min =T-10，T max =t+10, i.e. the tolerant frame number TR is 10;
s333, judging the current difference value stored in the circulation window, if the difference value is more than or equal to the minimum threshold value and less than or equal to the maximum threshold value, performing cycle counting,
if T min ≤T d (l) ≤T max
Count value Count incremented by 1
If the period count of the L frames in the circulation window is greater than the specified period transient interference event threshold T judge For example, count is greater than 10, a periodic transient disturbance event is determined to occur.
The invention relates to a method for detecting transient interference periodicity, which comprises the steps of calculating the conditional probability of frame transient interference by utilizing the conditional existence probability of transient interference obtained in the transient interference enhancing process, carrying out transient interference judgment on the average value of the conditional probability of frame transient interference calculated by continuous multiframes, and determining that the current frame is transient interference when the average value is larger than a given threshold value. And if the current frame is transient interference, detecting a periodic transient interference event.
And finally, judging the occurrence of the periodic transient interference event when the cycle count value is greater than the set threshold value, thereby judging the working state of the equipment.
Examples:
kitchen oil smoke pollution can seriously harm human health, and the kitchen oil smoke pollution is effectively reduced by using the range hood during cooking, but people may forget to turn on the range hood.
The periodic transient interference event detection method provided by the embodiment is applied to the intelligent range hood, and the range hood is awakened to be started by detecting the periodic transient interference formed by ignition of the gas stove, so that the situation that people forget to start the range hood is avoided.
According to the periodic transient interference event detection method, transient interference is enhanced by using an OM-LSA algorithm, and a voice signal and background noise of non-transient interference are restrained. The parameters of this embodiment are set as follows: first smoothing parameter alpha d =0.7, second smoothing parameter α p =0.3, third smoothing parameter α s =0.85, weight factor α t =0.8, empirical threshold δ=1.8, g min =0.02。
As shown in fig. 3, the abscissa is time, the ordinate is frequency domain, the upper part of fig. 3 is a spectrogram of a microphone receiving signal, which contains a transient interference signal and other background noise signals, and the lower part of the figure is an enhanced transient noise spectrogram.
As shown in the spectrogram below in fig. 3, the portion of the spectrogram with a color close to light color indicates no energy, i.e., no signal, and the portion with a color close to red indicates the presence of a voice signal, and the darker the color, the greater the signal energy. In the period of 13 s-15.5 s, a plurality of vertical lines with colors darker than the front and rear areas exist, each vertical line indicates that the energy is suddenly increased at the moment of the abscissa, and the characteristics of transient signals are met, so that transient interference is considered to exist.
In order to detect periodicity of transient interference, the embodiment obtains conditional probability of frame transient interference by using conditional existence probability of transient interference obtained in the transient interference enhancing process, and determines that the current frame is transient interference when the average value of the conditional probability of frame transient interference obtained by continuous 3 frames is larger than a given threshold value. If the current frame is transient interference, carrying out periodic transient interference event detection, firstly constructing an event judgment circulating window for storing the difference value between the current transient interference frame number and the last transient interference frame number, then determining the period of the event through priori information, setting the minimum value and the maximum value of a judgment threshold according to the period, finally judging the circulating window, if the difference value of the circulating window is within the range of the judgment threshold, carrying out period counting, and finally judging that the periodic transient interference event occurs when the period counting value is larger than the set threshold, thereby judging the working state of equipment. The parameters of this embodiment are set as follows:
designated band starting bin start =15, designated bin end point end =25, transient interference probability threshold G Th =0.3, interference threshold δ t =5，
The cyclic window length l=20, t 0 =100,len=32,F S =16000,T judge =10。
The current detection of the transient interference signals is indicated by serial port printing, and if a plurality of transient interference signals are continuously detected, the existence of ignition sound of the gas stove can be judged, and the range hood is automatically started by sending a wake-up word.
The foregoing description of the preferred embodiments of the present invention is not obvious contradiction or on the premise of a certain preferred embodiment, but all the preferred embodiments can be used in any overlapped combination, and the embodiments and specific parameters in the embodiments are only for clearly describing the invention verification process of the inventor and are not intended to limit the scope of the invention, and the scope of the invention is still subject to the claims, and all equivalent structural changes made by applying the specification and the content of the drawings of the present invention are included in the scope of the invention.
Claims (4)
1. A method for detecting a periodic transient interference event, comprising the steps of:
s1, setting a microphone receiving signal as
y(n)=x(n)+t(n)
Wherein n is a time scale, x (n) is a voice signal, and t (n) is transient noise;
time-frequency domain signals received by microphone through short-time Fourier transform
Y (k, l) =x (k, l) +t (k, l), where k is the frequency scale and l is the time domain frame index;
estimating a non-transient signal power spectrum lambda d (k, l) and conditional speech existence probability p (k, l);
s2, estimating transient noise power spectrum lambda t (k, l) and an optimal gain spectrum G (k, l)
The transient noise power spectrum is expressed as
λ t (k,l)=|G(k,l)*Y(k,l)| 2
Where G (k, l) represents the optimal gain spectrum for estimating the transient interference noise power spectrum, and its calculation formula is
G in min Is the preset spectrum gain without transient noise, G H1 Is gain in the presence of transient noise expressed as
Where e is a natural constant, ζ (k, l) is a priori signal-to-noise ratio of transient noise, αt represents a weight factor, αt e [0,1], γ (k, l) is a posterior signal-to-noise ratio, v (k, l) represents a relationship parameter between the a priori signal-to-noise ratio and the posterior signal-to-noise ratio, and the calculation formulas are respectively:
wherein Y (k, l) is the time-frequency domain signal received by the microphone in the S1 step, lambda d (k, l) is a non-transient signal power spectrum;
s3, judging whether the transient interference has periodicity according to the conditional existence probability of the transient interference:
s31, firstly, calculating the average value of the optimal gain function in the appointed frequency band according to the optimal gain frequency spectrum G (k, l) of the transient interference noise power spectrum, namely
Middle bin start To specify the band start of a band, bin end Is the end of the frequency band;
calculating the average value of the optimal gain function of continuous m frames taking the current frame as the end point
m is the set average optimal frame number;
s32, if
Greater than a set transient interference probability threshold G Th Judging the current frame as transient interference, otherwise, considering that the current frame does not contain transient interference noise;
when the current frame is determined to contain transient interference noise, carrying out periodic transient interference event detection, entering a step S33, otherwise repeating the step S32;
s33, judging whether a periodic transient interference event occurs or not;
s331, constructing an event judgment circulating window for storing a difference value T between the current transient interference frame number and the last transient interference frame number d The expression is
T d = T current - T last
T in current For the current transient interference frame number, T last The number of frames is the number of frames of the last transient interference;
constructing event judgment circulation window to circularly store all T between current transient interference frame number and last transient interference frame number d The frame, the window length of the circulation window is L, for each current frame in the circulation window, calculate the difference of the current first frame;
T d (l)= T current - T last
l=1,2…L
s332, setting the period time of the periodic transient interference event as T 0 Further converting the cycle time into a cycle frame number T
T=T 0 /（len/F s ）
Where len is the length of each frame of signal and Fs is the signal sampling frequency;
setting a minimum threshold T for judging periodic transient interference event min And a maximum threshold value T max And satisfy T min <T<T max ；T min ，T max Respectively representing a threshold minimum value and a threshold maximum value, wherein the unit is the number of frames;
s333, judging the current difference value stored in the circulation window, if the difference value is more than or equal to the minimum threshold value and less than or equal to the maximum threshold value, performing cycle counting,
if T min ≤T d (l) ≤T max Count value Count is incremented by 1;
if the period count of the L frames in the circulation window is greater than the specified period transient interference event threshold T judge And judging that a periodic transient interference event occurs.
2. The method for detecting a periodic transient interference event according to claim 1, wherein said step S331 further comprises determining inter-adjacent interference, and setting an interference threshold δ T When T d Greater than delta T When it is considered that there is no inter-adjacent interference, if it is considered that there is inter-adjacent interference, the subsequent step is not performed and step S331 is repeated.
3. The method for detecting a periodic transient interference event according to claim 1, wherein a minimum threshold T of the periodic transient interference event is determined min =t-TR and maximum threshold T max =t+tr, TR is the set number of tolerant frames.
4. The method for detecting periodic transient interference events according to claim 1, wherein in step S1, the non-transient signal power spectrum is estimated by using MCRA algorithm
Alpha in the formula d Is the first smoothing parameter, and alpha d ∈[0,1]，
Representing a time-varying smoothing parameter, p (k, l) is the conditional speech presence probability, expressed as
p(k,l)=α p *p(k,l-1)+( 1-α p )*I(k,l)
Alpha in the formula p Is the second smoothing parameter, and alpha p ∈[0,1]I (k, l) is an identification function, when I (k, l) =1, the kth bin of the first frame is marked as containing speech signals, otherwise not, wherein δ is an empirical threshold, S r (k, l) is the power spectrum ratio, and the expression is
S r （k,l）= S（k,l）/ S min (k, l) where S (k, l) is the time-domain smoothed power of the noisy speech power spectrum, S min (k, l) is the minimum power value searched, and the expressions are in turn
S（k,l）=α S *S（k,l-1）+(1-α S ) *| Y（k,l）| 2
S min （k,l）=min{ S min （k,l-1）, S（k,l）}
Alpha in the formula S Is a third smoothing parameter, and alpha S ∈[0,1]。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310198975.9A CN115954012B (en) | 2023-03-03 | 2023-03-03 | Periodic transient interference event detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310198975.9A CN115954012B (en) | 2023-03-03 | 2023-03-03 | Periodic transient interference event detection method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115954012A CN115954012A (en) | 2023-04-11 |
| CN115954012B true CN115954012B (en) | 2023-05-09 |
Family
ID=85892921
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310198975.9A Active CN115954012B (en) | 2023-03-03 | 2023-03-03 | Periodic transient interference event detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115954012B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116312545B (en) * | 2023-05-26 | 2023-07-21 | 北京道大丰长科技有限公司 | Speech recognition system and method in a multi-noise environment |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102084416A (en) * | 2008-02-21 | 2011-06-01 | 史诺有限公司 | Audio visual signature, method of deriving a signature, and method of comparing audio-visual data |
| CN103310800A (en) * | 2012-03-06 | 2013-09-18 | 中国科学院声学研究所 | Voiced speech detection method and voiced speech detection system for preventing noise interference |
| CN103456310A (en) * | 2013-08-28 | 2013-12-18 | 大连理工大学 | Transient noise suppression method based on spectrum estimation |
| CN104157295A (en) * | 2014-08-22 | 2014-11-19 | 中国科学院上海高等研究院 | Method used for detecting and suppressing transient noise |
| CN106170929A (en) * | 2014-02-10 | 2016-11-30 | 奥迪马科斯公司 | There is the communication system of the noise immunity of improvement, method and apparatus |
| CN107507630A (en) * | 2017-07-17 | 2017-12-22 | 嘉兴开泽电子设备有限公司 | A kind of non-cooperation voice communication receives data dead time section recognition methods |
| CN108922554A (en) * | 2018-06-04 | 2018-11-30 | 南京信息工程大学 | The constant Wave beam forming voice enhancement algorithm of LCMV frequency based on logarithm Power estimation |
| CN110739005A (en) * | 2019-10-28 | 2020-01-31 | 南京工程学院 | real-time voice enhancement method for transient noise suppression |
| CN114527321A (en) * | 2022-02-24 | 2022-05-24 | 国网北京市电力公司 | Anti-transient interference voltage sag detection method and device suitable for SSTS |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106997768B (en) * | 2016-01-25 | 2019-12-10 | 电信科学技术研究院 | Method and device for calculating voice occurrence probability and electronic equipment |
| US10891936B2 (en) * | 2019-06-05 | 2021-01-12 | Harman International Industries, Incorporated | Voice echo suppression in engine order cancellation systems |
| US12236958B2 (en) * | 2021-05-04 | 2025-02-25 | Cypress Semiconductor Corporation | Retransmission softbit decoding |
-
2023
- 2023-03-03 CN CN202310198975.9A patent/CN115954012B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102084416A (en) * | 2008-02-21 | 2011-06-01 | 史诺有限公司 | Audio visual signature, method of deriving a signature, and method of comparing audio-visual data |
| CN103310800A (en) * | 2012-03-06 | 2013-09-18 | 中国科学院声学研究所 | Voiced speech detection method and voiced speech detection system for preventing noise interference |
| CN103456310A (en) * | 2013-08-28 | 2013-12-18 | 大连理工大学 | Transient noise suppression method based on spectrum estimation |
| CN106170929A (en) * | 2014-02-10 | 2016-11-30 | 奥迪马科斯公司 | There is the communication system of the noise immunity of improvement, method and apparatus |
| CN104157295A (en) * | 2014-08-22 | 2014-11-19 | 中国科学院上海高等研究院 | Method used for detecting and suppressing transient noise |
| CN107507630A (en) * | 2017-07-17 | 2017-12-22 | 嘉兴开泽电子设备有限公司 | A kind of non-cooperation voice communication receives data dead time section recognition methods |
| CN108922554A (en) * | 2018-06-04 | 2018-11-30 | 南京信息工程大学 | The constant Wave beam forming voice enhancement algorithm of LCMV frequency based on logarithm Power estimation |
| CN110739005A (en) * | 2019-10-28 | 2020-01-31 | 南京工程学院 | real-time voice enhancement method for transient noise suppression |
| CN114527321A (en) * | 2022-02-24 | 2022-05-24 | 国网北京市电力公司 | Anti-transient interference voltage sag detection method and device suitable for SSTS |
Non-Patent Citations (1)
| Title |
|---|
| "基于矩阵补全的天波雷达瞬态干扰抑制算法";李茂;《电子与信息学报》;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115954012A (en) | 2023-04-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108831499B (en) | Speech enhancement method using speech existence probability | |
| CN110739005B (en) | Real-time voice enhancement method for transient noise suppression | |
| CN101354889B (en) | Method and apparatus for tonal modification of voice | |
| JP4279357B2 (en) | Apparatus and method for reducing noise, particularly in hearing aids | |
| CN103238183B (en) | Noise suppression device | |
| CN103632666B (en) | Audio recognition method, speech recognition apparatus and electronic equipment | |
| CA2404030A1 (en) | Relative noise ratio weighting techniques for adaptive noise cancellation | |
| CA2404027A1 (en) | Communication system noise cancellation power signal calculation techniques | |
| CN115954012B (en) | Periodic transient interference event detection method | |
| CN101320559A (en) | Sound activation detection apparatus and method | |
| JP2004254322A5 (en) | ||
| CN109360585A (en) | A kind of voice-activation detecting method | |
| CA2458428A1 (en) | System for suppressing wind noise | |
| CN103456310A (en) | Transient noise suppression method based on spectrum estimation | |
| KR20090012154A (en) | Noise reduction method with integrated pure tone reduction | |
| CN104599677B (en) | Transient noise suppressing method based on speech reconstructing | |
| CN111292758B (en) | Voice activity detection method and device and readable storage medium | |
| WO2001073751A1 (en) | Speech presence measurement detection techniques | |
| US9002030B2 (en) | System and method for performing voice activity detection | |
| WO2017128910A1 (en) | Method, apparatus and electronic device for determining speech presence probability | |
| CN112102818B (en) | Signal-to-noise ratio calculation method combining voice activity detection and sliding window noise estimation | |
| JPH05265479A (en) | Voice signal processor | |
| US20030078770A1 (en) | Method for detecting a voice activity decision (voice activity detector) | |
| CN110689905A (en) | Voice activity detection system for video conference system | |
| CN103310800A (en) | Voiced speech detection method and voiced speech detection system for preventing noise interference |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |