CN116962934B - Pickup noise reduction method and system - Google Patents
Pickup noise reduction method and system Download PDFInfo
- Publication number
- CN116962934B CN116962934B CN202311208195.4A CN202311208195A CN116962934B CN 116962934 B CN116962934 B CN 116962934B CN 202311208195 A CN202311208195 A CN 202311208195A CN 116962934 B CN116962934 B CN 116962934B
- Authority
- CN
- China
- Prior art keywords
- noise reduction
- sound
- signal
- noise
- uplink
- 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
- 238000000034 method Methods 0.000 title claims abstract description 33
- 238000012545 processing Methods 0.000 claims abstract description 75
- 230000008030 elimination Effects 0.000 claims abstract description 28
- 238000003379 elimination reaction Methods 0.000 claims abstract description 28
- 238000001228 spectrum Methods 0.000 claims description 65
- 238000001914 filtration Methods 0.000 claims description 39
- 238000011282 treatment Methods 0.000 claims description 21
- 230000007613 environmental effect Effects 0.000 claims description 19
- 238000005070 sampling Methods 0.000 claims description 14
- 238000001514 detection method Methods 0.000 claims description 10
- 230000005236 sound signal Effects 0.000 claims description 10
- 241000209140 Triticum Species 0.000 claims description 7
- 235000021307 Triticum Nutrition 0.000 claims description 7
- 238000011946 reduction process Methods 0.000 claims description 7
- 230000003111 delayed effect Effects 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 7
- 238000004891 communication Methods 0.000 abstract description 6
- 230000005540 biological transmission Effects 0.000 abstract description 2
- 230000006870 function Effects 0.000 description 73
- 230000009466 transformation Effects 0.000 description 7
- 230000003595 spectral effect Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 230000003247 decreasing effect Effects 0.000 description 2
- 238000009432 framing Methods 0.000 description 2
- 230000000737 periodic effect Effects 0.000 description 2
- 239000000654 additive Substances 0.000 description 1
- 230000000996 additive effect Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000001934 delay Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 210000005069 ears Anatomy 0.000 description 1
- 210000000214 mouth Anatomy 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 210000001260 vocal cord Anatomy 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
- H04R1/00—Details of transducers, loudspeakers or microphones
- H04R1/10—Earpieces; Attachments therefor ; Earphones; Monophonic headphones
- H04R1/1083—Reduction of ambient noise
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
- H04R2201/00—Details of transducers, loudspeakers or microphones covered by H04R1/00 but not provided for in any of its subgroups
- H04R2201/10—Details of earpieces, attachments therefor, earphones or monophonic headphones covered by H04R1/10 but not provided for in any of its subgroups
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Acoustics & Sound (AREA)
- Signal Processing (AREA)
- Cable Transmission Systems, Equalization Of Radio And Reduction Of Echo (AREA)
Abstract
The invention belongs to the technical field of digital information transmission, and particularly relates to a pickup noise reduction method and a system, wherein the method comprises signal noise reduction processing, echo cancellation processing and gain adjustment processing, wherein the echo cancellation processing comprises the following steps: based on the obtained downlink sound, comparing the downlink echo, and eliminating the downlink echo in the downlink sound through echo elimination processing and signal noise reduction processing. The invention solves the problems of poor uplink noise reduction performance and unstable volume in the uplink noise reduction earphone communication in the prior art, and has the technical effects of good pickup effect, good noise reduction effect and stable uplink volume.
Description
Technical Field
The invention belongs to the technical field of digital information transmission, and particularly relates to a pickup noise reduction method and system.
Background
In recent years, noise reduction headphones have become a common electronic product in people's work and life. One type of noise reducing earphone focuses on isolating the wearer's ears from external noise, providing a better hearing experience for the wearer, such as headphones using active noise reduction technology (ANC) and passive noise reduction technology (PNC). The other type of noise reduction earphone is focused on noise reduction during pickup, so that a wearer can be helped to more clearly transmit own sound to the opposite party in the scenes of voice communication, online conferences, electronic contests and the like, and the conversation quality is improved. Because the noise reduction earphone needs to process voice data obtained by pickup in real time, especially in recent years, with the development of artificial intelligence technology, deep learning is also used in the field of pickup noise reduction, and the algorithm has a higher demand on calculation force.
The prior art has the problems of poor uplink noise reduction performance and unstable volume when the uplink noise reduction earphone is used for communication.
The invention provides a relatively simple pick-up noise reduction method. When the noise reduction and pickup platform works, the conversation microphone and the noise reduction microphone pick up sound simultaneously, sounds outside a pickup space of the noise reduction and pickup platform can be restrained, and sounds inside the pickup space of the noise reduction and pickup platform can be reserved, including human voice of a wearer speaking and background noise and other human voice noise existing in the pickup space. Wherein, the talking microphone picks up more human voice components of the wearer, and the noise reduction microphone picks up more background noise and other human voice noise components.
Disclosure of Invention
The invention provides a pickup noise reduction method and a pickup noise reduction system, which are used for solving the problems of poor uplink noise reduction performance and unstable volume in the prior art when an uplink noise reduction earphone is used for talking.
The technical problems solved by the invention are realized by adopting the following technical scheme: a pickup noise reduction method comprising:
and (3) signal noise reduction: based on the human noise with the main sound component picked up in the pick-up space, comparing the picked-up background noise with the environment noise with the main sound component of other human noise, separating the main sound component from the background noise through the environment noise reduction treatment, and forming a noise-reduced human sound signal;
echo cancellation processing: based on the obtained downlink sound, comparing the downlink echo, and eliminating the downlink echo in the downlink sound through echo elimination processing and signal noise reduction processing;
gain adjustment processing: and based on the uplink sound sent by the noise-reduced voice signal, judging through the uplink sound volume, if the sound is smaller than the target amplitude, sending out the uplink sound after the uplink sound volume is increased through uplink gain adjustment, and if the sound is larger than the target amplitude, sending out the uplink sound after the uplink sound volume is reduced through uplink gain adjustment.
Further, the ambient noise reduction process includes:
outputting a high-frequency information emphasis signal through a pre-emphasis processing function based on the voice noise;
outputting a windowed voice signal through a time domain windowing function based on the high-frequency information emphasis signal;
outputting a frequency domain energy spectrum signal through a discrete Fourier transform function based on the windowed human voice signal;
outputting a filtered human voice signal through a first weighted moving average filter function based on the frequency domain energy spectrum signal;
based on the filtered human voice signals and the environmental noise, outputting a noise reduction human voice frequency spectrum through a background subtraction elimination function;
and outputting the noise-reduced human voice signal through an inverse Fourier transform function and amplitude limiting filtering based on the noise-reduced human voice frequency spectrum or the filtering noise-reduced human voice frequency spectrum.
Further, the pre-emphasis processing function is:
；
the saidAn original signal that is picked up human voice noise;
the saidIs the data after pre-emphasis treatment;
the independent variable is the number of sampling points;
the time domain windowing function is:
；
the saidIs the windowed data;
the saidIs a hamming window function;
the discrete fourier transform function is:
；
the saidIs obtained after discrete Fourier transformIs a frequency spectrum of (2);
the saidPoints that are fourier transforms;
the saidIs normalized frequency;
the first weighted moving average filter function is:
；
the saidFiltering the human voice signal spectrum after energy spectrum filtering;
the saidFiltering the energy spectrum of the filtered human voice signal after the first energy spectrum;
the background subtraction elimination function is:
；
the saidFiltering the energy spectrum of the filtered human voice signal;
the saidAn energy spectrum that is ambient noise;
the saidAn energy spectrum obtained after spectrum subtraction;
the inverse fourier transform function is:
；
the saidA time domain signal obtained after inverse Fourier transform;
the saidA frequency spectrum obtained based on the first spectrum subtraction is subjected to filtering;
the saidPoints are inverse fourier transforms.
Further, the ambient noise reduction process further includes: and outputting a filtered noise reduction human voice spectrum through a second weighted moving average filter function based on the noise reduction human voice spectrum.
Further, the second order weighted moving average filter function is:
。
further, the echo cancellation process includes: based on the delayed downlink sound, the echo-eliminated downlink sound is output through a weighted mixing function according to the environmental noise.
Further, the weighted mixing function is:
；
the saidNoise reduction wheat signals;
the saidIs a stope signal;
the saidThe data of the stope signal and the noise reduction wheat signal after weighted mixing processing;
the saidIs the sampling rate;
is the independent variable of the number of sampling points.
Further, the uplink gain adjustment includes:
based on the uplink sound volume, and based on a detection threshold, a target value, a default gain and a stepping gain, acquiring the average amplitude of the previous frame;
if the average amplitude of the previous frame is lower than the detection threshold, outputting the current uplink sound of the self-increasing gain through an automatic gain processing function and limiting filtering according to the stepping gain;
if the average amplitude of the previous frame is higher than the target value, the current uplink sound with the self-subtracting gain is output through an automatic gain processing function and limiting filtering according to the stepping gain.
Further, the automatic gain processing function is:
；
；
the saidIs the signal after gain;
the saidThe default gain is 1.0;
the saidIs the current gain;
the saidFor the step of gain adjustment, the step of gain adjustment is +.>Said->At a sampling rate of 48 kHz;
the saidAssigning an average value to the audio signal of the previous frame at the current time point;
the saidA detection threshold is effectively input;
the saidIs the target amplitude.
A pickup noise reduction system, comprising: a pickup noise reduction platform for implementing any one of the pickup noise reduction methods described above;
the pickup noise reduction platform comprises a signal noise reduction processing module;
the signal noise reduction processing module is used for: based on the human noise with the main sound component picked up in the pick-up space, comparing the picked-up background noise with the environment noise with the main sound component of other human noise, separating the main sound component from the background noise through the environment noise reduction treatment, and forming a noise-reduced human sound signal;
based on the obtained downlink sound, comparing the downlink echo, and eliminating the downlink echo in the downlink sound through echo elimination processing and signal noise reduction processing;
and based on the uplink sound sent by the noise-reduced voice signal, judging through the uplink sound volume, if the sound is smaller than the target amplitude, sending out the uplink sound after the uplink sound volume is increased through uplink gain adjustment, and if the sound is larger than the target amplitude, sending out the uplink sound after the uplink sound volume is reduced through uplink gain adjustment.
The beneficial effects are that:
the invention adopts environmental noise reduction (ENC), echo elimination (AEC), automatic Gain Control (AGC) and other processes, wherein the environmental noise reduction respectively adopts environmental noise reduction, pre-emphasis processing function, time domain windowing function, discrete Fourier transformation function, first weighted moving average filtering function, background subtraction elimination function, inverse Fourier transformation function and processes such as amplitude limiting filtering, the echo elimination respectively adopts echo elimination, signal noise reduction and weighting mixing function, and the automatic gain control respectively adopts uplink gain adjustment and automatic gain processing function, thus completing the environmental noise reduction, echo elimination and automatic gain control.
Drawings
FIG. 1 is a general flow chart of the pickup noise reduction method of the present invention;
FIG. 2 is a flow chart of a pickup noise reduction method of the present invention;
fig. 3 is a flowchart of the sound pickup noise reduction method of the present invention.
Detailed Description
The invention is further described below with reference to the accompanying drawings.
A sound pickup noise reduction method, as shown in fig. 1 and 2, includes:
step S101, signal noise reduction processing: step S2001 is based on the human noise with the main component of the picked-up host sound in the pickup space, comparing the picked-up background noise with the environmental noise with the main component of other human noise, separating the host sound component from the background noise through the environmental noise reduction process, and forming a noise-reduced human sound signal;
step S102, echo cancellation processing: step S2002 is based on the acquired downlink sound, and the downlink echo is compared, and the downlink echo in the downlink sound is eliminated through the echo elimination processing in step S102 and the signal noise reduction processing in step S101;
step S103 gain adjustment processing: step S2003 judges through the volume of the uplink sound based on the uplink sound sent from the noise-reduced human voice signal, if the sound is smaller than the target amplitude, the uplink sound is sent out after the uplink sound is increased through the uplink gain adjustment, and if the sound is larger than the target amplitude, the uplink sound is sent out after the uplink sound is decreased through the uplink gain adjustment.
The invention adopts environment noise reduction treatment (ENC), echo elimination treatment (AEC), automatic Gain Control (AGC) and other treatments respectively, wherein the environment noise reduction treatment adopts environment noise reduction treatment, pre-emphasis treatment function, time domain windowing function, discrete Fourier transformation function, first weighted moving average filtering function, background subtraction elimination function, inverse Fourier transformation function and the like, and is subjected to amplitude limiting filtering and other treatments, the echo elimination treatment adopts echo elimination treatment, signal noise reduction treatment, weighted mixing function and other treatments respectively, and the automatic gain control adopts uplink gain adjustment and automatic gain treatment function respectively, thus the invention has the characteristics of good sound pickup effect, good noise reduction effect and stable uplink volume.
As shown in fig. 3, the ambient noise reduction process includes:
outputting a high-frequency information emphasis signal through a pre-emphasis processing function based on the voice noise;
outputting a windowed voice signal through a time domain windowing function based on the high-frequency information emphasis signal;
outputting a frequency domain energy spectrum signal through a discrete Fourier transform function based on the windowed human voice signal;
outputting a filtered human voice signal through a first weighted moving average filter function based on the frequency domain energy spectrum signal;
based on the filtered human voice signals and the environmental noise, outputting a noise reduction human voice frequency spectrum through a background subtraction elimination function;
and outputting the noise-reduced human voice signal through an inverse Fourier transform function and amplitude limiting filtering based on the noise-reduced human voice frequency spectrum or the filtering noise-reduced human voice frequency spectrum.
The invention aims to realize a pickup method with noise reduction functions (AEC, ENC) and Automatic Gain Control (AGC), which can be used on an uplink noise reduction earphone for a wearer to use when talking, and the implementation scheme comprises the following steps: after the sound data picked up by the microphone are transmitted to the algorithm module in the form of analog signals, the algorithm module performs environmental noise reduction (ENC) to separate the speaker sound of the wearer in the pick-up space from other background noise so as to achieve the purpose of noise reduction; further, the algorithm module can acquire downlink sound data of the earphone, namely, sound which is transmitted from the far end through the far end and played through the earphone loudspeaker during communication, and perform echo cancellation (AEC) to eliminate a small amount of echo which possibly is picked up by the communication microphone and leaked from the loudspeaker; when the sound is smaller (smaller standard), the uplink gain is increased, when the sound of the speaking of the wearer is larger (larger standard), the uplink gain is reduced, so that the uplink volume during the conversation is ensured, the algorithm module can detect the volume of the voice of the speaking of the wearer picked up in a period of time and perform Automatic Gain Control (AGC), and when the wearer is stable, the problems of poor uplink noise reduction performance and unstable volume during the conversation of the existing uplink noise reduction earphone are solved.
The pre-emphasis processing function is:
；
the saidAn original signal that is picked up human voice noise;
the saidIs the data after pre-emphasis treatment;
the independent variable is the number of sampling points;
the time domain windowing function is:
；
the saidIs the windowed data;
the saidIs a hamming window function;
the discrete fourier transform function is:
；
the saidThe frequency spectrum is obtained after discrete Fourier transform;
the saidPoints that are fourier transforms;
the saidIs normalized frequency;
the first weighted moving average filter function is:
；
the saidFiltering the human voice signal spectrum after energy spectrum filtering;
the saidFiltering the energy spectrum of the filtered human voice signal after the first energy spectrum;
the background subtraction elimination function is:
；
the saidFiltering the energy spectrum of the filtered human voice signal;
the saidAn energy spectrum that is ambient noise;
the saidAn energy spectrum obtained after spectrum subtraction;
the inverse fourier transform function is:
；
the saidA time domain signal obtained after inverse Fourier transform;
the saidA frequency spectrum obtained based on the first spectrum subtraction is subjected to filtering;
the saidPoints are inverse fourier transforms.
Since a pre-emphasis processing function, a time-domain windowing function, a discrete fourier transform function, a background subtraction elimination function, an inverse fourier transform function, etc. are adopted, 10ms of data is designated as one frame, and the picked-up data is processed once every 10 ms. First, pre-emphasis processing is performed on picked-up talk microphone and noise reduction microphone data to highlight high frequency information. The purpose of this is that the amplitude of the voice signal above 900Hz is reduced due to the interference of the oral cavity structure when the person sounds, and the frequency spectrum can be flattened through pre-emphasis, so that the processing is convenient. The formula is:
；
when a person speaks, vowel energy is mainly concentrated below 1kHz, consonants do not cause vocal cord vibration, and the frequency is higher. The pre-emphasis is to promote the high frequency component in the voice signal, so that the signal is distributed more uniformly on the frequency spectrum;
and then time domain windowing is carried out, wherein the purpose of the windowing is to reduce spectrum leakage and aliasing which can occur after framing of a voice signal, and the formula is as follows:
；
wherein the method comprises the steps ofAs a window function, a Hamming window is selected for the characteristics of the voice signal in the module; the Hamming window can effectively avoid spectrum leakage and has the characteristic of smooth low pass; in real-time voice processing, voice data is divided into frames for processing, for example, 10ms is intercepted to be a frame; however, such sampling and truncation may result in signal discontinuities and also non-periodic, which may result in spectral leakage when fourier transformed, resulting in distortion of the waveform. The windowing can enable the signal to present partial characteristics of a periodic function, can obviously reduce spectrum leakage, and enables the overall situation to be more continuous, so that the Gibbs effect is avoided;
and then performing discrete Fourier transform on the windowed signal to obtain frequency domain information, and converting the frequency domain information from a time domain to a frequency domain for further processing, wherein the formula is as follows:
；
and is composed of frequency spectrumObtaining the energy spectrum +.>。
The energy spectrum is then weighted, moving average filtered to smooth the signal, as:
；
the additive background noise is then cancelled by spectral subtraction:
；
and is composed ofObtain spectrum->. Subsequently pair->Weighted moving average filtering is performed again to eliminate additional noise that may be introduced. Will then->Frequency spectrum of pickup signal of microphone for talking from beginning +.>Multiplied and inverse fourier transformed to obtain:
；
finally toPerforming amplitude limiting filtering to obtain noise-reduced data; through the flow and the function, the noise reduction performance of the earphone is improved.
The ambient noise reduction process further includes: and outputting a filtered noise reduction human voice spectrum through a second weighted moving average filter function based on the noise reduction human voice spectrum.
The second weighted moving average filter function is:
；
after the addition of background noise is cancelled using spectral subtraction,and is made up of->Obtain spectrum->. Subsequently pair->Weighted moving average filtering is performed again to eliminate additional noise that may be introduced.
The echo cancellation process in step S102 includes: based on the delayed downlink sound, the echo-eliminated downlink sound is output through a weighted mixing function according to the environmental noise.
The weighted mixing function is:
；
the saidNoise reduction wheat signals;
the saidIs a stope signal;
the saidWeighted mixing of stope and noise reduction wheat signalsProcessed data;
the saidIs the sampling rate;
is the independent variable of the number of sampling points.
As the sound pickup device applied to the noise reduction earphone is less affected by the echo, when the downlink playing volume is very large and the earphone is separated from normal wearing of a person, slight echo can still be picked up, but the echo can not generate a complex echo channel through wall reflection like products in forms of conference sound boxes and the like. Therefore, the algorithm module does not need to do complex LMS filtering processing when eliminating echo, and only needs to do proper time delay on the stoping signal and then to carry out weighted mixing on the stoping signal and the data acquired by the noise reduction microphone, so that the stoping can be eliminated together when eliminating background noise, and the balance of performance and efficiency is obtained. Let the noise-reducing wheat signal beThe stope signal is +.>The sampling rate is +.>The formula is: />Through echo cancellation processing, the stability of uplink volume is ensured.
The uplink gain adjustment includes:
based on the uplink sound volume, and based on a detection threshold, a target value, a default gain and a stepping gain, acquiring the average amplitude of the previous frame;
if the average amplitude of the previous frame is lower than the detection threshold, outputting the current uplink sound of the self-increasing gain through an automatic gain processing function and limiting filtering according to the stepping gain;
if the average amplitude of the previous frame is higher than the target value, the current uplink sound with the self-subtracting gain is output through an automatic gain processing function and limiting filtering according to the stepping gain.
The automatic gain processing function is:
；
；
the saidIs the signal after gain;
the saidThe default gain is 1.0;
the saidIs the current gain;
the saidFor the step of gain adjustment, the step of gain adjustment is +.>Said->At a sampling rate of 48 kHz;
the saidAssigning an average value to the audio signal of the previous frame at the current time point;
the saidA detection threshold is effectively input;
the saidIs the target amplitude.
Because the effective input detection threshold is preset when the automatic gain control is performedTarget value->Default gain->And step->. Initially, the current gain +.>Equal to the default gain->. In operation, the average amplitude of the input signal of the frame preceding the current time point is calculated>. If->Below->The current gain is made to approach the initial gain +_ in fixed steps>Is changed in direction up to +.>Equal; if->Above->But is lower than->The current gain is increased automatically according to fixed steps; if->Above->The method comprises the steps of carrying out a first treatment on the surface of the The current gain is self-reduced in fixed steps. The noise-reduced signal is then combined with +.>Multiplying and clipping filtering the output result. The formula is:
；
；
meanwhile, the invention also provides a pickup noise reduction system, comprising: a pickup noise reduction platform implementing the pickup noise reduction method as described in any one of the above;
the pickup noise reduction platform comprises a signal noise reduction processing module;
step S101, signal noise reduction processing: the signal noise reduction processing module in step S2001 is configured to: based on the human noise with the main sound component picked up in the pick-up space, comparing the picked-up background noise with the environment noise with the main sound component of other human noise, separating the main sound component from the background noise through the environment noise reduction treatment, and forming a noise-reduced human sound signal;
step S102, echo cancellation processing: step S2002 is based on the acquired downlink sound, and the downlink echo is compared, and the downlink echo in the downlink sound is eliminated through the echo elimination processing in step S102 and the signal noise reduction processing in step S101;
step S103 gain adjustment processing: step S2003 judges through the volume of the uplink sound based on the uplink sound sent from the noise-reduced human voice signal, if the sound is smaller than the target amplitude, the uplink sound is sent out after the uplink sound is increased through the uplink gain adjustment, and if the sound is larger than the target amplitude, the uplink sound is sent out after the uplink sound is decreased through the uplink gain adjustment.
Meanwhile, the invention also provides a pickup noise reduction system, which comprises a pickup noise reduction platform, and a signal noise reduction processing module, wherein the signal noise reduction processing module realizes processing methods such as signal noise reduction processing, echo cancellation processing, gain adjustment processing and the like, and provides a practical system for pickup noise reduction processing.
Taking a noise reduction aviation earphone for aviation field as an example, the noise reduction pickup device installed on the noise reduction aviation earphone comprises a call microphone, a noise reduction microphone and a noise reduction processing module, and is provided with a stoping passage. The conversation microphone collects the voice signal with noise which the wearer speaks, the noise reduction microphone collects the environmental noise signal, and the stoping passage collects the sound signal played in the earphone. The noise reduction processing module delays the received stope signal for 15ms, and mixes the stope signal with the environmental noise signal according to weight to form a new noise signal; and then, respectively and sequentially carrying out pre-emphasis processing, time domain framing windowing processing, discrete Fourier transformation and weighted moving average filtering on two paths of signals, namely the noise-carrying voice signals and the noise signals, acquired by the communication microphone, subtracting the energy spectrums of the two paths of signals subjected to the processing, and then carrying out normalization processing to obtain a spectrum subtraction factor. The spectral subtraction factor is multiplied by the original noisy speech signal to obtain a speech signal from which the ambient noise and echo are removed. For a noise-cancelled speech signal, a specific volume gain is multiplied. In order to stabilize the volume of the picked-up voice signal, the average amplitude of the voice signal within 10ms is counted and compared with the noise threshold value and the target amplitude successively. If the gain is smaller than the noise threshold value, the volume gain is gradually changed to be a default gain of 1.0; if the noise threshold is greater than the target amplitude, increasing the volume gain to increase the amplitude of the voice signal; if the amplitude is greater than the target amplitude, the volume gain is reduced to reduce the amplitude of the speech signal.
Working principle:
the invention adopts environmental noise reduction (ENC), echo elimination (AEC), automatic Gain Control (AGC) and other processes, wherein the environmental noise reduction respectively adopts environmental noise reduction, pre-emphasis processing function, time domain windowing function, discrete Fourier transformation function, first weighted moving average filtering function, background subtraction elimination function, inverse Fourier transformation function and processes such as amplitude limiting filtering, the echo elimination respectively adopts echo elimination, signal noise reduction and weighted mixing function, and the automatic gain control respectively adopts uplink gain adjustment and automatic gain processing functions, thereby completing the environmental noise reduction, echo elimination and automatic gain control.
By using the technical scheme of the invention or under the inspired by the technical scheme of the invention, a similar technical scheme is designed by a person skilled in the art, so that the technical effects are achieved, and the technical scheme is considered to fall into the protection scope of the invention.
Claims (8)
1. A sound pickup noise reduction method, comprising:
and (3) signal noise reduction: based on the human noise with the main sound component picked up in the pick-up space, comparing the picked-up background noise with the environment noise with the main sound component of other human noise, separating the main sound component from the background noise through the environment noise reduction treatment, and forming a noise-reduced human sound signal;
echo cancellation processing: based on the obtained downlink sound, comparing the downlink echo, and eliminating the downlink echo in the downlink sound through echo elimination processing and signal noise reduction processing;
gain adjustment processing: based on the uplink sound sent by the noise-reduced human voice signal, through uplink sound volume judgment, if the sound is smaller than the target amplitude, the uplink sound is sent out after the uplink sound volume is increased through uplink gain adjustment, and if the sound is larger than the target amplitude, the uplink sound is sent out after the uplink sound volume is reduced through uplink gain adjustment;
wherein the ambient noise reduction process includes:
outputting a high-frequency information emphasis signal through a pre-emphasis processing function based on the voice noise;
outputting a windowed voice signal through a time domain windowing function based on the high-frequency information emphasis signal;
outputting a frequency domain energy spectrum signal through a discrete Fourier transform function based on the windowed human voice signal;
outputting a filtered human voice signal through a first weighted moving average filter function based on the frequency domain energy spectrum signal;
based on the filtered human voice signals and the environmental noise, outputting a noise reduction human voice frequency spectrum through a background subtraction elimination function;
based on the noise reduction human voice frequency spectrum or the filtering noise reduction human voice frequency spectrum, outputting noise reduction human voice signals through an inverse Fourier transform function and amplitude limiting filtering;
the pre-emphasis processing function is:
；
the saidAn original signal that is picked up human voice noise;
the saidIs the data after pre-emphasis treatment;
the independent variable is the number of sampling points;
the time domain windowing function is:
；
the saidIs the windowed data;
the saidIs a hamming window function;
the discrete fourier transform function is:
；
the saidThe frequency spectrum is obtained after discrete Fourier transform;
the saidPoints that are fourier transforms;
the saidIs normalized frequency;
the first weighted moving average filter function is:
；
the saidFiltering the human voice signal spectrum after energy spectrum filtering;
the saidFiltering the energy spectrum of the filtered human voice signal after the first energy spectrum;
the background subtraction elimination function is:
；
the saidFiltering the energy spectrum of the filtered human voice signal;
the saidAn energy spectrum that is ambient noise;
the saidAn energy spectrum obtained after spectrum subtraction;
the inverse fourier transform function is:
；
the saidA time domain signal obtained after inverse Fourier transform;
the saidA frequency spectrum obtained based on the first spectrum subtraction is subjected to filtering;
the saidPoints are inverse fourier transforms.
2. The pickup noise reduction method according to claim 1, wherein:
the ambient noise reduction process further includes: and outputting a filtered noise reduction human voice spectrum through a second weighted moving average filter function based on the noise reduction human voice spectrum.
3. The pickup noise reduction method according to claim 2, wherein:
the second weighted moving average filter function is:
。
4. the pickup noise reduction method according to claim 1, wherein:
the echo cancellation process includes: based on the delayed downlink sound, the echo-eliminated downlink sound is output through a weighted mixing function according to the environmental noise.
5. The pickup noise reduction method according to claim 4, wherein:
the weighted mixing function is:
；
the saidNoise reduction wheat signals;
the saidIs a stope signal;
the saidThe data of the stope signal and the noise reduction wheat signal after weighted mixing processing;
the saidIs the sampling rate;
is the independent variable of the number of sampling points.
6. The pickup noise reduction method according to claim 1, wherein:
the uplink gain adjustment includes:
based on the uplink sound volume, and based on a detection threshold, a target value, a default gain and a stepping gain, acquiring the average amplitude of the previous frame;
if the average amplitude of the previous frame is lower than the detection threshold, outputting the current uplink sound of the self-increasing gain through an automatic gain processing function and limiting filtering according to the stepping gain;
if the average amplitude of the previous frame is higher than the target value, the current uplink sound with the self-subtracting gain is output through an automatic gain processing function and limiting filtering according to the stepping gain.
7. The pickup noise reduction method according to claim 6, wherein:
the automatic gain processing function is:
；
；
the saidIs the signal after gain;
the saidThe default gain is 1.0;
the saidIs the current gain;
the saidFor the step of gain adjustment, the step of gain adjustment is +.>Said->At a sampling rate of 48 kHz;
the saidAssigning an average value to the audio signal of the previous frame at the current time point;
the saidA detection threshold is effectively input;
the saidIs the target amplitude.
8. A pickup noise reduction system, characterized in that: comprising the following steps: a pickup noise reduction platform for implementing the pickup noise reduction method according to any one of claims 1 to 7;
the pickup noise reduction platform comprises a signal noise reduction processing module;
the signal noise reduction processing module is used for: based on the human noise with the main sound component picked up in the pick-up space, comparing the picked-up background noise with the environment noise with the main sound component of other human noise, separating the main sound component from the background noise through the environment noise reduction treatment, and forming a noise-reduced human sound signal;
based on the obtained downlink sound, comparing the downlink echo, and eliminating the downlink echo in the downlink sound through echo elimination processing and signal noise reduction processing;
and based on the uplink sound sent by the noise-reduced voice signal, judging through the uplink sound volume, if the sound is smaller than the target amplitude, sending out the uplink sound after the uplink sound volume is increased through uplink gain adjustment, and if the sound is larger than the target amplitude, sending out the uplink sound after the uplink sound volume is reduced through uplink gain adjustment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311208195.4A CN116962934B (en) | 2023-09-19 | 2023-09-19 | Pickup noise reduction method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311208195.4A CN116962934B (en) | 2023-09-19 | 2023-09-19 | Pickup noise reduction method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116962934A CN116962934A (en) | 2023-10-27 |
| CN116962934B true CN116962934B (en) | 2024-01-09 |
Family
ID=88454918
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311208195.4A Active CN116962934B (en) | 2023-09-19 | 2023-09-19 | Pickup noise reduction method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116962934B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118102155B (en) * | 2024-04-23 | 2024-06-25 | 深圳市万屏时代科技有限公司 | Gain method and system of microphone |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101262530A (en) * | 2008-04-29 | 2008-09-10 | 中兴通讯股份有限公司 | A device for eliminating echo of mobile terminal |
| JP2010028653A (en) * | 2008-07-23 | 2010-02-04 | Nippon Telegr & Teleph Corp <Ntt> | Echo canceling apparatus, echo canceling method, its program, and recording medium |
| TW201506913A (en) * | 2013-08-15 | 2015-02-16 | Aver Information Inc | Microphone system and sound processing method thereof |
| CN209419829U (en) * | 2019-03-29 | 2019-09-20 | Tcl通力电子（惠州）有限公司 | A kind of echo cancellation signal Dolby circuit, device and electronic product |
| CN110931034A (en) * | 2019-11-27 | 2020-03-27 | 深圳市悦尔声学有限公司 | Pickup noise reduction method for built-in earphone of microphone |
| CN111327985A (en) * | 2020-03-06 | 2020-06-23 | 华勤通讯技术有限公司 | Earphone noise reduction method and device |
| CN211630378U (en) * | 2020-04-21 | 2020-10-02 | 深圳市昂纬科技开发有限公司 | Stereo sound heater |
| CN114189781A (en) * | 2021-11-27 | 2022-03-15 | 苏州蛙声科技有限公司 | Noise reduction method and system for double-microphone neural network noise reduction earphone |
| CN114495962A (en) * | 2022-01-12 | 2022-05-13 | 合肥讯飞数码科技有限公司 | Audio noise reduction method, device and system and computer readable storage medium |
| CN115802224A (en) * | 2022-10-18 | 2023-03-14 | 昆山联滔电子有限公司 | Noise elimination method and device, electronic equipment, earphone and storage medium |
-
2023
- 2023-09-19 CN CN202311208195.4A patent/CN116962934B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101262530A (en) * | 2008-04-29 | 2008-09-10 | 中兴通讯股份有限公司 | A device for eliminating echo of mobile terminal |
| JP2010028653A (en) * | 2008-07-23 | 2010-02-04 | Nippon Telegr & Teleph Corp <Ntt> | Echo canceling apparatus, echo canceling method, its program, and recording medium |
| TW201506913A (en) * | 2013-08-15 | 2015-02-16 | Aver Information Inc | Microphone system and sound processing method thereof |
| CN209419829U (en) * | 2019-03-29 | 2019-09-20 | Tcl通力电子（惠州）有限公司 | A kind of echo cancellation signal Dolby circuit, device and electronic product |
| CN110931034A (en) * | 2019-11-27 | 2020-03-27 | 深圳市悦尔声学有限公司 | Pickup noise reduction method for built-in earphone of microphone |
| CN111327985A (en) * | 2020-03-06 | 2020-06-23 | 华勤通讯技术有限公司 | Earphone noise reduction method and device |
| CN211630378U (en) * | 2020-04-21 | 2020-10-02 | 深圳市昂纬科技开发有限公司 | Stereo sound heater |
| CN114189781A (en) * | 2021-11-27 | 2022-03-15 | 苏州蛙声科技有限公司 | Noise reduction method and system for double-microphone neural network noise reduction earphone |
| CN114495962A (en) * | 2022-01-12 | 2022-05-13 | 合肥讯飞数码科技有限公司 | Audio noise reduction method, device and system and computer readable storage medium |
| CN115802224A (en) * | 2022-10-18 | 2023-03-14 | 昆山联滔电子有限公司 | Noise elimination method and device, electronic equipment, earphone and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116962934A (en) | 2023-10-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8751224B2 (en) | Combined microphone and earphone audio headset having means for denoising a near speech signal, in particular for a “hands-free” telephony system | |
| TWI433553B (en) | Automatic volume and dynamic range adjustment for mobile audio devices,and methods thereof | |
| CN108376548B (en) | Echo cancellation method and system based on microphone array | |
| TWI463817B (en) | System and method for adaptive intelligent noise suppression | |
| EP3791565B1 (en) | Method and apparatus utilizing residual echo estimate information to derive secondary echo reduction parameters | |
| KR101444100B1 (en) | Noise cancelling method and apparatus from the mixed sound | |
| US8180064B1 (en) | System and method for providing voice equalization | |
| US9699554B1 (en) | Adaptive signal equalization | |
| EP3282678B1 (en) | Signal processor with side-tone noise reduction for a headset | |
| US20040170284A1 (en) | Sound reinforcement system having an echo suppressor and loudspeaker beamformer | |
| CN108447496B (en) | Speech enhancement method and device based on microphone array | |
| KR20130108063A (en) | Multi-microphone robust noise suppression | |
| US9343073B1 (en) | Robust noise suppression system in adverse echo conditions | |
| CN111696567B (en) | Noise estimation method and system for far-field call | |
| CN116962934B (en) | Pickup noise reduction method and system | |
| CN111683319A (en) | Call pickup noise reduction method, earphone and storage medium | |
| WO2022218093A1 (en) | Audio signal compensation method and apparatus, earphone, and storage medium | |
| CN117321681A (en) | Speech optimization in noisy environments | |
| US8804981B2 (en) | Processing audio signals | |
| CN112637721A (en) | Bluetooth headset conversation noise reduction method and Bluetooth noise reduction headset | |
| CN114023352B (en) | Voice enhancement method and device based on energy spectrum depth modulation | |
| CN109874088A (en) | A kind of method and apparatus adjusting sound pressure level | |
| CN115866474A (en) | Transparent transmission noise reduction control method and system of wireless earphone and wireless earphone | |
| CN113038315A (en) | Voice signal processing method and device | |
| US11527232B2 (en) | Applying noise suppression to remote and local microphone signals |
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 |