CN1121684C - System for adaptively filtering audio signals to enhance speech intelligibility in noisy environmental conditions - Google Patents

System for adaptively filtering audio signals to enhance speech intelligibility in noisy environmental conditions Download PDF

Info

Publication number
CN1121684C
CN1121684C CN96198008A CN96198008A CN1121684C CN 1121684 C CN1121684 C CN 1121684C CN 96198008 A CN96198008 A CN 96198008A CN 96198008 A CN96198008 A CN 96198008A CN 1121684 C CN1121684 C CN 1121684C
Authority
CN
China
Prior art keywords
noise
filter
frame
value
filter circuit
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired - Fee Related
Application number
CN96198008A
Other languages
Chinese (zh)
Other versions
CN1201547A (en
Inventor
T·W·索尔维
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Ericsson Inc
Original Assignee
Ericsson Inc
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Ericsson Inc filed Critical Ericsson Inc
Publication of CN1201547A publication Critical patent/CN1201547A/en
Application granted granted Critical
Publication of CN1121684C publication Critical patent/CN1121684C/en
Anticipated expiration legal-status Critical
Expired - Fee Related legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L21/00—Processing of the speech or voice signal to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
    • G10L21/02—Speech enhancement, e.g. noise reduction or echo cancellation
    • G10L21/0208—Noise filtering
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L21/00—Processing of the speech or voice signal to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
    • G10L21/02—Speech enhancement, e.g. noise reduction or echo cancellation
    • G10L21/0208—Noise filtering
    • G10L21/0216—Noise filtering characterised by the method used for estimating noise
    • G10L2021/02168—Noise filtering characterised by the method used for estimating noise the estimation exclusively taking place during speech pauses
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
    • G10L25/78—Detection of presence or absence of voice signals
    • G10L2025/783—Detection of presence or absence of voice signals based on threshold decision
    • G10L2025/786—Adaptive threshold
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L21/00—Processing of the speech or voice signal to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
    • G10L21/02—Speech enhancement, e.g. noise reduction or echo cancellation
    • G10L21/0208—Noise filtering
    • G10L21/0216—Noise filtering characterised by the method used for estimating noise
    • G10L21/0232—Processing in the frequency domain

Abstract

A method and system are provided for adaptively reducing noise in frames of digitized audio signals that include both speech and background noise. Frames of digitized audio signals are passed through an adjustable, high-pass filter circuit to filter a portion of background noise located in a low frequency range of the digitized signal. The filter circuit is adjusted by a filter control circuit adapted for a current frame to exhibit a selected frequency response curve. The filter control circuit includes a speech detector for detecting the presence or absence of speech in the frames of digitized audio signals. The filter circuit is adjusted when no speech is detected in the current frame. In a first preferred embodiment, the filter control circuit controls the filter circuit by calculating a noise estimate corresponding to the background noise, and adjusting the filter circuit based on the noise estimate. As the noise estimates increase, the filter circuit is adjusted to extract increasing amounts of energy falling in low frequency ranges of speech. In a second preferred embodiment, the filter circuit is adjusted as a function of a noise profile estimate. A noise profile estimate for a current frame is determined as a function of speech detection and is compared to a reference noise profile. Based on this comparison, the filter circuit is adaptively adjusted.

Description

Be used for optionally changing the method and apparatus of a frame of digital signal
Technical field
This invention relates to the noise-cut system, the particularly a kind of adaptive voice sharpness enhanced system of portable digital cordless phones and method and apparatus that is used for optionally changing a frame of digital signal of being used for.
Background technology
In the commercial operation of the U.S. and other parts of the world, cellular telephone industry has obtained significant progress.At main urban place, the demand of cellular service is being surpassed the capacity of existing system.Suppose that this trend continues, the cellular radio communication will touch even minimum rural market.Capacity of cellular network must strengthen when therefore, keeping high-quality service with reasonable cost.Towards one step of weight of improving capacity is that cellular system is from the conversion of simulation transmission to digital delivery.This conversion also is important, this is because first generation person-to-person communication net (PCNs) may be provided by the honeycomb substrate that uses digital cellular foundation structure of future generation, wherein, first generation person-to-person communication net adopted be easy to carry and be convenient at home, the low cost that make a phone call or answer the call in ground in the office, the street, automobile etc., the wireless phone of pocket size.
Digital communication system has utilized strong Digital Signal Processing.Digital signal processing is commonly referred to as the mathematics or the otherwise processing of digitized signal.For example, be after the digital form with analog signal conversion (digitizing), may utilize the simple mathematical routine in the digital signal processor (DSP) to come filtering, amplification and this digital signal that decays.Digital signal processor generally is manufactured to high speed integrated circuit, makes data processing operation carry out in real time basically.Digital signal processor also can be used to reduce the bit transfer rate of digitize voice, and its result shows as the reduction of spectrum occupation rate and the power system capacity that send radio signal and enlarges.If for example use the linear pulse code modulated of 14 bits (PCM) to come digitized voice signal, and, can produce 112K bps serial bit rate with the sampling of the sampling rate of 8KHZ.In addition, cut down with the bit transfer rate that obtains 14: 1 by utilize the redundant characteristic and the measurable characteristic of other human speech, acoustic coding technology can be used for that 112K bps serial bit rate is compressed to 7.95K bps from mathematics.Transfer rate is cut down and is meaned the bandwidth that get more.
In the U.S., a kind of popular voice compression technique that is used as the digitizing standard of second generation cell phone system (i.e.IS-54) by TIA is the linear predictive coding (VSELP) of vector source code book excitation.Unfortunately, when the sound signal that comprises voice and be mixed with high level neighbourhood noise (especially " coloured noise ") is utilized VSELP coding/compression, may comprise the undesirable characteristic audio signal of part in the result.For example, if use digital mobile phone in noise circumstance (for example, wherein in the car of moving vehicle), the voice of neighbourhood noise and hope all are used the compression of VSELP encryption algorithm, and being sent to the base station, the signal in the base station after the compression is decoded and be reconfigured as the voice that can listen.When ground unrest is reconfigured as analog form, listened to the distortion that is not hoped of noise, and accidental these situations that betide in the voice are introduced into.This distortion is stinking for general audience.
This distortion major part is caused by the environment that uses mobile phone.Mobile phone generally is used for vehicle interior, the noise that wagon flow caused around the neighbourhood noise that usually has car engine to produce there reached.The neighbourhood noise of this vehicle interior concentrates on the bass scope usually, and noise amplitude can since the speed of vehicle and acceleration and on every side the such factor of the big or small degree of vehicle flowrate change.This low-frequency noise also has such trend: the serious speech intelligibility that reduces from teller in the automobile.In the communication system that adopts the VSELP vocoder, the reduction of this speech intelligibility that is caused by low-frequency noise may be especially remarkable, but this phenomenon also may betide in the communication system that does not comprise the VSELP vocoder.
Neighbourhood noise also may work because of the mode of using mobile phone to the influence of mobile phone.Especially mobile phone can be used for hands-free way, and promptly the phone user is facing to the mobile phone speech that is placed in the carriage.This makes mobile phone user's hand can spare driving, arrives the distance that must pass by before the mobile phone microphone input end but also increased the words that the user tells.The distance that has strengthened between this user and the mobile phone adds the neighbourhood noise of variation, signal portion of the sound signal general power spectrum energy that can cause noise to become being input to mobile phone.
The description of prior art is included in EP 0 645 756, and EP 0 558 312, and EP 0 665530, DE 4 012 349, U.S. Patent No. 4,811,404,4,461,025, and in 5,251,263, all these has described the mode that signal content is not wished in filtering.
In theory, can utilize digital signal processor to realize the ground unrest of various digital signal processing algorithms with filtering VSELP coding.Yet these solutions usually need be carried out the very big digital signal processing expense that millions of instructions (MIPS) are calculated with per second, and this has expended the valuable processing time, memory headroom and power consumption.Yet in the portable radiotelephone phone, each in these signal processing resources all is limited.Therefore, for the ground unrest that minimizes the VSLEP coding and the ground unrest of other types, the processing burden that increases DSP simply is not a best solution.
Brief summary of the invention
This invention has provided a kind of adaptive noise reduction system, and this system has cut down the effect of undesirable coding ground unrest any negative effect to the encoded voice quality being minimized and any of digital signal processing resources being increased consumption the minimized while.The method and system of this invention has increased the sharpness of voice in the digital audio signal, and its method is by a filter circuit with the digital audio signal frame.This filter circuit plays the scalable Hi-pass filter, and the digitized signal of its filtering part bass scope is by dropping on the digitized signal part in the high-frequency range.Because the noise in the vehicle is tending towards concentrating on the bass scope, and has only sub-fraction lamprophonia value to drop on this low-frequency range, so filtering circuit is when only filtering out the inessential part of voice, filtering the most of noise in the digital audio signal.This makes and partly compares with the speech energy of removing that the noise energy of a bigger part is removed relatively.By adjusting the also frequency response curve of selective filter circuit adaptively, the speech volume that filters out is limited, and the speech intelligibility of radio output is had minimum influence.
Filter control circuit is used to adjust filter circuit, makes it to show different frequency response curves with the form of a certain noise estimation value and/or spectrum envelope function, and noise estimation value wherein and/or spectrum envelope are corresponding to the noise in the sound signal.Noise estimation value and/or spectrum envelope adjust at digital signal on basis frame by frame and adjust with the form of speech detection function.If do not detect voice, that is petty to be present frame correction noise estimation value and/or spectrum envelope.If detect voice, just do not adjust noise estimation value and/or spectrum envelope.
In the first embodiment, filter circuit is at digitized audio signal frame calculating noise estimated value.This noise estimation value is corresponding to the amount of background noise in the digital audio signal frame.When the ground unrest in the voice low-frequency range increased the relative quantity of voice, noise estimation value increased.When the ground unrest in the voice low-frequency range increased the relative quantity of voice, filter control circuit used noise estimation value to adjust filter circuit with the more most low-frequency range voice of filtering.When not having ground unrest, there is not voice signal by filtering.When having higher noise level, more most noise and voice messaging are extracted.Because noise trends towards concentrating in the low-frequency range and has only relative smaller portions lamprophonia value to drop in this low-frequency range, when noise estimation value increases, by strengthening by the low frequency energy part of filtering, the whole sharpness of sound signal can be enhanced.
In second embodiment, an amended filter control circuit is used to adjust filter circuit, make it to demonstrate different frequency response curves with the form of a certain noise envelope function, noise envelope wherein is a noise envelope of selecting noise estimation value on the frequency range in the sound signal.This filter control circuit comprises a spectralyzer, and this analyzer is determined a noise envelope estimated value with the form that detects the voice function.Determining a noise envelope estimated value for present frame also compares this estimated value with the reference noise envelope.Based on this relatively, filter circuit is adjusted the low frequency energy that is used for extracting varying number from present frame adaptively.
Cut down system according to the self-adaptation of this invention and can be advantageously applied to radio communication system, in this radio communication system, pass through the communication of RF channel between portable/mobile radio transceiver and between radio transceiver and the fixation phone line user.Each transceiver comprises an antenna, and a radio signal that is used for receiving by antenna on the RF channel converts the receiver of simulated audio signal to, and a transmitter.Transmitter comprises that a volume-demoder (codec) is used for the simulated audio signal that will be sent out is digitized as the digitized speech information frame, and this voice messaging had both comprised that voice also comprised ground unrest.Handle present frame on the basis of digital signal processor speech detection in ground unrest estimated value and present frame and come the minimum background noise.Digitized speech information frame after modulator will be handled is modulated to and is used for follow-up transmission by antenna on the RF carrier wave.
Accompanying drawing is briefly described
According to the description of writing below also in conjunction with the accompanying drawings, for those skilled in the art, all features and the advantage of this invention will be readily understood that.
Fig. 1 is a general utility functions block scheme of this invention.
Fig. 2 illustrates frame and the positional structure of the U.S. digital standard IS-54 that is used for the cellular radio communication;
Fig. 3 is that this that utilize that digital signal processor realizes invented the block scheme of first preferred embodiment;
Fig. 4 is the functional-block diagram of an exemplary embodiment of this invention, and this embodiment is applied in the radio communication system in a plurality of portable radio transceivers one.
Fig. 5 A and 5B are a process flow diagram, and it illustrates, and this is invented in the first preferred embodiment process in realization, function/operation that digital signal processor is carried out.
To be graphic extension invent first exemplary plot of decay-frequency characteristic of the filtering circuit of first preferred embodiment according to this to Fig. 6 A.
To be graphic extension invent second exemplary plot of decay-frequency characteristic of the filtering circuit of first preferred embodiment according to this to Fig. 6 B.
Fig. 7 can be invented a sample query table of filter controller circuit access in first preferred embodiment by this.
Fig. 8 A and 8B illustrate the amplitude-frequency characteristic of example input audio signal.
Fig. 9 A and 9B illustrate input audio signal among Fig. 8 A and the 8B respectively by the amplitude-frequency characteristic after the filter circuit filtering of this invention;
Figure 10 is that this that utilize that digital signal processor realizes invented the block scheme of second preferred embodiment;
Figure 11 is a process flow diagram, and corresponding to the process flow diagram of Fig. 5 B, it illustrates, and this invents function/operation that digital signal processor is carried out in the second preferred embodiment process in realization.
Figure 12 can be invented a sample query table of the second preferred embodiment median filter control circuit access by this.
Accompanying drawing is described in detail
In the following description, for purposes of explanation and not limitation, in order to provide the complete understanding to this invention, concrete details such as special circuit, circuit component, technology, process flow diagram or the like are stated.Yet these those skilled in the art can understand that this invention can be practiced in other embodiments that depart from these details.In other examples, the detailed description of well-known method, equipment and circuit all is omitted so that can not blur description to this invention with unnecessary details.
Fig. 1 is the general block diagram according to the adaptive noise reduction system 100 of this invention.Adaptive noise reduction system 100 comprises a filter control circuit 105 that is connected to filter circuit 115.Filter control circuit 105 is that the present frame of digital audio signal produces a filter control signal.This filter control signal is output to filter circuit 115, and filter circuit 115 is regulated to demonstrate a high-pass equipment response curve according to filter control signal, and this curve is to select on the basis of filter control signal.The present frame of the filtering circuit 115 filtering figure sound signals after the adjusting.Filtering signal is handled the coded signal that produces the expression digital audio signal by vocoder 120.
Be applied in cellular radio communication systems in the exemplary scheme of this invention of portable/mobile radiotelephone transceiver, Fig. 2 illustrates the frame structure of the time-division multiple access (TDMA) that the IS-54 standard adopts for the digital cellular radio communication.One " frame " is one 20 milliseconds time period, and it comprises that sends a piece TX, and a reception piece RX and a signal strength measurement piece are used for Mobile Assisted Handover (MAHO).Two successive frames expressing among Fig. 2 were sent out in 40 milliseconds time period.Digitized voice and background noise information are as further describing on basis frame by frame processed and filtering below.
Best is, the filter control circuit 105 among Fig. 1, and the function of filter circuit 115 and vocoder 120 realizes with a high speed digital signal processor.A kind of suitable digital signal processor is can be from the TMS320C53 DSP of TI company acquisition.This TMS320C53 DSP comprises one 16 bit microprocessor on integrated an of list, be used to store RAM on the sheet of the data as speech frame that will be processed, the ROM that is used for the store various kinds of data Processing Algorithm, algorithm wherein comprises the VSELP voice compression algorithm, and following other algorithm that is used for finishing the functional block of being carried out by filter control circuit 105 and filter circuit 115 that will describe.
First embodiment of this invention is expressed out in Fig. 3.In the first embodiment, filter circuit 115 is adjusted with the form of ground unrest estimated value function.This ground unrest estimated value is determined by filter control circuit.The audio-frequency information frame of pulse code modulated by sequential storage on the sheet of DSP among the RAM.Can use other digitizing technique to come digital audio information.Each PCM digital frame is removed the RAM from the DSP sheet and is handled by frame Energy Estimation device 210, is temporarily stored in then in the interim frame memory 220.The energy of the present frame of being determined by frame Energy Estimation device 210 is provided for noise estimator 230 and speech detector 240 functional blocks.When frame Energy Estimation value surpass last noise estimation value and a voice threshold value and the time, speech detector 240 represents that voice are present in the present frame.If speech detector determines not have voice to exist, digital signal processor 200 calculates the noise estimation value of a correction with the form of current noise estimation value and present frame energy function so.
Revised noise estimation value is output to filter selector 235.Filter selector 235 produces a filter control signal based on noise estimation value.In preferred embodiments, filter selector 235 reads question blank in producing the filter control signal process.Question blank comprises a series of FILTER TO CONTROL value, and each controlling value all is complementary with the scope of a noise estimation value or noise estimation value.On the basis of noise estimation value after the correction, the FILTER TO CONTROL value in the question blank is selected to be gone out, and this FILTER TO CONTROL value is represented that by a filter control signal this control signal is that filter circuit 115 outputs to bank of filters 265.In order to stablize this process, and avoid the continuous switching between different wave filters, be provided with the switching time of a N frame for the selection of new wave filter.New wave filter can only be selected once by every N frame, wherein N be one greater than 1, and more preferably greater than 10 integer.
Filter circuit 115 is adjusted to demonstrate the high-pass equipment response curve corresponding to input filter control signal and noise estimation value according to filter control signal.Well-known various dissimilar filter circuits can be used to show the frequency response curve of choosing according to filter control signal in the prior art.The wave filter of these prior aries comprises iir filter, as Butterworth, and Chebyshev or elliptic filter, because lower processing requirements also can use the FIR wave filter, but preferred iir filter.
Filtered signal is handled by vocoder 120, and 120 are used to the bit rate of signal behind the compression filtering.In preferred embodiments, vocoder 120 uses vector source code book excited linear prediction encoding (VSELP) technology to come coding audio signal.Other acoustic coding technology and algorithm also can be used, for example Code Excited Linear Prediction (CELP) coding, residual impulse Excited Linear Prediction (RPE-LTP) coding, improved many band excitations (IMBE) coding.By before acoustic coding according to this invention filtering audio signals frame, ground unrest is minimized, any undesirable noise effect in the voice when voice reconstruct has been cut down in this processing basically.It has prevented that also voice from " being flooded " in low-frequency noise.
The digital signal processor of describing in conjunction with Fig. 3 200 can be used in the such device of the transceiver of digital portable/mobile radiotelephone of using in the radio communication system for example.Fig. 4 illustrates such digital radio transceiver, and it can be used for the cellular radio communication network.
The sound signal that comprises voice and ground unrest is input to a preferably special IC (ASIC) of volume-demoder 402,402 from microphone 400.At the detected band limited audio signals in microphone 400 places by codec 402 with the sampling of the sampling rate of 8000 samples of per second and be blocked into frame.According to above-mentioned, each 20 milliseconds of frame comprises 160 speech samples.These samples are quantized and are converted into for example such coded digital form of 14 bit linear PCM.In case 160 samples of the digitize voice of present frame are stored on the sheet that sends among the DSP200 among the RAM, send DSP200 as above describe in conjunction with Fig. 3, according to the VSELP algorithm, carry out the chnnel coding function, the frame Energy Estimation, Noise Estimation, speech detection, FFT, filter function and digital speech code/compression.
Monitoring microprocessor 432 is being controlled the whole operation of all elements in the transceiver of expressing among Fig. 4.Be provided and be used for orthogonal modulation and transmission by sending filtered PCM data stream that DSP200 produces.So far, based on PCM data stream after the filtering that derives from DSP 200, ASIC gate array 404 produces homophase (I) information channel and quadrature (Q) information channel.The low-pass filter 406 that I and Q bit stream are complementary and 408 is handled and is sent in the IQ mixer in the balanced modulator 410.Provide a transmission intermediate frequency (IF) with reference to oscillator 412 and multiplier 414.I signal mixes with homophase IF, and Q signal mixes (being that homophase IF is by phase-shifter 416 quadrature laggings) with quadrature IF.The I and the Q signal that mix are added, and by " on " be transformed on the RF channel frequency of selecting by channel synthesizer 430, send on the radio frequency channel of selecting by duplexer 420 and antenna 422 then.
Receiving the limit, the signal that receives by antenna 422 and duplexer 420 is converted to an IF frequency downwards by choosing from mixer 424 on the receive channel frequency, an IF frequency has wherein been used the local oscillator signal that is synthesized by channel synthesizer 430 on the output basis of reference oscillator 428.Filtered and its frequency of the output of the one IF mixer 424 is converted to the 2nd IF downwards, and this conversion is carried out on the basis of another output of channel synthesizer 430 and detuner 426.Then, receiving gate array 434 becomes a series of phase sample and a series of frequency samples with the 2nd IF conversion of signals.Receive DSP436 and carry out demodulation, filtering, gain/attenuation, channel-decoding and voice expansion to the received signal.Then, the speech data after the processing is sent to codec 402 and is converted into base-band audio signal and is used to drive loudspeaker 438.
Now will be in conjunction with the flow chart description digital signal processor among Fig. 5 A, the 5B 200 for realizing filter control circuit 105, the function of filter circuit 115 and vocoder 120 and the operation carried out.Frame Energy Estimation device 210 is determined the energy of each frame sound signal.By calculate each PCM sample square value in the frame and (step 505), frame Energy Estimation device 210 is determined the energy of present frames.Because for the sampling rate of per second 8000 samples, each 20 milliseconds of long frame has 160 samples, and that is petty just to have 160 PCM samples square to be added.Express according to mathematical way, the frame Energy Estimation is determined according to equation 1: Equation 1
The frame energy value that calculates for present frame is stored on the sheet of DSP200 among the RAM202 (step 510).
The function of speech detector 240 comprises takes out one by the noise estimator 230 previous noise estimation value of determining (step 515) the RAM202 from the sheet of DSP200.Certainly, when transceiver powers at first, there is not noise estimation value to exist.Decision block 520 expects this situation and provides a noise estimation value in step 525.For what will describe below resembling, force correction to noise estimation value, preferably arrange any high value as noise estimation value, for example the 20dB on the normal voice level.The frame energy of being determined by frame Energy Estimation device 210 is taken out (square frame 530) the RAM202 from the sheet of DSP210.In square frame 535, determine frame Energy Estimation value whether surpassed the noise estimation value that detects add a predetermined voice threshold value and, as following equation 2 expressions:
Frame Energy Estimation value＞(noise estimation value+voice threshold value) (equation 2)
The voice threshold value can be a fixed value, and this fixed value is determined short-time energy variance greater than general ground unrest by experience, and can be set to for example 9dB.In addition, the voice threshold value can be revised the voice condition that reflects variation adaptively, for example, and the voice condition when the talker enters a more noisy or more quiet environment.If frame Energy Estimation value surpassed in the equation 2 and, that is petty to be provided with a zone bit and to represent that voice exist in square frame 570.If speech detector 240 detects voice and exists, that petty noise estimator 230 is crossed, and the noise estimation value of calculating for former digitized audio frame is retrieved and be used as current noise estimation value.Opposite, if the frame Energy Estimation less than in the equation 2 and, at square frame 540 with the zero clearing of voice sign.
Also can use other to detect the system of voice in the present frame.For example, European telecommunication standards body (ETSI) has developed a kind of standard that GPS GSM voice activity detects (VAD) that is used for.And at the ETSI list of references: be described among the RE/SMG-020632P, the document is quoted at this by reference.
If voice do not exist, the noise estimation value correction routine in the noise estimator 230 is performed.In the time that does not have voice to exist, it is online average that noise estimation value comes down to of frame energy.As described above, if the selecteed enough height of initial startup noise estimation value, those petty voice are not detected, and the voice sign by therefore zero clearing to force correction to noise estimation value.
In the Noise Estimation routine of carrying out by noise estimator 230, in square frame 545, determined a difference/error (Δ), according to equation, this error is the frame noise energy of frame Energy Estimation device 210 generations and the difference between the noise estimator 230 former noise estimation value of calculating:
Δ=present frame energy-former Noise Estimation (equation 3)
Decision block 550 determines whether that Δ has surpassed 0.If Δ is born, take place during as the strong noise estimated value, recomputated in square frame 560 according to the equation noise estimation value so:
Noise Estimation=former Noise Estimation+Δ/2 (equation 4)
Because Δ is born, this causes noise estimation value to be proofreaied and correct downwards.Here selected relatively large step delta/2 to proofread and correct fast to reduce noise level.Yet, if the frame energy has surpassed noise estimation value, providing one greater than 0 Δ, noise is by according to the equation correction in square frame 555:
Noise Estimation=former noise estimation value+Δ/256 (equation 5)
Because Δ is positive, noise estimation value is bound to increase.Yet, chosen littler step delta/256 (comparing) here and strengthened noise estimation value gradually and instantaneous noise is made substantial elimination with Δ/2.
The noise estimation value of calculating for present frame is output to filter selector 235.In first preferred embodiment, filter selector 235 reads question blank and utilizes current noise estimation value to choose a FILTER TO CONTROL value (step 572).Filter circuit 115 (step 574) demonstrates a frequency response curve with the form adjustment of selecting the FILTER TO CONTROL value function then, and this response curve is intended to strengthen the noisiness of filtering when noise estimation value and ground unrest increase.Then, the PCM sample that is stored among the DSP RAM comes filtering PCM sample to remove denoising (step 576) by adjusted filter circuit 265.Handled (step 578) by vocoder 120 after the filtered PCM sample, then, the sample behind the coding is output to RF transtation mission circuit (step 580).
Fig. 6 A has provided several examples that show different frequency response curve F1-F4 for the different filter control signals that are input to filter circuit 115 of how adjusting about filter circuit with 6B.As shown in Figure 6A, filter circuit 115 can be selected to show a series of different frequency response curves, and frequency response curve F1-F4 has cutoff frequency F1c-F4c respectively.In preferred embodiments, the scope of the cutoff frequency of filter circuit 115 can be that 300HZ is to 800HZ.When noise estimation value increased, filter circuit 115 was designed to show to have the more frequency response curve of higher cutoff frequency.This higher cutoff frequency causes more, and major part drops on filtered device circuit 115 extractions of the interior frame energy of voice low-frequency range.
Same, shown in Fig. 6 B, filter circuit 115 can be selected to show a series of different frequency response curve F1-F4, and each frequency response curve has the different gradients and identical cutoff frequency.In the scope that the cutoff frequency of frequency response curve F1-F4 is mentioned in the above.When noise estimation value increased, filter circuit 115 was adjusted to show to have the more frequency response curve of steep gradient.This steeper gradient causes more, and major part drops on interior filtered device circuit 115 extractions of frame energy of low frequency ranges of voice.
Filter circuit 115 comes the filtering present frame with the form of a certain Noise Estimation value function, and noise estimation value is wherein calculated for present frame.The filtered noise that makes of present frame has been passed through the major part of voice by reduction.Do not provided discernible voice output by filtering and the voice major part passed through, quality of speech signal has only very little reduction.The combination of different cutoff frequencys and different gradient can be used to extract adaptively the part of selecting that drops on the interior frame energy of voice low-frequency range.
Fig. 7 has described a sample query table that is read by filter selector 235, so that select one for filter circuit 115 from filter response curve F1-F4.This question blank comprises a series of possible noise estimation value N1-Nn and FILTER TO CONTROL value F1-Fn, the possible response curve that these values show corresponding to filter circuit 115.Among the noise estimation value N1-Nn each can be represented the noise estimation value of a scope, and each all is complementary with a specific FILTER TO CONTROL value F1-F4.Filter control circuit 105 produces a filter control signal, and its method is to calculate a noise estimation value and detect associated FILTER TO CONTROL value from question blank.
Fig. 8 A﹠amp; B and 9A﹠amp; B represented each frame of two frame sound signals be how by auto adapted filtering to provide an improvement sound signal that outputs to the RF transmitter.Fig. 8 A and 8B have represented to comprise respectively speech components s1, s2 and noise component n1, one first frame and one second frame of the sound signal of n2.As shown in the figure, noise energy n1 in two frames and n2 concentrate in the bass scope.And speech energy s1 and s2 concentrate in the high audio scope.Fig. 9 A has provided the noise signal n1 and the voice signal s1 of first frame after the filtering.Fig. 9 B has provided the noise signal n2 and the voice signal s2 of second frame after the filtering.
As what discussed, adaptive audio noise reduction system 100 is designed to calculate the difference of noise level between first frame and second frame, and noise estimation value was adjusted filter control circuit 105 after its method was based on the calculating of present frame.For example, filter control circuit 105 calculates Noise Estimation N1 and spectrum envelope s1 and is that first frame is selected a FILTER TO CONTROL value F1.In preferred embodiments, based on FILTER TO CONTROL value F1, filter circuit 115 is adjusted and resembles shown in Fig. 6 A, demonstrates a frequency response curve with cutoff frequency F1c.Then, first frame is by this adjusted filter circuit 115.Filter circuit 115 selected most noise n1 and the only a fraction of voice s1 of making drop under the cutoff frequency F1c of frequency response curve F1.This cause noise n1 by effective filtering and only some unessential relatively voice s1 by filtering.The filtered first frame sound signal is illustrated among Fig. 9 A.
In second frame that shows, have higher ground unrest in Fig. 8 B, and suppose that voice are not detected, that petty filter control circuit 105 can calculate a higher noise estimation value n2.Based on this strong noise estimated value more, be that second frame is determined a higher respective filter controlling value F2.In first preferred embodiment, adjust filter circuit 115 according to higher FILTER TO CONTROL value F2 and resembling that Fig. 6 A represents, show that has the more frequency response curve of higher cutoff frequency F2c.Then, the subsequent frame of sound signal has passed through adjusted filter circuit 115.Because for subsequent frame, the cutoff frequency F2c of frequency response curve F2 is higher, so most noise n2 and voice s2 are by filtering.(still), the sharpness information that voice s2 is comprised by the part of filtering and this frame is compared still relative not remarkable, so this only has slight influence to voice.The shortcoming of the more most voice s2 of filtering is offset by the advantage that noise n2 removal amount in second frame increases.The speech manual part that is filtered out does not significantly act on the sharpness of voice.Filtered audio signal is expressed in Fig. 9 B in second frame.
Provided second preferred embodiment of adaptive noise reduction system 100 among Figure 10-12.In second preferred embodiment, filter control circuit 105 is adjusted filter circuit 115 with the form of noise envelope estimated value function.The noise envelope estimated value be calculate at each frame and with reference noise envelope estimated value relatively.Based on this relatively, filter circuit 115 is adjusted the low frequency energy that extracts varying number from present frame adaptively.
With reference to Figure 10, provided DSP 200 according to the configuration of second preferred embodiment.As shown in the figure, except the frame Energy Estimation device of describing with reference to first preferred embodiment 210, noise estimator 230, outside speech detector 240 and the filter selector 235, filter control circuit 105 also comprises spectralyzer 270.Described represented with process flow diagram 5A and 5B as first embodiment, filter control circuit 105 is determined noise estimation value for the frame that receives and is detected the existence of voice.When detecting voice for present frame, spectralyzer 270 is revised the noise envelope estimated value, and uses this value in adjusting filter circuit 115.
With reference to Figure 11, provided the step of revising the noise envelope estimated value and adjusting filter circuit 115.Figure 11 has provided the step that spectralyzer 270 is carried out, and is cited in the whole process of describing among the process flow diagram 5A of former first preferred embodiment of these steps and the 5B.
If do not detect voice in the present frame, spectralyzer 270 is at first determined a noise envelope (step 600) for present frame.For comprising the energy calculated value on the different frequency (being Frequency point) in the definite noise envelope of present frame, these frequencies are arranged in the voice low-frequency range of selecting for present frame.In preferred embodiments, the frequency range of choosing is approximately 300 to 800HZ.The noise envelope of present frame can be handled present frame by the fast Fourier transform (FFT) that utilization has a N Frequency point and determine.Utilizing the FFT processing digital signal is well-known in the prior art field, and it is superior to be in being confined to less relatively Frequency point as FFT, and for example 32 the time, it needs processing power seldom.FFT with N Frequency point calculates N different frequency place produce power.The energy calculated value that drops on the Frequency point in the frequency range of choosing has formed the noise envelope of present frame.
In order to determine the noise envelope estimated value (step 604) of present frame, with the noise envelope of present frame with do on average for the definite noise envelope estimated value of the former frame of sound signal.When the noise envelope estimated value before not having can obtain, for example after the initialization, can use the initial noise envelope estimated value of storage.The noise envelope estimated value comprise be positioned at continuous low frequency (that is, and in the frequency range of choosing, e 1Be the noise energy estimated value of highest frequency and e nBe the noise energy estimated value of low-limit frequency) on noise energy estimated value e i(i=1 wherein, 2 ... n).In preferred embodiments, each noise energy estimated value e iCorresponding to the mean value of energy calculated value on a certain characteristic frequency, this characteristic frequency is a Frequency point in a large amount of selected frequencies scopes that wherein do not have on the successive frame that voice are detected.By using a large amount of frames to determine the noise envelope estimated value, filter circuit 115 is adjusted on a more progressive basis.In another embodiment, the noise envelope estimated value can equal the noise envelope of present frame.
Then, the Energy Estimation value e of noise envelope estimated value iCompare (step 604) with the reference noise envelope.The reference noise envelope comprises reference energy threshold value e Ri(i=1 wherein, 2 ... n), these threshold values are positioned at the noise energy estimated value e corresponding to the noise envelope estimated value iFrequency point on.Reference energy threshold value e RiCan determine by experience.According to from highest frequency Energy Estimation value e 1To low-limit frequency Energy Estimation value e nOrder, noise energy estimated value e iBy continuous and corresponding reference energy threshold e RiCompare.
More specifically, noise energy is estimated e 1At first with reference noise threshold value e R1Compare.If e 1Greater than reference noise threshold value e R1, that petty fiducial value C1 is selected and be input to filter selector 235.If noise estimation value e 1Less than reference noise threshold value e Rl, that petty noise energy estimated value e 2(this value is for to be lower than e 1The noise energy estimated value that obtains of frequency place) with reference threshold e R2Compare.If noise energy estimated value e 2Greater than reference noise threshold value e R2, that petty fiducial value C 2Selected and be input to filter selector 235.To fiducial value C i(i=1 wherein, 2 ... n) till selected, comparison procedure continues always.
Filter circuit 235 uses the fiducial value Ci that determines to determine a FILTER TO CONTROL value.Select in the question blank that this FILTER TO CONTROL value provides from Figure 12 for example.Question blank comprises a series of fiducial value Ci and corresponding FILTER TO CONTROL value Fi.Filter circuit 115 is adjusted with the form of choosing the FILTER TO CONTROL value function.Filter circuit 115 is adjusted to show that a frequency response curve is so that extract low frequency energy from present frame.When the noise energy estimated value on the continuous higher frequency surpassed their corresponding reference energy threshold, filter circuit 115 was adjusted to extract more low frequency energy.Fig. 6 A and 6B have provided the example frequency responses curve of choosing the FILTER TO CONTROL value.
The use of noise envelope estimated value helps to have improved the ability that filter circuit extracts low frequency energy of adjusting adaptively, and its mode of taking helps to improve the voice oeverall quality.Because automotive environment is not to use the unique environment of mobile radio communication device.Therefore, the noise envelope in a certain environment may tend to higher frequency.When the noise energy in the low frequency was very little, spectralyzer 270 can optionally be under an embargo.And, when a big chunk of noise frequency spectrum is positioned at low frequency,, also to use the steeper filtering gradient even that petty some processing power is sacrificed.This extra process requires to remain very little.
Apparent according to top description, the auto system of this invention is realized simply.And the calculated amount of DSP does not significantly increase.Cut down the more complicated method of noise, for example " spectrum is cut down " requires a large amount of storeies that several MIPS that relate to calculating and storage data and program code use.By relatively, this invention can sub-fraction " spectrum is cut down " the desired MIPS of algorithm and storer realize that spectrum is wherein cut down algorithm and also introduced more voice distortion simultaneously by only using.The storer that capacity reduces has reduced the size of DSP integrated circuit; The MIPS that reduces has reduced power consumption.These characteristics for battery powered portable/mobile radiotelephone all is very desirable.
With reference to its preferred embodiment, although this invention is represented especially and describe out that it is not limited only to these embodiments.For example, although DSP is described to carry out frame Energy Estimation device 210, noise estimator 230, speech detector 240, the function of filter selector 235 and filter circuit 265, these functions can realize by the numeral and/or the analog element that use other.In addition, when filter circuit 115 estimated that with Noise Estimation and noise envelope the form of the two function is adjusted, Avaptive filtering system 100 also can be implemented.

Claims (10)

1. method that is used for optionally changing a frame of digital signal, digital signal wherein is made of a plurality of successive frames, the sound signal that this digital signal representation receives at the transmitter place, this sound signal or constitute by speech components, or constitute by noise component, perhaps be made of jointly speech components and noise component, described method is characterised in that and may further comprise the steps:
The energy sizes values of estimative figure signal frame;
Determine whether comprise speech components in the digital signal frame according to the estimated value that obtains in the described estimating step;
When determining that in described determining step speech components does not constitute frame a part of, revise noise estimation value with the form of the energy sizes values function estimated in previous noise estimation value and the described estimating step;
Read a record in the question blank, this question blank has the filter characteristic according to the noise estimation value size index, and the record of reading is corresponding to the noise estimation value of revising in the described correction step;
Select to want the filter circuit filtering characteristic of filtered device demonstration, the filtering characteristic of choosing is corresponding to the filter characteristic of storage in reading to note down in the described read step;
Use the filter filtering digital data frames, wave filter wherein shows the filter circuit filtering characteristic, changes digital data frames according to the filter circuit filtering characteristic thus.
2. according to the method in the claim 1, its feature is also the intermediate steps of adding not comprise speech components if digital data frames is determined that this additional step is determined the noise envelope estimated value of digital signal frame so.
3. according to the method in the claim 2, the noise envelope estimated value that wherein is determined in described definite noise envelope estimated value step is used to revise noise estimation value in described correction step.
4. according to the method in the claim 1, wherein question blank is read in described read step, and it is characterized in that has a lot of records in the look-up table, and each record all comprises an independently filter characteristic.
5. according to the method in the claim 4, wherein the separate filter characteristic of a plurality of records comprises independently high pass filter characteristic in the question blank, and each high-pass filtering characteristic is by an independent cutoff frequency definition.
6. according to the method in the claim 4, wherein the separate filter characteristic of a plurality of records comprises independently high pass filter characteristic in the question blank, and each high-pass filtering characteristic is defined by an independent frequency response curve gradient.
7. according to the method in the claim 1, it is characterized in that further step: thus Counter Value statistics frame number of estimated energy sizes values for it in described estimating step increased.
8. according to the method in the claim 7, when every N counting value increases, just carry out the step of described selective filter circuit median filter characteristic, N be one greater than 1 integer.
9. device (100 that is used for optionally changing digital signal frame; 200), digital signal wherein is made of a plurality of successive frames, the sound signal that this digital signal representation receives at the transmitter place, this sound signal can be made of speech components, or constitute by noise component, perhaps constituting jointly by these two components, described device is characterised in that:
Be coupled the energy value estimator (210) of receiving digital signals frame identification, described energy value estimator is used for the energy value of estimative figure signal frame;
Be coupled to the speech detector (240) of described energy value estimator, described speech components determiner is used for determining whether digital signal frame comprises speech components;
Determine not exercisable noise estimator (230) during configuration frame a part of of speech components when described speech components determiner, described noise estimator is revised noise estimation value with the form of the energy value function that front noise estimation value and described estimator are estimated;
The question blank that comprises a lot of records, wherein every record comes index with noise estimation value, and the record in the question blank reads according to the noise estimation value that is formed by described noise estimator;
Be coupled the wave filter (265) of receiving digital data frame, described wave filter shows the filtering characteristic of selectable filter circuit, selection to the filter circuit filtering characteristic of wave filter is to determine according to the record of question blank, and the record of this question blank reads according to the noise estimation value by described noise estimator correction.
10. according to the device in the claim 9, its feature also is a noise envelope estimator (270), if described speech components determiner determines that digital data frames does not comprise speech components, this noise envelope estimator is determined the noise envelope estimated value of this digital data frames so.
CN96198008A 1995-09-14 1996-09-13 System for adaptively filtering audio signals to enhance speech intelligibility in noisy environmental conditions Expired - Fee Related CN1121684C (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US52800595A 1995-09-14 1995-09-14
US08/528,005 1995-09-14

Publications (2)

Publication Number Publication Date
CN1201547A CN1201547A (en) 1998-12-09
CN1121684C true CN1121684C (en) 2003-09-17

Family

ID=24103874

Family Applications (1)

Application Number Title Priority Date Filing Date
CN96198008A Expired - Fee Related CN1121684C (en) 1995-09-14 1996-09-13 System for adaptively filtering audio signals to enhance speech intelligibility in noisy environmental conditions

Country Status (15)

Country Link
EP (1) EP0852052B1 (en)
JP (1) JPH11514453A (en)
KR (1) KR100423029B1 (en)
CN (1) CN1121684C (en)
AU (1) AU724111B2 (en)
BR (1) BR9610290A (en)
CA (1) CA2231107A1 (en)
DE (1) DE69613380D1 (en)
EE (1) EE03456B1 (en)
MX (1) MX9801857A (en)
NO (1) NO981074L (en)
PL (1) PL185513B1 (en)
RU (1) RU2163032C2 (en)
TR (1) TR199800475T1 (en)
WO (1) WO1997010586A1 (en)

Families Citing this family (168)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE19747885B4 (en) 1997-10-30 2009-04-23 Harman Becker Automotive Systems Gmbh Method for reducing interference of acoustic signals by means of the adaptive filter method of spectral subtraction
JP2001508197A (en) * 1997-10-31 2001-06-19 コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィ Method and apparatus for audio reproduction of speech encoded according to the LPC principle by adding noise to a constituent signal
KR20000074236A (en) * 1999-05-19 2000-12-15 정몽규 Auto audio volume control means
US8645137B2 (en) 2000-03-16 2014-02-04 Apple Inc. Fast, language-independent method for user authentication by voice
JP2001318694A (en) * 2000-05-10 2001-11-16 Toshiba Corp Device and method for signal processing and recording medium
US6983242B1 (en) * 2000-08-21 2006-01-03 Mindspeed Technologies, Inc. Method for robust classification in speech coding
KR20030010432A (en) * 2001-07-28 2003-02-05 주식회사 엑스텔테크놀러지 Apparatus for speech recognition in noisy environment
IL148592A0 (en) 2002-03-10 2002-09-12 Ycd Multimedia Ltd Dynamic normalizing
AU2003242921A1 (en) * 2002-07-01 2004-01-19 Koninklijke Philips Electronics N.V. Stationary spectral power dependent audio enhancement system
KR100978015B1 (en) * 2002-07-01 2010-08-25 코닌클리케 필립스 일렉트로닉스 엔.브이. Stationary spectral power dependent audio enhancement system
WO2004008801A1 (en) * 2002-07-12 2004-01-22 Widex A/S Hearing aid and a method for enhancing speech intelligibility
US7242763B2 (en) 2002-11-26 2007-07-10 Lucent Technologies Inc. Systems and methods for far-end noise reduction and near-end noise compensation in a mixed time-frequency domain compander to improve signal quality in communications systems
DE10305369B4 (en) * 2003-02-10 2005-05-19 Siemens Ag User-adaptive method for noise modeling
US7127076B2 (en) 2003-03-03 2006-10-24 Phonak Ag Method for manufacturing acoustical devices and for reducing especially wind disturbances
EP1339256B1 (en) * 2003-03-03 2017-12-27 Sonova AG Method for manufacturing acoustical devices and for reducing wind disturbances
CN101048813B (en) 2004-08-30 2012-08-29 高通股份有限公司 Adaptive de-jitter buffer for voice IP transmission
KR100640865B1 (en) 2004-09-07 2006-11-02 엘지전자 주식회사 method and apparatus for enhancing quality of speech
US8085678B2 (en) 2004-10-13 2011-12-27 Qualcomm Incorporated Media (voice) playback (de-jitter) buffer adjustments based on air interface
EP1840874B1 (en) * 2005-01-11 2019-04-10 NEC Corporation Audio encoding device, audio encoding method, and audio encoding program
GB2429139B (en) 2005-08-10 2010-06-16 Zarlink Semiconductor Inc A low complexity noise reduction method
US8677377B2 (en) 2005-09-08 2014-03-18 Apple Inc. Method and apparatus for building an intelligent automated assistant
KR100667852B1 (en) * 2006-01-13 2007-01-11 삼성전자주식회사 Apparatus and method for eliminating noise in portable recorder
RU2453986C2 (en) * 2006-01-27 2012-06-20 Долби Интернэшнл Аб Efficient filtering with complex modulated filterbank
US9318108B2 (en) 2010-01-18 2016-04-19 Apple Inc. Intelligent automated assistant
KR101414233B1 (en) 2007-01-05 2014-07-02 삼성전자 주식회사 Apparatus and method for improving speech intelligibility
KR100883896B1 (en) * 2007-01-19 2009-02-17 엘지전자 주식회사 Speech intelligibility enhancement apparatus and method
KR100876794B1 (en) 2007-04-03 2009-01-09 삼성전자주식회사 Apparatus and method for enhancing intelligibility of speech in mobile terminal
US8977255B2 (en) 2007-04-03 2015-03-10 Apple Inc. Method and system for operating a multi-function portable electronic device using voice-activation
US8583426B2 (en) 2007-09-12 2013-11-12 Dolby Laboratories Licensing Corporation Speech enhancement with voice clarity
CN101904097B (en) 2007-12-20 2015-05-13 艾利森电话股份有限公司 Noise suppression method and apparatus
US20110137646A1 (en) * 2007-12-20 2011-06-09 Telefonaktiebolaget L M Ericsson Noise Suppression Method and Apparatus
US9330720B2 (en) 2008-01-03 2016-05-03 Apple Inc. Methods and apparatus for altering audio output signals
CN101221767B (en) * 2008-01-23 2012-05-30 晨星半导体股份有限公司 Voice boosting device and method used on the same
US8996376B2 (en) 2008-04-05 2015-03-31 Apple Inc. Intelligent text-to-speech conversion
KR101238731B1 (en) 2008-04-18 2013-03-06 돌비 레버러토리즈 라이쎈싱 코오포레이션 Method and apparatus for maintaining speech audibility in multi-channel audio with minimal impact on surround experience
US10496753B2 (en) 2010-01-18 2019-12-03 Apple Inc. Automatically adapting user interfaces for hands-free interaction
US20100030549A1 (en) 2008-07-31 2010-02-04 Lee Michael M Mobile device having human language translation capability with positional feedback
WO2010067118A1 (en) 2008-12-11 2010-06-17 Novauris Technologies Limited Speech recognition involving a mobile device
DE102009011583A1 (en) 2009-03-06 2010-09-09 Krones Ag Method and device for producing and filling thin-walled beverage containers
US10241644B2 (en) 2011-06-03 2019-03-26 Apple Inc. Actionable reminder entries
US10241752B2 (en) 2011-09-30 2019-03-26 Apple Inc. Interface for a virtual digital assistant
US9858925B2 (en) 2009-06-05 2018-01-02 Apple Inc. Using context information to facilitate processing of commands in a virtual assistant
US20120311585A1 (en) 2011-06-03 2012-12-06 Apple Inc. Organizing task items that represent tasks to perform
US9431006B2 (en) 2009-07-02 2016-08-30 Apple Inc. Methods and apparatuses for automatic speech recognition
US10679605B2 (en) 2010-01-18 2020-06-09 Apple Inc. Hands-free list-reading by intelligent automated assistant
US10276170B2 (en) 2010-01-18 2019-04-30 Apple Inc. Intelligent automated assistant
US10553209B2 (en) 2010-01-18 2020-02-04 Apple Inc. Systems and methods for hands-free notification summaries
US10705794B2 (en) 2010-01-18 2020-07-07 Apple Inc. Automatically adapting user interfaces for hands-free interaction
WO2011089450A2 (en) 2010-01-25 2011-07-28 Andrew Peter Nelson Jerram Apparatuses, methods and systems for a digital conversation management platform
US8682667B2 (en) 2010-02-25 2014-03-25 Apple Inc. User profiling for selecting user specific voice input processing information
CN102202038B (en) * 2010-03-24 2015-05-06 华为技术有限公司 Method and system for realizing voice energy display, conference server and terminal
US9837097B2 (en) 2010-05-24 2017-12-05 Nec Corporation Single processing method, information processing apparatus and signal processing program
CN101859569B (en) * 2010-05-27 2012-08-15 上海朗谷电子科技有限公司 Method for lowering noise of digital audio-frequency signal
US8639516B2 (en) 2010-06-04 2014-01-28 Apple Inc. User-specific noise suppression for voice quality improvements
US10762293B2 (en) 2010-12-22 2020-09-01 Apple Inc. Using parts-of-speech tagging and named entity recognition for spelling correction
CN102128976B (en) * 2011-01-07 2013-05-15 钜泉光电科技(上海)股份有限公司 Energy pulse output method and device of electric energy meter and electric energy meter
US9262612B2 (en) 2011-03-21 2016-02-16 Apple Inc. Device access using voice authentication
US10057736B2 (en) 2011-06-03 2018-08-21 Apple Inc. Active transport based notifications
US8994660B2 (en) 2011-08-29 2015-03-31 Apple Inc. Text correction processing
AU2012232977A1 (en) * 2011-09-30 2013-04-18 Apple Inc. Using context information to facilitate processing of commands in a virtual assistant
US10134385B2 (en) 2012-03-02 2018-11-20 Apple Inc. Systems and methods for name pronunciation
US9483461B2 (en) 2012-03-06 2016-11-01 Apple Inc. Handling speech synthesis of content for multiple languages
US9280610B2 (en) 2012-05-14 2016-03-08 Apple Inc. Crowd sourcing information to fulfill user requests
US9721563B2 (en) 2012-06-08 2017-08-01 Apple Inc. Name recognition system
CN102737646A (en) * 2012-06-21 2012-10-17 佛山市瀚芯电子科技有限公司 Real-time dynamic voice noise reduction method for single microphone
US9495129B2 (en) 2012-06-29 2016-11-15 Apple Inc. Device, method, and user interface for voice-activated navigation and browsing of a document
US9576574B2 (en) 2012-09-10 2017-02-21 Apple Inc. Context-sensitive handling of interruptions by intelligent digital assistant
US9547647B2 (en) 2012-09-19 2017-01-17 Apple Inc. Voice-based media searching
KR102516577B1 (en) 2013-02-07 2023-04-03 애플 인크. Voice trigger for a digital assistant
US10652394B2 (en) 2013-03-14 2020-05-12 Apple Inc. System and method for processing voicemail
US9368114B2 (en) 2013-03-14 2016-06-14 Apple Inc. Context-sensitive handling of interruptions
WO2014144949A2 (en) 2013-03-15 2014-09-18 Apple Inc. Training an at least partial voice command system
WO2014144579A1 (en) 2013-03-15 2014-09-18 Apple Inc. System and method for updating an adaptive speech recognition model
CN104095640A (en) * 2013-04-03 2014-10-15 达尔生技股份有限公司 Oxyhemoglobin saturation detecting method and device
WO2014197334A2 (en) 2013-06-07 2014-12-11 Apple Inc. System and method for user-specified pronunciation of words for speech synthesis and recognition
US9582608B2 (en) 2013-06-07 2017-02-28 Apple Inc. Unified ranking with entropy-weighted information for phrase-based semantic auto-completion
WO2014197336A1 (en) 2013-06-07 2014-12-11 Apple Inc. System and method for detecting errors in interactions with a voice-based digital assistant
WO2014197335A1 (en) 2013-06-08 2014-12-11 Apple Inc. Interpreting and acting upon commands that involve sharing information with remote devices
EP3008641A1 (en) 2013-06-09 2016-04-20 Apple Inc. Device, method, and graphical user interface for enabling conversation persistence across two or more instances of a digital assistant
US10176167B2 (en) 2013-06-09 2019-01-08 Apple Inc. System and method for inferring user intent from speech inputs
CN105265005B (en) 2013-06-13 2019-09-17 苹果公司 System and method for the urgent call initiated by voice command
EP2816557B1 (en) * 2013-06-20 2015-11-04 Harman Becker Automotive Systems GmbH Identifying spurious signals in audio signals
US9697831B2 (en) * 2013-06-26 2017-07-04 Cirrus Logic, Inc. Speech recognition
WO2015020942A1 (en) 2013-08-06 2015-02-12 Apple Inc. Auto-activating smart responses based on activities from remote devices
US9620105B2 (en) 2014-05-15 2017-04-11 Apple Inc. Analyzing audio input for efficient speech and music recognition
US10592095B2 (en) 2014-05-23 2020-03-17 Apple Inc. Instantaneous speaking of content on touch devices
US9502031B2 (en) 2014-05-27 2016-11-22 Apple Inc. Method for supporting dynamic grammars in WFST-based ASR
US9633004B2 (en) 2014-05-30 2017-04-25 Apple Inc. Better resolution when referencing to concepts
US9734193B2 (en) 2014-05-30 2017-08-15 Apple Inc. Determining domain salience ranking from ambiguous words in natural speech
US9715875B2 (en) 2014-05-30 2017-07-25 Apple Inc. Reducing the need for manual start/end-pointing and trigger phrases
US9842101B2 (en) 2014-05-30 2017-12-12 Apple Inc. Predictive conversion of language input
US9760559B2 (en) 2014-05-30 2017-09-12 Apple Inc. Predictive text input
EP3149728B1 (en) 2014-05-30 2019-01-16 Apple Inc. Multi-command single utterance input method
US10170123B2 (en) 2014-05-30 2019-01-01 Apple Inc. Intelligent assistant for home automation
US9785630B2 (en) 2014-05-30 2017-10-10 Apple Inc. Text prediction using combined word N-gram and unigram language models
US10289433B2 (en) 2014-05-30 2019-05-14 Apple Inc. Domain specific language for encoding assistant dialog
US10078631B2 (en) 2014-05-30 2018-09-18 Apple Inc. Entropy-guided text prediction using combined word and character n-gram language models
US9430463B2 (en) 2014-05-30 2016-08-30 Apple Inc. Exemplar-based natural language processing
US10659851B2 (en) 2014-06-30 2020-05-19 Apple Inc. Real-time digital assistant knowledge updates
US9338493B2 (en) 2014-06-30 2016-05-10 Apple Inc. Intelligent automated assistant for TV user interactions
EP2980801A1 (en) * 2014-07-28 2016-02-03 Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V. Method for estimating noise in an audio signal, noise estimator, audio encoder, audio decoder, and system for transmitting audio signals
US10446141B2 (en) 2014-08-28 2019-10-15 Apple Inc. Automatic speech recognition based on user feedback
US9818400B2 (en) 2014-09-11 2017-11-14 Apple Inc. Method and apparatus for discovering trending terms in speech requests
US10789041B2 (en) 2014-09-12 2020-09-29 Apple Inc. Dynamic thresholds for always listening speech trigger
US9886432B2 (en) 2014-09-30 2018-02-06 Apple Inc. Parsimonious handling of word inflection via categorical stem + suffix N-gram language models
US9646609B2 (en) 2014-09-30 2017-05-09 Apple Inc. Caching apparatus for serving phonetic pronunciations
US10074360B2 (en) 2014-09-30 2018-09-11 Apple Inc. Providing an indication of the suitability of speech recognition
US10127911B2 (en) 2014-09-30 2018-11-13 Apple Inc. Speaker identification and unsupervised speaker adaptation techniques
US9668121B2 (en) 2014-09-30 2017-05-30 Apple Inc. Social reminders
US10552013B2 (en) 2014-12-02 2020-02-04 Apple Inc. Data detection
US9711141B2 (en) 2014-12-09 2017-07-18 Apple Inc. Disambiguating heteronyms in speech synthesis
RU2589298C1 (en) * 2014-12-29 2016-07-10 Александр Юрьевич Бредихин Method of increasing legible and informative audio signals in the noise situation
US9865280B2 (en) 2015-03-06 2018-01-09 Apple Inc. Structured dictation using intelligent automated assistants
US9886953B2 (en) 2015-03-08 2018-02-06 Apple Inc. Virtual assistant activation
US9721566B2 (en) 2015-03-08 2017-08-01 Apple Inc. Competing devices responding to voice triggers
US10567477B2 (en) 2015-03-08 2020-02-18 Apple Inc. Virtual assistant continuity
US9899019B2 (en) 2015-03-18 2018-02-20 Apple Inc. Systems and methods for structured stem and suffix language models
US9842105B2 (en) 2015-04-16 2017-12-12 Apple Inc. Parsimonious continuous-space phrase representations for natural language processing
US10083688B2 (en) 2015-05-27 2018-09-25 Apple Inc. Device voice control for selecting a displayed affordance
US10127220B2 (en) 2015-06-04 2018-11-13 Apple Inc. Language identification from short strings
US10101822B2 (en) 2015-06-05 2018-10-16 Apple Inc. Language input correction
US10186254B2 (en) 2015-06-07 2019-01-22 Apple Inc. Context-based endpoint detection
US10255907B2 (en) 2015-06-07 2019-04-09 Apple Inc. Automatic accent detection using acoustic models
US11025565B2 (en) 2015-06-07 2021-06-01 Apple Inc. Personalized prediction of responses for instant messaging
US10671428B2 (en) 2015-09-08 2020-06-02 Apple Inc. Distributed personal assistant
US10747498B2 (en) 2015-09-08 2020-08-18 Apple Inc. Zero latency digital assistant
US9697820B2 (en) 2015-09-24 2017-07-04 Apple Inc. Unit-selection text-to-speech synthesis using concatenation-sensitive neural networks
US11010550B2 (en) 2015-09-29 2021-05-18 Apple Inc. Unified language modeling framework for word prediction, auto-completion and auto-correction
US10366158B2 (en) 2015-09-29 2019-07-30 Apple Inc. Efficient word encoding for recurrent neural network language models
US11587559B2 (en) 2015-09-30 2023-02-21 Apple Inc. Intelligent device identification
US10691473B2 (en) 2015-11-06 2020-06-23 Apple Inc. Intelligent automated assistant in a messaging environment
US10726859B2 (en) 2015-11-09 2020-07-28 Invisio Communication A/S Method of and system for noise suppression
US10049668B2 (en) 2015-12-02 2018-08-14 Apple Inc. Applying neural network language models to weighted finite state transducers for automatic speech recognition
US10223066B2 (en) 2015-12-23 2019-03-05 Apple Inc. Proactive assistance based on dialog communication between devices
CN105869650B (en) * 2015-12-28 2020-03-06 乐融致新电子科技(天津)有限公司 Digital audio data playing method and device
US10446143B2 (en) 2016-03-14 2019-10-15 Apple Inc. Identification of voice inputs providing credentials
CN106060717A (en) * 2016-05-26 2016-10-26 广东睿盟计算机科技有限公司 High-definition dynamic noise-reduction pickup
US9934775B2 (en) 2016-05-26 2018-04-03 Apple Inc. Unit-selection text-to-speech synthesis based on predicted concatenation parameters
US9972304B2 (en) 2016-06-03 2018-05-15 Apple Inc. Privacy preserving distributed evaluation framework for embedded personalized systems
US10249300B2 (en) 2016-06-06 2019-04-02 Apple Inc. Intelligent list reading
US10049663B2 (en) 2016-06-08 2018-08-14 Apple, Inc. Intelligent automated assistant for media exploration
DK179309B1 (en) 2016-06-09 2018-04-23 Apple Inc Intelligent automated assistant in a home environment
US10586535B2 (en) 2016-06-10 2020-03-10 Apple Inc. Intelligent digital assistant in a multi-tasking environment
US10192552B2 (en) 2016-06-10 2019-01-29 Apple Inc. Digital assistant providing whispered speech
US10509862B2 (en) 2016-06-10 2019-12-17 Apple Inc. Dynamic phrase expansion of language input
US10067938B2 (en) 2016-06-10 2018-09-04 Apple Inc. Multilingual word prediction
US10490187B2 (en) 2016-06-10 2019-11-26 Apple Inc. Digital assistant providing automated status report
DK179415B1 (en) 2016-06-11 2018-06-14 Apple Inc Intelligent device arbitration and control
DK179343B1 (en) 2016-06-11 2018-05-14 Apple Inc Intelligent task discovery
DK179049B1 (en) 2016-06-11 2017-09-18 Apple Inc Data driven natural language event detection and classification
DK201670540A1 (en) 2016-06-11 2018-01-08 Apple Inc Application integration with a digital assistant
US9748929B1 (en) * 2016-10-24 2017-08-29 Analog Devices, Inc. Envelope-dependent order-varying filter control
US10593346B2 (en) 2016-12-22 2020-03-17 Apple Inc. Rank-reduced token representation for automatic speech recognition
CN107039044B (en) * 2017-03-08 2020-04-21 Oppo广东移动通信有限公司 Voice signal processing method and mobile terminal
DK179745B1 (en) 2017-05-12 2019-05-01 Apple Inc. SYNCHRONIZATION AND TASK DELEGATION OF A DIGITAL ASSISTANT
DK201770431A1 (en) 2017-05-15 2018-12-20 Apple Inc. Optimizing dialogue policy decisions for digital assistants using implicit feedback
US10157627B1 (en) 2017-06-02 2018-12-18 Bose Corporation Dynamic spectral filtering
JP6948609B2 (en) * 2018-03-30 2021-10-13 パナソニックＩｐマネジメント株式会社 Noise reduction device
RU2680735C1 (en) * 2018-10-15 2019-02-26 Акционерное общество "Концерн "Созвездие" Method of separation of speech and pauses by analysis of the values of phases of frequency components of noise and signal
WO2020107269A1 (en) * 2018-11-28 2020-06-04 深圳市汇顶科技股份有限公司 Self-adaptive speech enhancement method, and electronic device
US11438452B1 (en) 2019-08-09 2022-09-06 Apple Inc. Propagating context information in a privacy preserving manner
US11501758B2 (en) 2019-09-27 2022-11-15 Apple Inc. Environment aware voice-assistant devices, and related systems and methods
CN111370033B (en) * 2020-03-13 2023-09-22 北京字节跳动网络技术有限公司 Keyboard sound processing method and device, terminal equipment and storage medium
US20230305590A1 (en) * 2020-03-13 2023-09-28 University Of South Australia A data processing method
CN111402916B (en) * 2020-03-24 2023-08-04 青岛罗博智慧教育技术有限公司 Voice enhancement system, method and handwriting board
CN111916106B (en) * 2020-08-17 2021-06-15 牡丹江医学院 Method for improving pronunciation quality in English teaching
CN112927715A (en) * 2021-02-26 2021-06-08 腾讯音乐娱乐科技（深圳）有限公司 Audio processing method and device and computer readable storage medium
CN114550740B (en) * 2022-04-26 2022-07-15 天津市北海通信技术有限公司 Voice definition algorithm under noise and train audio playing method and system thereof

Family Cites Families (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4461025A (en) * 1982-06-22 1984-07-17 Audiological Engineering Corporation Automatic background noise suppressor
US4630305A (en) * 1985-07-01 1986-12-16 Motorola, Inc. Automatic gain selector for a noise suppression system
US4811404A (en) * 1987-10-01 1989-03-07 Motorola, Inc. Noise suppression system
DE4012349A1 (en) * 1989-04-19 1990-10-25 Ricoh Kk Noise elimination device for speech recognition system - uses spectral subtraction of sampled noise values from sampled speech values
JP3065739B2 (en) * 1991-10-14 2000-07-17 三菱電機株式会社 Voice section detection device
US5412735A (en) * 1992-02-27 1995-05-02 Central Institute For The Deaf Adaptive noise reduction circuit for a sound reproduction system
JPH05259928A (en) * 1992-03-09 1993-10-08 Oki Electric Ind Co Ltd Method and device for canceling adaptive control noise
US5251263A (en) * 1992-05-22 1993-10-05 Andrea Electronics Corporation Adaptive noise cancellation and speech enhancement system and apparatus therefor
JPH0695693A (en) * 1992-09-09 1994-04-08 Fujitsu Ten Ltd Noise reducing circuit for voice recognition device
JP3270866B2 (en) * 1993-03-23 2002-04-02 ソニー株式会社 Noise removal method and noise removal device
US5485522A (en) * 1993-09-29 1996-01-16 Ericsson Ge Mobile Communications, Inc. System for adaptively reducing noise in speech signals
US5657422A (en) * 1994-01-28 1997-08-12 Lucent Technologies Inc. Voice activity detection driven noise remediator

Also Published As

Publication number Publication date
MX9801857A (en) 1998-11-29
AU7078496A (en) 1997-04-01
NO981074D0 (en) 1998-03-11
EE03456B1 (en) 2001-06-15
TR199800475T1 (en) 1998-06-22
WO1997010586A1 (en) 1997-03-20
BR9610290A (en) 1999-03-16
KR100423029B1 (en) 2004-07-01
KR19990044659A (en) 1999-06-25
AU724111B2 (en) 2000-09-14
EE9800068A (en) 1998-08-17
PL185513B1 (en) 2003-05-30
PL325532A1 (en) 1998-08-03
RU2163032C2 (en) 2001-02-10
CA2231107A1 (en) 1997-03-20
DE69613380D1 (en) 2001-07-19
EP0852052B1 (en) 2001-06-13
NO981074L (en) 1998-05-13
JPH11514453A (en) 1999-12-07
EP0852052A1 (en) 1998-07-08
CN1201547A (en) 1998-12-09

Similar Documents

Publication Publication Date Title
CN1121684C (en) System for adaptively filtering audio signals to enhance speech intelligibility in noisy environmental conditions
CN1188835C (en) System and method for reducing noise
CA2117587C (en) System for adaptively reducing noise in speech signals
CN1158644C (en) Reliable text conversion of voice in a radio communication system and method
CN101080766A (en) Noise reduction and comfort noise gain control using BARK band WEINER filter and linear attenuation
CN1311426C (en) Coding of stereo signals
CN1647154A (en) Coding of stereo signals
CN1228895A (en) Method and apparatus for compensating for varying D. C. offset in sampled signal
CN104981870B (en) Sound enhancing devices
CN101989430A (en) Audio mixing processing system and audio mixing processing method
CN1092866C (en) Method and arrangement in a transmission system
CN1051393C (en) Method and apparatus for reducing an undesirable characteristic of a spectral estimate of a noise signal between occurrences of voice signals
CN1121764C (en) Transmission device and system
JPH0946233A (en) Sound encoding method/device and sound decoding method/ device
CN112951251B (en) LC3 audio mixing method, device and storage medium
CN1313718A (en) Device for shaping signal, especially speech signal
CN101039305A (en) Balancing technique-based receiver and receiving method
CN1256000A (en) Method and device forr emphasizing pitch
CN112908350B (en) Audio processing method, communication device, chip and module equipment thereof
CN1809197A (en) Method of voice processing for TETRA cluster mobile telephone
CN1212555A (en) Direction transform echo canceller and method
RU2144222C1 (en) Method for compressing sound information and device which implements said method
CN1787565A (en) Speech loudspeaking device for mobile communication terminal and method thereof
CN114333885A (en) Voice noise reduction method and device and storage medium
CN1672193A (en) Speech communication unit and method for error mitigation of speech frames

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant
C19 Lapse of patent right due to non-payment of the annual fee
CF01 Termination of patent right due to non-payment of annual fee