CN109147759A - A kind of shortwave voice signal diversity merging method of reseptance based on marking algorithm - Google Patents
A kind of shortwave voice signal diversity merging method of reseptance based on marking algorithm Download PDFInfo
- Publication number
- CN109147759A CN109147759A CN201811172837.9A CN201811172837A CN109147759A CN 109147759 A CN109147759 A CN 109147759A CN 201811172837 A CN201811172837 A CN 201811172837A CN 109147759 A CN109147759 A CN 109147759A
- Authority
- CN
- China
- Prior art keywords
- signal
- voice
- speech
- marking
- voice signal
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 17
- 238000001514 detection method Methods 0.000 claims description 7
- 230000006870 function Effects 0.000 claims description 2
- 230000002045 lasting effect Effects 0.000 claims description 2
- 238000013077 scoring method Methods 0.000 abstract description 10
- 230000003321 amplification Effects 0.000 abstract description 2
- 230000007547 defect Effects 0.000 abstract description 2
- 238000003199 nucleic acid amplification method Methods 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 3
- 238000004088 simulation Methods 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 238000005562 fading Methods 0.000 description 2
- 238000012074 hearing test Methods 0.000 description 2
- 230000037361 pathway Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 238000007596 consolidation process Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000007850 degeneration Effects 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012806 monitoring device Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000001303 quality assessment method Methods 0.000 description 1
- 230000035807 sensation Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L13/00—Speech synthesis; Text to speech systems
- G10L13/02—Methods for producing synthetic speech; Speech synthesisers
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/04—Segmentation; Word boundary detection
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; 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
- G10L25/48—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use
- G10L25/69—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for evaluating synthetic or decoded voice signals
Landscapes
- Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Radio Transmission System (AREA)
Abstract
A kind of shortwave voice signal diversity based on marking algorithm of the disclosure of the invention merges method of reseptance and is handled mainly for the enhanced single-side belt analogue voice signal of multipath reception, is related to a kind of shortwave voice signal diversity and merges method of reseptance.The present invention proposes a kind of diversity merging method of reseptance based on marking algorithm for defect existing for existing folding.The present invention is merged in segment of speech using the weighting after marking, uses equal gain combining in silent section.Compared with existing folding, the intelligibility of voice signal is dramatically improved.The present invention is merged in segment of speech using the weighting based on scoring method, silent section compensated for by the way of equal gain combining the possible speech of equal gain combining really become estranged noise amplification, simultaneously by way of the weighting merging based on marking result, the quality of speech segment signal is improved.
Description
Technical field
Present invention is generally directed to the enhanced single-side belt analogue voice signals of multipath reception to be handled, and be related to a kind of short
Wave voice signal diversity merges method of reseptance.
Background technique
Speech terminals detection refers to isolates voice segments and non-speech segment from the signal comprising voice, also includes in signal
The confirmation of starting point and terminating point, effective end-point detection technology can not only exclude the noise jamming of unvoiced segments, improve system
Processing real-time, and can be reduced the processing time of system, to enable the larger raising of subsequent recognition performance.The present invention adopts
It is more accurate with the end-point detecting method based on energy entropy ratio, testing result.
Marking is the objective voice evaluation of programme of reference source-free, which is suitable for the voice matter without independent reference signal
Amount prediction.For this reason that the method is proposed as the voice quality assessment in the unknown voice source of phone distal end, scene
Network monitoring and assessment.Marking can predict the voice quality based on physiological sensation, which is not only restricted to end-to-end measurement, energy
It is used for any position of chain.It should be noted that the algorithm is not the comprehensive assessment to transmission quality, voice matter is only measured
The one direction voice distortion of amount and the influence of noise, it can be studied by hearing test, and hearing test assessment is
The quality received in absolute scope range of value.Because the algorithm is to combine receiving end simulation mankind's mass perception, receive
The degeneration that end and other true monitoring devices generate should not be considered.Simultaneously because the algorithm predicts sense of hearing score, so all
Reducing the influence spoken with conversational quality cannot be considered., it is intended that loudness reduction, sidetone, delay, echo and it is other and
The damage of the two-way interaction of speech quality is not reacted in the algorithm.Accordingly, it is possible to there is very high score, but do not represent best
Quality.
Equal gain combining is also referred to as phase equalization, does not carry out amplitude weighting to each branch, and only corrects the phase of each branch
Position guarantees with addition.Although reducing estimation parameter, merge performance still influenced by evaluated error, in addition, also by
The influence of each branch's disequilibrium, when each branch's performance differs greatly, due to signal, weak branch is also amplified same multiple
It participates in merging afterwards, causes to introduce more noises, so that is obtained after merging output may not be to merge gain but merge damage
It loses.
Summary of the invention
The present invention proposes a kind of diversity merging reception based on marking algorithm for defect existing for existing folding
Method.The present invention is merged in segment of speech using the weighting after marking, uses equal gain combining in silent section.With existing folding
It compares, dramatically improves the intelligibility of voice signal.
Technical scheme is as follows:
When merging multi-path voice signal, according to voice signal end-point detection as a result, using scoring method, segment of speech is believed
It number gives a mark, and merging is weighted according to marking result；Equal gain combining is carried out to silent section.
Technical solution of the present invention are as follows: a kind of shortwave voice signal diversity merging method of reseptance based on marking algorithm, the party
Method includes:
Step 1, according to voice signal end-point detection as a result, determining the position of segment of speech and silent section；
Step 2 gives a mark to the current voice section of each road signal, wi(i=1,2 ... L)=P (x, length)
Wherein, length is to need to give a mark the length of signal x, wiMarking is indicated as a result, i indicates that the i-th road signal, L indicate
The shared road L voice signal, p () indicate scoring functions；The realization of the marking algorithm: first by input signal, according to algorithm
Calculate eight required key parameters, comprising: pitch period, the kurtosis value of linear predictor coefficient, signal-to-noise ratio, mechanic sound ginseng
Number, the suddenly lasting length of weak number, voice interruption, mute length and estimating part signal-to-noise ratio；Secondly, being closed according to this eight
Bond parameter matches corresponding six kinds of type of distortion；There are 12 characteristic parameters, different type of distortion pair in each type of distortion
The characteristic parameter answered is not exactly the same, is weighted respectively to 12 characteristic parameters, obtains a median；Finally, according to this
A median, then be weighted with other 11 characteristic parameters, marking result can be obtained；
Step 3 carries out equal gain combining to the silent section before the current voice section of each road speech；
W at this timeiIt is that 1, y (t) indicates the voice signal exported after weighting merging, xi(t) enhanced single channel voice is indicated
Signal；
Step 4 is weighted merging to the current voice section of each road signal, and give a mark result wiAs weighting coefficient；
Step 5, index, if index is not above the segment of speech number of first via signal, goes to step 2 and continue to hold from increasing
Row；
If step 6, index exceed the segment of speech number of first via signal, just to the gains conjunction such as last silent section carries out
And；
Step 7 carries out end-point detection to merging what happened latter sound again, and place is normalized to the voice signal after merging
Reason.
Beneficial effect of the present invention are as follows: the present invention is merged in segment of speech using the weighting based on scoring method, in silent section
It compensates for the possible speech of equal gain combining by the way of equal gain combining really to become estranged noise amplification, while by being based on
The weighting for result of giving a mark merges mode, improves the quality of speech segment signal.
Detailed description of the invention
Fig. 1 is scoring method flow chart of the present invention；
Fig. 2 is that the present invention is based on the diversity of scoring method to merge block diagram；
Fig. 3 is the performance comparison figure that artificial voice signal diversifying of the present invention merges front and back；
Fig. 4 is the performance comparison figure that actual voice signal diversifying of the present invention merges front and back.
Specific embodiment
With reference to the accompanying drawings and examples, technical solution of the present invention is described in detail.But it is above-mentioned that this should not be interpreted as to the present invention
The range of main body is only limitted to following embodiment, all to be all belonged to the scope of the present invention based on the technology that the content of present invention is realized.
It is present invention marking algorithm flow chart shown in Fig. 1.
Voice distortion is divided into 6 classifications by scoring method, it is according to the sizes of 8 key parameters, according to the preferential of setting
Grade realizes the judgement of type of distortion.
The type of distortion of highest priority is ambient noise, it is determined according to the signal-to-noise ratio of signal.Ambient noise can be serious
Influence voice quality, most of voice quality Mean Opinion Score (Mean Opinion Scores, MOS) containing ambient noise
Value is generally in the range of 1~3.The interruption distortion of voice signal refers to that signal has mute or interrupts, i.e. the level value of signal
It is mutated.Multiplicative noise distortion refers to noise related with signal envelope in voice signal, such distortion only occurs in activity
Phonological component.The mechanic sound of voice and the tone of voice are closely related.The minimum type of distortion of priority be voice entirety not
Naturalness, since the output quality of audio coder & decoder (codec) is gender-related, scoring method is based on fundamental frequency for the type of distortion
It is divided into two kinds of male voice, female voice situations.
Since subjective feeling of the human ear to different type voice distortion is different, scoring method is according to specific type of distortion pair
As a result the different value of mapping model parameter setting.Each type of distortion includes 12 different phonetic features, scoring method
According to the type of distortion of voice to be measured, linear combination is done to 12 features by perceiving weight accordingly and is obtained in evaluation result
Between be worth, then by this intermediate result combine 11 characteristic parameters obtain final result.
It is that the present invention is based on the diversity of scoring method to merge block diagram shown in Fig. 2.
The principle of diversity and combining techniques is exactly the same signal copy difference that will be carried on two or more pieces independent pathway
Strategy be combined, with increase receive signal instantaneous signal-to-noise ratio and average signal-to-noise ratio, improve system performance.It is based in this way
For one physical phenomenon when deep fade occurs in the signal of a paths, other independent pathways are also in the probability of deep fade very simultaneously
It is low, therefore one or more signals can be selected to merge in multiple signals, the output of receiving end thus can be improved
Signal-to-noise ratio.
Diversity and combining techniques include two aspect meanings: first is that distributed transmission, enables receiving end to obtain multiple mutual statisticals
Fading signal that is independent and carrying same information；Second is that concentrating merging treatment, multiple mutual statisticals that receiver is received
Independent fading signal is merged according to different strategies, to reduce the influence of decline.Therefore, to obtain diversity most heavy
It is " uncorrelated " between each signal that the condition wanted, which is ensuring that,.
A plurality of mutually independent tributary signal can be obtained in receiving end by diversity technique, but receiving end is with which kind of side
Formula combines multiple signals to achieve the purpose that improve output signal-to-noise ratio, and here it is folding problems to be solved.Fig. 2
L branch combinatorial construction block diagram is given, wherein wi(i=1,2 ... L) is the weighting coefficient of i-th receiving branch.If i-th
It is x that branch, which receives signal,i(t), then output end signal y (t) is represented by after merging
By choosing different weighting coefficient wi, different consolidation strategies can be formed.
Simulation comparison is carried out using the performance that Matlab simulation software merges front and back to emulation signal and actual signal diversity
Analysis, simulation result difference are as shown in Figure 3 and Figure 4.Fig. 3 illustrates 12 sections of emulation signal diversifyings of the present invention and merges front and back marking knot
The difference of fruit.As seen from the figure, the marking result of each section of voice signal is apparently higher than each section of speech before merging after the present invention merges
Signal, the speech marking result after merging improve 1 point or so；Before Fig. 4 illustrates 12 sections of actual signal diversity merging of the present invention
The difference for result of giving a mark afterwards.As seen from the figure, the marking result of each section of voice signal is slightly above each before merging after the present invention merges
Section voice signal, the speech marking result after merging improve 0.5~1 point or so.Therefore, the present invention is on the basis that speech enhances
On improve speech quality, achieved the effect that ideal.
Claims (1)
1. a kind of shortwave voice signal diversity based on marking algorithm merges method of reseptance, this method comprises:
Step 1, according to voice signal end-point detection as a result, determining the position of segment of speech and silent section；
Step 2 gives a mark to the current voice section of each road signal, wi(i=1,2 ... L)=P (x, length)
Wherein, length is to need to give a mark the length of signal x, wiMarking is indicated as a result, i indicates that the i-th road signal, L indicate shared L
Road voice signal, p () indicate scoring functions；The realization of the marking algorithm: it first by input signal, is calculated according to algorithm
Eight required key parameters, comprising: pitch period, the kurtosis value of linear predictor coefficient, signal-to-noise ratio, mechanic sound parameter, suddenly
The lasting length of the weak number in ground, voice interruption, mute length and estimating part signal-to-noise ratio；Secondly, according to this eight key parameters
Match corresponding six kinds of type of distortion；There are 12 characteristic parameters, the corresponding spy of different type of distortion in each type of distortion
It is not exactly the same to levy parameter, 12 characteristic parameters are weighted respectively, obtain a median；Finally, according among this
Value, then be weighted with other 11 characteristic parameters, marking result can be obtained；
Step 3 carries out equal gain combining to the silent section before the current voice section of each road speech；
W at this timeiIt is that 1, y (t) indicates the voice signal exported after weighting merging, xi(t) enhanced single channel voice letter is indicated
Number；
Step 4 is weighted merging to the current voice section of each road signal, and give a mark result wiAs weighting coefficient；
Step 5, index, if index is not above the segment of speech number of first via signal, goes to step 2 and continue to execute from increasing；
If step 6, index exceed the segment of speech number of first via signal, equal gain combining just is carried out to last silent section；
Step 7 carries out end-point detection to merging what happened latter sound again, and the voice signal after merging is normalized.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811172837.9A CN109147759A (en) | 2018-10-09 | 2018-10-09 | A kind of shortwave voice signal diversity merging method of reseptance based on marking algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811172837.9A CN109147759A (en) | 2018-10-09 | 2018-10-09 | A kind of shortwave voice signal diversity merging method of reseptance based on marking algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109147759A true CN109147759A (en) | 2019-01-04 |
Family
ID=64811099
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811172837.9A Pending CN109147759A (en) | 2018-10-09 | 2018-10-09 | A kind of shortwave voice signal diversity merging method of reseptance based on marking algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109147759A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110336764A (en) * | 2019-07-04 | 2019-10-15 | 电子科技大学 | A kind of blind symbol timing synchronization method of short wave channel based on diversity signal decoding feedback |
| CN110444228A (en) * | 2019-07-03 | 2019-11-12 | 深圳华海尖兵科技有限公司 | A kind of short-wave reception method and system |
| CN111341294A (en) * | 2020-02-28 | 2020-06-26 | 电子科技大学 | Method for converting text into voice with specified style |
| CN111696572A (en) * | 2019-03-13 | 2020-09-22 | 富士通株式会社 | Speech separation apparatus, method and medium |
| CN112634927A (en) * | 2020-12-03 | 2021-04-09 | 电子科技大学 | Short wave channel voice enhancement method |
| CN112634926A (en) * | 2020-11-24 | 2021-04-09 | 电子科技大学 | Short wave channel voice anti-fading auxiliary enhancement method based on convolutional neural network |
| CN114401168A (en) * | 2021-12-17 | 2022-04-26 | 郑州中科集成电路与系统应用研究院 | Voice enhancement method suitable for short-wave Morse signals in complex strong noise environment |
| CN114842864A (en) * | 2022-04-19 | 2022-08-02 | 电子科技大学 | Short wave channel signal diversity combining method based on neural network |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1237304A (en) * | 1996-09-13 | 1999-12-01 | 夸尔柯姆股份有限公司 | Conference calling system and method for wireless communications channel |
| CN105471489A (en) * | 2015-11-13 | 2016-04-06 | 中国电子科技集团公司第五十四研究所 | Realization method of selective diversity combining |
-
2018
- 2018-10-09 CN CN201811172837.9A patent/CN109147759A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1237304A (en) * | 1996-09-13 | 1999-12-01 | 夸尔柯姆股份有限公司 | Conference calling system and method for wireless communications channel |
| CN105471489A (en) * | 2015-11-13 | 2016-04-06 | 中国电子科技集团公司第五十四研究所 | Realization method of selective diversity combining |
Non-Patent Citations (3)
| Title |
|---|
| 刘青龙: "短波广域分集技术研究", 《电子技术应用》 * |
| 杨波: "基于P.563的话音质量客观评价", 《中国优秀硕士学位论文全文数据库》 * |
| 郭冬梅: "无线通信中分集合并技术的研究", 《中国博士学位论文全文数据库》 * |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111696572B (en) * | 2019-03-13 | 2023-07-18 | 富士通株式会社 | Voice separation device, method and medium |
| CN111696572A (en) * | 2019-03-13 | 2020-09-22 | 富士通株式会社 | Speech separation apparatus, method and medium |
| CN110444228B (en) * | 2019-07-03 | 2021-07-13 | 深圳华海尖兵科技有限公司 | Short wave receiving method and system |
| CN110444228A (en) * | 2019-07-03 | 2019-11-12 | 深圳华海尖兵科技有限公司 | A kind of short-wave reception method and system |
| CN110336764B (en) * | 2019-07-04 | 2021-03-30 | 电子科技大学 | Short wave channel blind symbol synchronization method based on diversity signal decoding feedback |
| CN110336764A (en) * | 2019-07-04 | 2019-10-15 | 电子科技大学 | A kind of blind symbol timing synchronization method of short wave channel based on diversity signal decoding feedback |
| CN111341294B (en) * | 2020-02-28 | 2023-04-18 | 电子科技大学 | Method for converting text into voice with specified style |
| CN111341294A (en) * | 2020-02-28 | 2020-06-26 | 电子科技大学 | Method for converting text into voice with specified style |
| CN112634926A (en) * | 2020-11-24 | 2021-04-09 | 电子科技大学 | Short wave channel voice anti-fading auxiliary enhancement method based on convolutional neural network |
| CN112634926B (en) * | 2020-11-24 | 2022-07-29 | 电子科技大学 | Short wave channel voice anti-fading auxiliary enhancement method based on convolutional neural network |
| CN112634927B (en) * | 2020-12-03 | 2022-07-29 | 电子科技大学 | Short wave channel voice enhancement method |
| CN112634927A (en) * | 2020-12-03 | 2021-04-09 | 电子科技大学 | Short wave channel voice enhancement method |
| CN114401168A (en) * | 2021-12-17 | 2022-04-26 | 郑州中科集成电路与系统应用研究院 | Voice enhancement method suitable for short-wave Morse signals in complex strong noise environment |
| CN114401168B (en) * | 2021-12-17 | 2023-11-03 | 郑州中科集成电路与系统应用研究院 | Voice enhancement method applicable to short wave Morse signal under complex strong noise environment |
| CN114842864A (en) * | 2022-04-19 | 2022-08-02 | 电子科技大学 | Short wave channel signal diversity combining method based on neural network |
| CN114842864B (en) * | 2022-04-19 | 2023-05-23 | 电子科技大学 | Short wave channel signal diversity combining method based on neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109147759A (en) | A kind of shortwave voice signal diversity merging method of reseptance based on marking algorithm | |
| CN109859767B (en) | Environment self-adaptive neural network noise reduction method, system and storage medium for digital hearing aid | |
| Jeub et al. | Model-based dereverberation preserving binaural cues | |
| JP4307557B2 (en) | Voice activity detector | |
| EP3360250B1 (en) | A sound signal processing apparatus and method for enhancing a sound signal | |
| US8041054B2 (en) | Systems and methods for selectively switching between multiple microphones | |
| JP5006279B2 (en) | Voice activity detection apparatus, mobile station, and voice activity detection method | |
| CN105261359B (en) | The noise-canceling system and noise-eliminating method of mobile microphone | |
| US8954324B2 (en) | Multiple microphone voice activity detector | |
| CN101510426B (en) | Method and system for eliminating noise | |
| KR100304666B1 (en) | Speech enhancement method | |
| CN104067339B (en) | Noise-suppressing device | |
| US20040252852A1 (en) | Hearing system beamformer | |
| CN105469785A (en) | Voice activity detection method in communication-terminal double-microphone denoising system and apparatus thereof | |
| CN103440869A (en) | Audio-reverberation inhibiting device and inhibiting method thereof | |
| CN101192411A (en) | Large distance microphone array noise cancellation method and noise cancellation system | |
| Gamper | Blind C50 estimation from single-channel speech using a convolutional neural network | |
| US9015044B2 (en) | Formant based speech reconstruction from noisy signals | |
| Fukumori et al. | Estimation of speech recognition performance in noisy and reverberant environments using PESQ score and acoustic parameters | |
| KR20220130744A (en) | Method and device for wind noise attenuation | |
| May | Robust speech dereverberation with a neural network-based post-filter that exploits multi-conditional training of binaural cues | |
| Huckvale et al. | ELO-SPHERES intelligibility prediction model for the Clarity Prediction Challenge 2022 | |
| CN110444228B (en) | Short wave receiving method and system | |
| Pfeifenberger et al. | Blind source extraction based on a direction-dependent a-priori SNR. | |
| CN114023352A (en) | Voice enhancement method and device based on energy spectrum depth modulation |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190104 |
|
| RJ01 | Rejection of invention patent application after publication |