JPH03212697A - Signal processor - Google Patents
Signal processorInfo
- Publication number
- JPH03212697A JPH03212697A JP2008593A JP859390A JPH03212697A JP H03212697 A JPH03212697 A JP H03212697A JP 2008593 A JP2008593 A JP 2008593A JP 859390 A JP859390 A JP 859390A JP H03212697 A JPH03212697 A JP H03212697A
- Authority
- JP
- Japan
- Prior art keywords
- noise
- voice
- signal
- section
- output
- 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
- 238000001514 detection method Methods 0.000 abstract description 9
- 206010002953 Aphonia Diseases 0.000 description 11
- 238000010586 diagram Methods 0.000 description 7
- 230000001629 suppression Effects 0.000 description 1
Abstract
Description
【発明の詳細な説明】
産業上の利用分野
本発明は音声処理に用いる信号処理装置に関するもので
ある。DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application The present invention relates to a signal processing device used for audio processing.
従来の技術
第３図は従来の信号処理装置のブロック図である。この
装置は図示のように音声に雑音が重畳された入力信号は
音声判別部９と、雑音予測部１０と、キャンセル部１１
とに供給される。音声判別部９ではその入力信号から音
声の有無を判別して判別結果を雑音予測部１ｏに供給す
る。雑音予測部１ｏでは音声に雑音が重畳された入力信
号と、その入力信号の音声の有無を示す音声判別部９か
ら供給される信号を用いて雑音を予測する。キャンセル
部１１では音声に雑音が重畳された入力信号から、雑音
予測部１０で得られた雑音の予測信号を除去して、音声
信号の出方を得るものである。Prior Art FIG. 3 is a block diagram of a conventional signal processing device. As shown in the figure, an input signal in which noise is superimposed on speech is sent to a speech discriminator 9, a noise predictor 10, and a canceler 11.
and will be supplied. The voice discrimination section 9 discriminates the presence or absence of voice from the input signal and supplies the discrimination result to the noise prediction section 1o. The noise prediction section 1o predicts noise using an input signal in which noise is superimposed on speech and a signal supplied from the speech discrimination section 9 indicating the presence or absence of speech in the input signal. The canceling unit 11 removes the predicted noise signal obtained by the noise predicting unit 10 from the input signal in which noise is superimposed on the voice to obtain the output of the voice signal.
発明が解決しようとする課題
しかしながら、このような従来の信号処理装置では音声
判別部９の入力に雑音が多く重畳されたときに正しい音
声の有無を判別することが困難であった。さらに音声判
別部９での音声の有無が正確に判別されなければ雑音予
測１ｏで正しい雑音成分が予測されず、キャンセル部１
１から良好な音声出力が得られないという課題がある。Problems to be Solved by the Invention However, in such a conventional signal processing device, when a lot of noise is superimposed on the input of the speech discrimination section 9, it is difficult to discriminate whether or not there is correct speech. Furthermore, if the presence or absence of voice is not accurately determined by the voice discrimination section 9, the correct noise component will not be predicted by the noise prediction section 1o, and the cancellation section 1
There is a problem that good audio output cannot be obtained from 1.
本発明は上記課題を解決するもので、入力信号に雑音成
分の多いときにも良好な雑音除去ができる信号処理装置
を提供することを目的とする。The present invention has been made to solve the above-mentioned problems, and an object of the present invention is to provide a signal processing device that can perform good noise removal even when an input signal has many noise components.
課題を解決するだめの手段
本発明は上記目的を達成するために、入力信号から音声
信号の有無を検出してその有無に対応する制御信号を出
力する音声判別部を備え、音声／雑音信号入力と前記制
御信号とにょシ雑音予測信号を出力する雑音予測部と、
前記音声／雑音信号と前記雑音予測信号が入力されて雑
音を除去した音声出力を得るキャンセル部とをもち、そ
の音声出力を前記音声判別部の入力信号とする構成とす
る。Means for Solving the Problems In order to achieve the above object, the present invention includes a voice discriminating section that detects the presence or absence of a voice signal from an input signal and outputs a control signal corresponding to the presence or absence of the voice/noise signal input. and a noise prediction unit that outputs the control signal and a noise prediction signal,
The present invention has a canceling section which receives the voice/noise signal and the noise prediction signal and obtains a voice output from which noise has been removed, and uses the voice output as an input signal to the voice discrimination section.
作 用
本発明は上記した構成により、音声／雑音信号は、キャ
ンセル部に入力され、キャンセル部にて雑音予測部で得
られた雑音予測信号をもとに雑音が除去されて音声判別
部に供給されるとともに、音声出力とされる。音声判別
部ではキャンセル部から供給される信号から音声の有無
を検出して、その結果を雑音予測部の制御信号として供
給するように動作する。したがって音声判別部には音声
／雑音信号よりあらかじめ雑音が除去されて入力される
ため、正確な音声の有無の判別が行なえるようになると
ともに、雑音の予測を正しく行なえ、良好な音声出力を
得る。According to the above-described configuration, the present invention inputs the speech/noise signal to the cancellation section, removes the noise based on the noise prediction signal obtained by the noise prediction section in the cancellation section, and supplies it to the speech discrimination section. It is also used as audio output. The voice discrimination section operates to detect the presence or absence of voice from the signal supplied from the cancellation section and supply the result as a control signal to the noise prediction section. Therefore, the noise is removed from the voice/noise signal beforehand and input to the voice discriminator, so it is possible to accurately determine the presence or absence of voice, and it is also possible to predict noise correctly and obtain good voice output. .
実施例
以下本発明の一実施例について第１図を参照しながら説
明する。EXAMPLE An example of the present invention will be described below with reference to FIG.
第１図は本発明の一実施例における信号処理装置のブロ
ック図を示すものである。図に示すように雑音予測部１
は音声／雑音入力と音声判別部３によって供給される制
御信号とにより雑音を予測してキャンセル部２に雑音予
測信号を供給する。FIG. 1 shows a block diagram of a signal processing device in an embodiment of the present invention. As shown in the figure, the noise prediction unit 1
predicts noise based on the voice/noise input and the control signal supplied by the voice discrimination section 3 and supplies a noise prediction signal to the canceling section 2 .
キャンセル部２は、音声／雑音入力から雑音予測部１よ
り供給される雑音予測信号に応じて雑音の除去を行い、
音声出力を得るとともに、音声判別部３にその音声出力
を供給する。音声判別部３はその音声出力より実際の音
声の有無を検出し、音声検出出力とするとともに、雑音
予測部１にその音声検出出力を制御信号として供給する
ような構成となっている。The cancellation unit 2 removes noise according to the noise prediction signal supplied from the noise prediction unit 1 from the voice/noise input,
It obtains a voice output and supplies the voice output to the voice discrimination section 3. The voice discrimination section 3 is configured to detect the presence or absence of actual voice from the voice output, output the voice detection output, and supply the voice detection output to the noise prediction section 1 as a control signal.
上記構成における動作を説明する。音声に雑音が重畳さ
れた音声／雑音入力はキャンセル部２に供給されて、雑
音予測部１より供給される雑音予測信号に応じて、雑音
の除去が行なわれて音声出力を得る。そしてキャンセル
部２で雑音を除去された音声／雑音入力は、音声判別部
３に供給されて音声の有無を検出して音声検出出力を得
る。ここで雑音予測部１は音声判別部３より供給される
音声の有無を示す音声検出出力を制倒信号として用いて
、音声／雑音入力信号の雑音を予測し、その雑音予測信
号をキャンセル部２に供給するように動作する。The operation in the above configuration will be explained. The speech/noise input in which noise is superimposed on the speech is supplied to the canceling section 2, where the noise is removed according to the noise prediction signal supplied from the noise prediction section 1, and a speech output is obtained. The speech/noise input from which noise has been removed by the cancellation section 2 is then supplied to the speech discrimination section 3, which detects the presence or absence of speech and obtains a speech detection output. Here, the noise prediction unit 1 predicts the noise of the voice/noise input signal using the voice detection output indicating the presence or absence of voice supplied from the voice discrimination unit 3 as a suppression signal, and transmits the noise prediction signal to the cancellation unit 2. It operates to supply.
このように本発明の実施例の信号処理装置によれば、音
声／雑音入力からあらかじめ雑音を除去した信号から音
声判別を行なうことにより、雑音に関係なく正確に音声
の有無を判別することが可能となる。またこれにより、
雑音予測も正確に行なうことができ、よシ効果的に音声
／雑音入力から雑音の除去が行なえ、クリアな音声出力
を得ることができる。As described above, according to the signal processing device of the embodiment of the present invention, by performing voice discrimination from a signal from which noise has been removed from the voice/noise input, it is possible to accurately determine the presence or absence of voice regardless of noise. becomes. Also, this allows
Noise prediction can also be performed accurately, noise can be more effectively removed from voice/noise input, and clear voice output can be obtained.
次に本発明の第２の実施例について第２図を参照しなが
ら説明する。Next, a second embodiment of the present invention will be described with reference to FIG.
第２図は本発明の他の実施例における信号処理装置のブ
ロック図である。図示のように第１のキャンセル部６は
音声／雑音入力からの第１の雑音予測部４で予測された
雑音を除去した信号を音声判別部６．第２の雑音予測部
７．および第２のキャンセル部８に供給する。音声判別
部６は第１のキャンセル部６より供給される信号の音声
の有無を検出して音声検出出力とするとともに、第１の
雑音予測部４．第２の雑音予測部７にその音声検出出力
を制御信号として供給する。第２のキャンセル部８は第
１のキャンセル部５より供給される信号から第２の雑音
予測部子で予測された雑音を除去し、音声出力とする。FIG. 2 is a block diagram of a signal processing device in another embodiment of the present invention. As shown in the figure, the first canceling unit 6 removes the noise predicted by the first noise predicting unit 4 from the voice/noise input, and converts the signal into a voice discriminating unit 6. Second noise prediction unit 7. and is supplied to the second canceling unit 8. The voice discriminator 6 detects the presence or absence of voice in the signal supplied from the first canceller 6 and outputs a voice detection output. The voice detection output is supplied to the second noise prediction unit 7 as a control signal. The second canceling unit 8 removes the noise predicted by the second noise predicting unit from the signal supplied from the first canceling unit 5, and outputs the signal as audio output.
第１の雑音予測部４と。and a first noise prediction unit 4.
第２の雑音予測部７は、ともに音声判別部６からの制御
信号を用いて、それぞれ音声／雑音入力。The second noise prediction unit 7 uses the control signal from the voice discrimination unit 6 to input voice/noise, respectively.
第１のキャンセル部５より供給される信号の雑音予測を
行う。そして第２の雑音予測部７はその結果を第２のキ
ャンセル部８に供給し、その第２のキャンセル部８はそ
の結果を音声出力とするような構成となっている。Noise prediction of the signal supplied from the first canceling unit 5 is performed. The second noise prediction section 7 then supplies the result to the second cancellation section 8, and the second cancellation section 8 is configured to output the result as audio.
上記構成における動作を説明すると、音声に雑音が重畳
された音声／雑音入力は第１のキャンセル部５に供給さ
れて、第１の雑音予測部４より供給される雑音予測信号
に応じて雑音の除去が行なわれる。そして第１のキャン
セル部６で雑音をあらかじめ除去された第１の音声出力
は第２のキャンセル部８に供給されて、第２の雑音予測
部７より供給される第２の雑音予測信号に応じてさらに
雑音の除去を正確に行い音声出力を得る。また第１のキ
ャンセル部６で雑音をあらかじめ除去された第１の音声
出力は音声判別部６に供給されて音声の有無を検出して
音声検出出力（制御信号）を得る。ここで、第１の雑音
予測部４は、音声判別部６から供給される音声の有無を
示す制御信号を用いて音声／雑音入力信号の雑音を予測
し、その第１の雑音予測信号を第１のキャンセル部６に
供給する。さらに、第２の雑音予測部７は同様に音声判
別部６より供給される音声の有無を示す制御信号を用い
て、第１のキャンセル部５によってあらかじめ雑音を除
去された第１の音声高力信号からさらに精度よく雑音を
予測し、その第１の雑音予測信号を第２のキャンセル部
８に供給するように動作する。このように本発明の第２
の実施例の信号処理装置によれば、雑音に関係なく正確
に音声の有無を判別できるとともに、あらかじめ雑音を
除去した第１の音声出力からさらに雑音を正確に予測し
て除去することで、−初伝レベルでかつ急激な変動を有
する非定常的な雑音の除去を行なうことが可能となる。To explain the operation of the above configuration, the voice/noise input in which noise is superimposed on the voice is supplied to the first canceling section 5, and the noise is reduced according to the noise prediction signal supplied from the first noise predicting section 4. Removal takes place. The first audio output from which noise has been removed in advance by the first canceling unit 6 is supplied to the second canceling unit 8, and is responsive to the second noise prediction signal supplied from the second noise predicting unit 7. Then, noise is removed accurately and audio output is obtained. Further, the first audio output from which noise has been removed in advance by the first canceling unit 6 is supplied to the audio determining unit 6 to detect the presence or absence of audio to obtain a audio detection output (control signal). Here, the first noise prediction unit 4 predicts the noise of the voice/noise input signal using the control signal indicating the presence or absence of voice supplied from the voice discrimination unit 6, and converts the first noise prediction signal into the first noise prediction signal. The signal is supplied to the cancellation unit 6 of No. 1. Further, the second noise prediction unit 7 similarly uses a control signal indicating the presence or absence of a voice supplied from the voice discrimination unit 6 to detect the first voice high intensity from which noise has been removed in advance by the first cancellation unit 5. It operates to more accurately predict noise from the signal and supply the first noise prediction signal to the second canceling unit 8. In this way, the second aspect of the present invention
According to the signal processing device of the embodiment, it is possible to accurately determine the presence or absence of voice regardless of noise, and by further accurately predicting and removing noise from the first voice output from which noise has been removed, - It becomes possible to remove non-stationary noise that has rapid fluctuations at the initial level.
発明の効果
以上の実施例から明らかなように、本発明によれば、音
声に雑音が重畳された入力信号をキャンセル部であらか
じめ雑音の除去した信号を用いて音声判別部で音声の有
無の判別を行なうように構成しているので、音声に重畳
された雑音の大きさに関係なく、音声の有無を正確に判
別することが可能になるという効果を得ることができる
。さらに音声の有無を正確にすることで雑音の予測能力
が高まり、正確に雑音の除去を行なえて、クリアな音声
出力を得ることができるという効果が得られる。Effects of the Invention As is clear from the above-described embodiments, according to the present invention, the presence or absence of speech is determined in the speech discrimination section using a signal from which noise has been removed in advance from an input signal in which noise is superimposed on speech in the cancellation section. Since the present invention is configured to perform this, it is possible to obtain the effect that the presence or absence of voice can be accurately determined regardless of the magnitude of the noise superimposed on the voice. Furthermore, by accurately detecting the presence or absence of voice, the ability to predict noise increases, noise can be removed accurately, and clear voice output can be obtained.
さらに、あらかじめ雑音除去したキャンセル部の音声出
力を入力とする、第２のキャンセル部と第２の雑音予測
部を備え、その第２のキャンセル部によってさらに雑音
除去した信号を音声出力とする構成にすることによって
、−層低レベルでかつ急激な変動を有する非定常的な雑
音の除去を可能にするという効果がある。Furthermore, the configuration includes a second canceling section and a second noise predicting section which receive as input the audio output of the canceling section from which noise has been removed in advance, and outputs a signal from which the noise has been further removed by the second canceling section. This has the effect of making it possible to remove non-stationary noise that is low level and has rapid fluctuations.
第１図は本発明の第１の実施例の信号処理装置のブロッ
ク図、第２図は本発明の第２の実施例の信号処理装置の
ブロック図、第３図は従来の信号処理装置のブロック図
である。
１・・・・・・雑音予測部、２・・・・・・キャンセル
部、３・・・・・・音声判別部。FIG. 1 is a block diagram of a signal processing device according to a first embodiment of the present invention, FIG. 2 is a block diagram of a signal processing device according to a second embodiment of the present invention, and FIG. 3 is a block diagram of a conventional signal processing device. It is a block diagram. 1...Noise prediction unit, 2...Cancellation unit, 3...Speech discrimination unit.
Claims (2)
に対応する制御信号を出力する音声判別部を備え、音声
／雑音信号入力と前記制御信号とにより雑音予測信号を
出力する雑音予測部と、前記音声／雑音信号と前記雑音
予測信号が入力されて雑音を除去した音声出力を得るキ
ャンセル部をもち、その音声出力を前記音声判別部の入
力信号とした信号処理装置。(1) A noise prediction unit that includes a voice discrimination unit that detects the presence or absence of a voice signal from an input signal and outputs a control signal corresponding to the presence or absence, and outputs a noise prediction signal based on the voice/noise signal input and the control signal. and a canceling unit that receives the voice/noise signal and the noise prediction signal and obtains a noise-free voice output, and uses the voice output as an input signal of the voice discrimination unit.
に対応する制御信号を出力する音声判別部を備え、音声
／雑音信号入力と前記制御信号とにより第１の雑音予測
信号を出力する第１の雑音予測部と、前記音声／雑音信
号と前記第１の雑音予測信号が入力されて雑音を除去し
た第１の音声出力を得る第１のキャンセル部と、その第
１の音声出力を前記音声判別部の入力信号とする信号処
理装置に、前記第１の音声出力と前記制御信号とにより
第２の雑音予測信号を出力する第２の雑音予測部とを設
け、前記第１の音声出力と前記第２の雑音予測信号が入
力されて雑音を除去した音声出力を得るようにした信号
処理装置。(2) A voice discrimination unit that detects the presence or absence of a voice signal from an input signal and outputs a control signal corresponding to the presence or absence of the voice signal, and outputs a first noise prediction signal based on the voice/noise signal input and the control signal. a first noise prediction unit; a first cancellation unit that receives the voice/noise signal and the first noise prediction signal and obtains a first voice output from which noise is removed; A second noise prediction unit that outputs a second noise prediction signal based on the first voice output and the control signal is provided in a signal processing device that is used as an input signal of the voice discrimination unit, and the second noise prediction unit outputs a second noise prediction signal using the first voice output and the control signal. A signal processing device that receives the output and the second noise prediction signal and obtains a noise-free audio output.
Priority Applications (14)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2008593A JPH03212697A (en) | 1990-01-18 | 1990-01-18 | Signal processor |
| AU68687/91A AU633673B2 (en) | 1990-01-18 | 1991-01-07 | Signal processing device |
| KR1019910000600A KR950011964B1 (en) | 1990-01-18 | 1991-01-16 | Signal processing device |
| CA002034354A CA2034354C (en) | 1990-01-18 | 1991-01-17 | Signal processing device |
| EP94108160A EP0637012B1 (en) | 1990-01-18 | 1991-01-18 | Signal processing device |
| DE69105760T DE69105760T2 (en) | 1990-01-18 | 1991-01-18 | Device for signal processing. |
| EP91100591A EP0438174B1 (en) | 1990-01-18 | 1991-01-18 | Signal processing device |
| NO910220A NO306800B1 (en) | 1990-01-18 | 1991-01-18 | The signal processing device |
| FI910292A FI104663B (en) | 1990-01-18 | 1991-01-18 | signal Processing device |
| DE69131883T DE69131883T2 (en) | 1990-01-18 | 1991-01-18 | Noise reduction device |
| US08/095,179 US6038532A (en) | 1990-01-18 | 1993-07-23 | Signal processing device for cancelling noise in a signal |
| HK184895A HK184895A (en) | 1990-01-18 | 1995-12-07 | Signal processing device |
| HK98110578A HK1010009A1 (en) | 1990-01-18 | 1998-09-10 | Signal processing device |
| FI990865A FI118022B (en) | 1990-01-18 | 1999-04-16 | Signal processor for noise suppression system - has canceller for subtracting protected noise from frequency analyser signal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2008593A JPH03212697A (en) | 1990-01-18 | 1990-01-18 | Signal processor |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| JPH03212697A true JPH03212697A (en) | 1991-09-18 |
Family
ID=11697281
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| JP2008593A Pending JPH03212697A (en) | 1990-01-18 | 1990-01-18 | Signal processor |
Country Status (1)
| Country | Link |
|---|---|
| JP (1) | JPH03212697A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2012176932A1 (en) * | 2011-06-24 | 2012-12-27 | 日本電気株式会社 | Speech processing device, speech processing method, and speech processing program |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPS59124397A (en) * | 1982-12-29 | 1984-07-18 | 松下電器産業株式会社 | Non-voice section detecting circuit |
| JPS60254100A (en) * | 1984-05-30 | 1985-12-14 | 沖電気工業株式会社 | Voice recognition system |
-
1990
- 1990-01-18 JP JP2008593A patent/JPH03212697A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPS59124397A (en) * | 1982-12-29 | 1984-07-18 | 松下電器産業株式会社 | Non-voice section detecting circuit |
| JPS60254100A (en) * | 1984-05-30 | 1985-12-14 | 沖電気工業株式会社 | Voice recognition system |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2012176932A1 (en) * | 2011-06-24 | 2012-12-27 | 日本電気株式会社 | Speech processing device, speech processing method, and speech processing program |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP4236726B2 (en) | Voice activity detection method and voice activity detection apparatus | |
| US6266398B1 (en) | Method and apparatus for facilitating speech barge-in in connection with voice recognition systems | |
| US6449361B1 (en) | Control method and device for echo canceller | |
| WO2005094397A3 (en) | Tone event detector and method therefor | |
| JP4438720B2 (en) | Echo canceller and microphone device | |
| EP3354004B1 (en) | Acoustic echo path change detection apparatus and method | |
| JPH03212697A (en) | Signal processor | |
| CA2112767A1 (en) | Tone-Detecting Method for Detecting at Least one Tone in a Dual-Tone Multifrequency Signal, Call-Progress Method Using the Same | |
| KR960041026A (en) | Method and apparatus for detecting waiting passengers in an elevator | |
| WO2005036915A3 (en) | Cell selection techniques for frequency division multiple access systems | |
| JPH07226779A (en) | Frequency detection circuit | |
| JPH04227338A (en) | Voice signal processing unit | |
| JP2000295641A (en) | Tone signal detection method and tone signal detector | |
| JPH06101666B2 (en) | Adaptive noise eliminator | |
| JP2980141B2 (en) | Center clipper circuit | |
| US5668858A (en) | Control device for a facsimile apparatus | |
| CN112822593B (en) | Adaptive noise reduction control method, adaptive noise reduction control device and earphone | |
| KR0176884B1 (en) | Optical disk discrimination method | |
| KR970017174A (en) | Speech recognition system and echo recognition method using echo cancellation | |
| JPH07175498A (en) | Device for recognizing and responding voice | |
| JPH07131828A (en) | Video signal monitor device | |
| JPH0556178A (en) | Communication information processor | |
| KR970048588A (en) | Input circuit of digital overcurrent relay | |
| JPH07104781A (en) | Method and device for deciding voice | |
| KR19980056044A (en) | A method for removing back noise of a television signal and a circuit for performing the same |