CN85100180B - Recognition method of chinese sound using computer - Google Patents
Recognition method of chinese sound using computer Download PDFInfo
- Publication number
- CN85100180B CN85100180B CN85100180A CN85100180A CN85100180B CN 85100180 B CN85100180 B CN 85100180B CN 85100180 A CN85100180 A CN 85100180A CN 85100180 A CN85100180 A CN 85100180A CN 85100180 B CN85100180 B CN 85100180B
- Authority
- CN
- China
- Prior art keywords
- voiced sound
- speech
- computing machine
- sampling
- chinese speech
- 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
Links
Images
Abstract
The present invention relates to a method for computers to recognize Chinese speech, which belongs to the field of speech recognition. Computers are used for recognizing Chinese speech. Speech recognition devices are not independent of specific users by the method without the limitation of vocabulary. The present invention recognizes Chinese speech according to phonemes, syllables and tones. The fundamental frequency of voiced sound is extracted, frequency multiplication pulses are generated by a phase-locked loop, and then analysis, characteristic extraction and recognition are carried out to a signal sequence for sampling by a synchronous sampling technology. The method can be applied to man-machine systems which use natural Chinese speech as input.
Description
The invention belongs to field of speech recognition, utilize computing machine that Chinese speech is discerned.
General speech analysis and recognition methods now all is that voice signal is sampled by the mode of even time interval, divides frame by the time, and the time ordinal series of every frame is asked for feature, discerns then.This recognition methods depends critically upon intonation and speech speed, and therefore the recognition device of making in this way depends on specific people, and the vocabulary of its identification also is very limited.The second phase in 1985 " international electronics newspaper ", the listed various speech recognition plug-in cards that now put goods on the market promptly belonged to this example.
The present invention proposes a kind of recognition device that can not rely on specific end user and be not subjected to concrete vocabulary restriction.This apparatus features takes into full account characteristics and people's the sounding and the mechanism of the sense of hearing of Chinese speech to the processing of voice signal, analysis and identification the time.The present invention will discern by phoneme, syllable and tone Chinese speech.For the voice signal that is sent by vocal cord vibration, the present invention proposes to adopt the technology of extracting fundamental frequency and synchronized sampling, then the burst of sampling is analyzed, and extracts phonetic feature, discerns.
Chinese speech is monosyllabic, and each syllable is formed to several phonemes by one.The quantity of syllable and phoneme all is limited.Voiced sound phoneme by the vocal cord vibration pronunciation in the four tones of standard Chinese pronunciation intonation of Chinese and the syllable occupies an important position.To take into full account these characteristics of Chinese in the present invention, the voiced sound signal has the characteristic of cycle or quasi-periodic signal, and its fundamental frequency changes when intonation changes.If adopt the Sampling techniques of even time interval, then data volume is very big and introduce information fuzzy such as leakage errors inevitably.The used synchronous sampling technique quantity of information of compress voiced significantly among the present invention, it can also provide the feature of intonation and the variation of intonation fully.
The present invention can develop into the input of the Chinese speech of usefulness nature as the person machine system.Recognition methods among the present invention can be widely used in various fields, for example various semiautomatic plants of term sound control system and work mechanism; Term sound control false making limb, nursing machinery; With voice computing machine is carried out program composition; Sound-controlled typewriter; Secret device that discriminates one's identification with voice etc.
The speech recognition equipment block diagram that the present invention proposes as shown in Figure 1.A is voice, and it is detected by microphone (1), changes electric signal into.Then by a prime amplifier (2).Voice telecommunication after the amplification number is by a low-pass filter (3), and the fundamental frequency of voiced sound can be searched for and follow the tracks of to the cutoff frequency of this wave filter automatically.(4) be judgment means, then, voiced sound fundamental frequency C triggered a phaselocking frequency multiplier (5), obtain 64 frequencys multiplication for example) with the sampling pulse sequence d(of voiced sound fundamental frequency frequency multiplication to the voiced sound fundamental frequency.(6) be a frequency divider, it provides feedback for the phase-locked loop.Voice telecommunication b simultaneously again by a frequency overlapped-resistable filter (7), uses A/D converter (8) to carry out synchronized sampling then, for voiced sound, samples in the mode of external trigger with the double frequency pulse sequence d of its fundamental frequency.For voiceless sound, then still sample with time clock.The information of the sample sequence of voice telecommunication number and fundamental frequency all delivered in the computing machine (9) analyze, extract feature and also discern.(10) are to use the mode transfer plate among Fig. 1, and (11) are phoneme and syllable template, and template all presets, and the output e of computing machine is the identification to phoneme, syllable; F is the identification to the intonation four tones of standard Chinese pronunciation, and g is the identification to speaker characteristic.
Claims (1)
1, a kind of device that utilizes computing machine that Chinese speech is discerned, comprise low-pass filter (3), A/D converter (8), computing machine (9) and some tone templates (10), phoneme syllable template (11) etc., it is characterized in that utilizing wave filter to extract the fundamental frequency of voiced sound, trigger a frequency multiplication of phase locked loop device (5) to obtain the sampling pulse sequence of a frequency multiplication, this sampling pulse sequence trigger A/D converter (8) carries out synchronized sampling to the voiced sound signal, send into computing machine through the voiced sound signal of sampling and discern, meanwhile the fundamental frequency information of voiced sound is also sent into the identification four tones of standard Chinese pronunciation of computing machine.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN85100180A CN85100180B (en) | 1985-04-01 | 1985-04-01 | Recognition method of chinese sound using computer |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN85100180A CN85100180B (en) | 1985-04-01 | 1985-04-01 | Recognition method of chinese sound using computer |
Publications (2)
Publication Number | Publication Date |
---|---|
CN85100180A CN85100180A (en) | 1986-10-01 |
CN85100180B true CN85100180B (en) | 1987-05-13 |
Family
ID=4790952
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN85100180A Expired CN85100180B (en) | 1985-04-01 | 1985-04-01 | Recognition method of chinese sound using computer |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN85100180B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1835075B (en) * | 2006-04-07 | 2011-06-29 | 安徽中科大讯飞信息科技有限公司 | Speech synthetizing method combined natural sample selection and acaustic parameter to build mould |
JP4946293B2 (en) * | 2006-09-13 | 2012-06-06 | 富士通株式会社 | Speech enhancement device, speech enhancement program, and speech enhancement method |
-
1985
- 1985-04-01 CN CN85100180A patent/CN85100180B/en not_active Expired
Also Published As
Publication number | Publication date |
---|---|
CN85100180A (en) | 1986-10-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Daubechies et al. | A nonlinear squeezing of the continuous wavelet transform based on auditory nerve models | |
US4181813A (en) | System and method for speech recognition | |
US4284846A (en) | System and method for sound recognition | |
US3770892A (en) | Connected word recognition system | |
US4783807A (en) | System and method for sound recognition with feature selection synchronized to voice pitch | |
JPS58130393A (en) | Voice recognition equipment | |
JPS6466698A (en) | Voice recognition equipment | |
US20220238118A1 (en) | Apparatus for processing an audio signal for the generation of a multimedia file with speech transcription | |
US4707857A (en) | Voice command recognition system having compact significant feature data | |
EP0472578B1 (en) | Apparatus and methods for the generation of stabilised images from waveforms | |
WO1983002190A1 (en) | A system and method for recognizing speech | |
CN85100180B (en) | Recognition method of chinese sound using computer | |
Nagaraja et al. | Mono and cross lingual speaker identification with the constraint of limited data | |
JP2580768B2 (en) | Voice recognition device | |
Wolf | Speech signal processing and feature extraction | |
Dersch | A decision logic for speech recognition | |
JPS60149098A (en) | Voice input unit | |
JPH0731508B2 (en) | Speech recognition response device | |
WO1987003127A1 (en) | System and method for sound recognition with feature selection synchronized to voice pitch | |
Barger et al. | Evaluation of discrete transforms for use in digital speech recognition | |
Chen et al. | Implementation of speech recognition system based on VC++ | |
Ashton | Speech recognition with the Apple macintosh | |
JPS61249099A (en) | Voice recognition equipment | |
Lienard | An over-view of speech synthesis | |
JPS62299899A (en) | Contracted sound-direct sound speech evaluation system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
PB01 | Publication | ||
C06 | Publication | ||
C13 | Decision | ||
GR02 | Examined patent application | ||
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 |