JPH0682280B2 - Speech coding system - Google Patents
Speech coding systemInfo
- Publication number
- JPH0682280B2 JPH0682280B2 JP60169882A JP16988285A JPH0682280B2 JP H0682280 B2 JPH0682280 B2 JP H0682280B2 JP 60169882 A JP60169882 A JP 60169882A JP 16988285 A JP16988285 A JP 16988285A JP H0682280 B2 JPH0682280 B2 JP H0682280B2
- Authority
- JP
- Japan
- Prior art keywords
- pulse
- power
- white noise
- coding system
- speech coding
- 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 - Lifetime
Links
Landscapes
- Analogue/Digital Conversion (AREA)
Description
【発明の詳細な説明】 〔産業上の利用分野〕 本発明は、電話帯域の音声信号のデイジタル信号への変
換・逆変換を行なう音声符号化方式に関する。Description: TECHNICAL FIELD The present invention relates to a voice encoding method for converting / reversing a voice signal in a telephone band into a digital signal.
従来、このような音声信号の符号化にはPCM方式が使わ
れ、１チヤンネル当り64Kビツト／秒のデイジタル信号
の形で伝送されていたが、最近はデイジタル回線の効率
的運用の要望が強く、32Kビツト／秒のADPCM方式が開発
されたが、なおそれ以下、すなわち16KBビツトあるいは
さらにそれ以下の音声符号化のニーズが強い。このよう
な16Kビツトあるいはそれ以下での音声符号化法の有望
な方式としてマルチパルス符号化方式がある（小澤，荒
関他「マルチパルス駆動形音声符号化法の検討」電子通
信学会、通方研資cs82−161）。Conventionally, the PCM method was used for encoding such a voice signal, and it was transmitted in the form of a digital signal of 64 Kbit / sec per channel, but recently, there is a strong demand for efficient operation of a digital line, A 32K bit / sec ADPCM system was developed, but there is a strong need for voice coding at or below the limit, that is, at 16KB or even less. Multipulse coding is a promising method for speech coding at 16K bits or less (Ozawa, Araseki et al. "Study on multipulse-driven speech coding", IEICE, Tsukuba Lab. Resources cs82-161).
上述したマルチパルス法は、受信側に合成フイルタを持
ち、そのフイルタを複数のパルスで駆動する形式となつ
ている。しかし、そのパルスの数が有限であるため、音
声波形の単純な有声音の場合は大きいパルスが少ない数
あればよいことからさして問題とならないが、無声音の
場合は小さなパルスが数多く必要となるため具合が悪
い。つまり、伝送するパルスの数が一定であるため無声
音の劣化が耳につきやすいという欠点がある。The multi-pulse method described above has a form in which the receiving side has a composite filter and the filter is driven by a plurality of pulses. However, since the number of that pulse is finite, it is not a problem for a simple voiced sound with a small number of large pulses, but for unvoiced sound many small pulses are required. Bad condition. That is, since the number of pulses to be transmitted is constant, unvoiced sound is easily deteriorated.
このような問題点を解決するために、本発明による音声
符号化方式は、送信側にマルチパルスの電力和を求める
手段と、LPC分析による残差電力を求める手段とを備
え、両者の差信号を送出するとともに、その差信号に比
例する白色雑音を受信例でマルチパルス列に加えるよう
にしたものである。In order to solve such a problem, the speech coding method according to the present invention comprises means for obtaining the power sum of multi-pulses on the transmitting side and means for obtaining the residual power by LPC analysis, and a difference signal between them. Is transmitted, and white noise proportional to the difference signal is added to the multi-pulse train in the reception example.
送信側でマルチパルスによつて表わされなかつた不定電
力分が受信側で白色雑音として加えられる。Uncertain power that is not represented by multipulses on the transmitter side is added as white noise on the receiver side.
図は、本発明の一実施例を示すブロツク図である。同図
において、送信側１において入力端子100から入力され
た音声信号はLPC分析部10およびパルス探索部20とに入
力される。LPC分析部10により求められたLPC係数はパル
ス探索に使われると同時に残差電力検出部30に送られ
る。通常LPC分析を行なうと、残差電力はその副産物と
して得られるため、それをそのまま利用してもよい。ま
たパルス探索部20の出力は受信側２に送られると同時に
パルスレベル計算部40へ送られる。残差電力計算部30と
パルスレベル計算部40のそれぞれの出力は減算器50に送
られ、不足電力が求められる。減算器50の出力すなわち
不足電力情報は、受信側２に送られ、白色雑音発生部60
に入力され不足電力に比例した白色雑音となる。この白
色雑音とパルス組立部70の出力つまりマルチパルスとが
加算器80で加算され、送られて来たLPC係数で作られる
合成フイルタ90に入力される。合成フイルタ90の出力が
復元信号となり、出力端子200から送出される。The drawing is a block diagram showing an embodiment of the present invention. In the figure, the voice signal input from the input terminal 100 on the transmission side 1 is input to the LPC analysis unit 10 and the pulse search unit 20. The LPC coefficient obtained by the LPC analysis unit 10 is used for pulse search and at the same time sent to the residual power detection unit 30. Usually, when LPC analysis is performed, the residual power is obtained as a by-product, and it can be used as it is. The output of the pulse searching unit 20 is sent to the receiving side 2 and at the same time, sent to the pulse level calculating unit 40. The respective outputs of the residual power calculation unit 30 and the pulse level calculation unit 40 are sent to the subtractor 50, and the power shortage is obtained. The output of the subtractor 50, that is, the power shortage information is sent to the receiving side 2 and the white noise generating unit 60
Is input to and becomes white noise proportional to the power shortage. The white noise and the output of the pulse assembling unit 70, that is, the multi-pulses are added by the adder 80 and input to the synthesis filter 90 made of the LPC coefficient sent. The output of the synthesis filter 90 becomes a restoration signal and is sent out from the output terminal 200.
なお、上記LPC係数は、予測係数、PARCOR係数のいずれ
であつてもよい。The LPC coefficient may be either a prediction coefficient or a PARCOR coefficient.
このように、パルスで表わせなかつた電力（不足電力）
情報をも伝送し、受信側でそれに比例した白色雑音を発
生させて印加することにより、マルチパルス符号化方式
の宿命ともいえる、単位時間当りのパルス数が有限であ
ることによる子音入力時の音質劣化という欠点をカバー
することができる。もちろん、従来方式で問題のなかつ
た母音部については不足電力はほぼ零となるから、余計
な白色雑音を与えて悪影響を及ぼすというような懸念は
ない。In this way, power that cannot be represented by pulses (insufficient power)
Sound quality at the time of consonant input due to the finite number of pulses per unit time, which can be said to be the fate of the multi-pulse coding method, by transmitting information and generating white noise proportional to it at the receiving side and applying it. It can cover the drawback of deterioration. Of course, in the vowel part which has no problem in the conventional method, the power shortage becomes almost zero, so that there is no concern that extra white noise is given to exert an adverse effect.
以上説明したように、本発明によれば、送信例でマルチ
パルスの電力和とLPC分析による残差電力との差信号を
送出する一方、受信側ではその差に比例した白色雑音を
加えるようにしたことにより、従来マルチパルス方式の
欠点であつた無声音すなわち子音部の音質劣化の問題を
解消することができる。As described above, according to the present invention, in the transmission example, while transmitting the difference signal between the power sum of the multi-pulses and the residual power by the LPC analysis, the receiving side adds white noise proportional to the difference. By doing so, it is possible to solve the problem of unvoiced sound, that is, the deterioration of the sound quality of the consonant portion, which has been a drawback of the conventional multi-pulse method.
図は本発明の一実施例を示すブロツク図である。 １……送信側、２……受信側、10……LPC分析部、20…
…パルス探索部、30……残差電力計算部、40……パルス
電力計算部、50……減算器、60……白色雑音発生器、70
……パルス発生器、80……加算器、90……合成フイル
タ。The figure is a block diagram showing an embodiment of the present invention. 1 ... Sending side, 2 ... Reception side, 10 ... LPC analysis section, 20 ...
… Pulse search unit, 30 …… Residual power calculation unit, 40 …… Pulse power calculation unit, 50 …… Subtractor, 60 …… White noise generator, 70
...... Pulse generator, 80 …… Adder, 90 …… Synthesis filter.
Claims (1)
え、受信側にパルス発生器と合成フイルタとを備えたマ
ルチパルス符号化方式において、送信側にLPCの残差電
力を求める手段と、探索されたパルスの電力和を求める
手段と、減算器とを備え、前記残差電力とパルス電力と
の差信号を伝送するとともに、受信側に前記差信号に比
例した白色雑音を出力する手段と、この白色雑音を前記
パルス発生器の出力に加える加算器とを備え、この加算
器の出力を前記合成フイルタに入力するように構成した
ことを特徴とする音声符号化方式。1. A means for obtaining a residual power of LPC on the transmitting side in a multi-pulse coding system comprising an LPC analyzing section and a pulse searching section on the transmitting side and a pulse generator and a synthesis filter on the receiving side. And a subtracter for obtaining the sum of power of the searched pulses, transmitting a difference signal between the residual power and the pulse power, and outputting white noise proportional to the difference signal to the receiving side. A speech coding system comprising means and an adder for adding the white noise to the output of the pulse generator, and the output of the adder is input to the synthesis filter.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP60169882A JPH0682280B2 (en) | 1985-08-02 | 1985-08-02 | Speech coding system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP60169882A JPH0682280B2 (en) | 1985-08-02 | 1985-08-02 | Speech coding system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| JPS6231222A JPS6231222A (en) | 1987-02-10 |
| JPH0682280B2 true JPH0682280B2 (en) | 1994-10-19 |
Family
ID=15894699
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| JP60169882A Expired - Lifetime JPH0682280B2 (en) | 1985-08-02 | 1985-08-02 | Speech coding system |
Country Status (1)
| Country | Link |
|---|---|
| JP (1) | JPH0682280B2 (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5263119A (en) * | 1989-06-29 | 1993-11-16 | Fujitsu Limited | Gain-shape vector quantization method and apparatus |
| JPH0365822A (en) * | 1989-08-04 | 1991-03-20 | Fujitsu Ltd | Vector quantization coder and vector quantization decoder |
-
1985
- 1985-08-02 JP JP60169882A patent/JPH0682280B2/en not_active Expired - Lifetime
Also Published As
| Publication number | Publication date |
|---|---|
| JPS6231222A (en) | 1987-02-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JPH0353300A (en) | Sound encoding and decoding system | |
| EP1420389A1 (en) | Speech bandwidth extension apparatus and speech bandwidth extension method | |
| JP3446764B2 (en) | Speech synthesis system and speech synthesis server | |
| EP0477960B1 (en) | Linear prediction speech coding with high-frequency preemphasis | |
| JPH0990995A (en) | Speech coding device | |
| JP2903533B2 (en) | Audio coding method | |
| JP3137176B2 (en) | Audio coding device | |
| JP3266178B2 (en) | Audio coding device | |
| US5448680A (en) | Voice communication processing system | |
| JP2002268686A (en) | Voice coder and voice decoder | |
| JPH09319398A (en) | Signal encoder | |
| JP3092653B2 (en) | Broadband speech encoding apparatus, speech decoding apparatus, and speech encoding / decoding apparatus | |
| JP2615548B2 (en) | Highly efficient speech coding system and its device. | |
| JPH0682280B2 (en) | Speech coding system | |
| EP1154407A2 (en) | Position information encoding in a multipulse speech coder | |
| JPH052159B2 (en) | ||
| JP3299099B2 (en) | Audio coding device | |
| Ozawa et al. | Low bit rate multi-pulse speech coder with natural speech quality | |
| JPH0258100A (en) | Voice encoding and decoding method, voice encoder, and voice decoder | |
| JP3552201B2 (en) | Voice encoding method and apparatus | |
| JP2560682B2 (en) | Speech signal coding / decoding method and apparatus | |
| JP2615862B2 (en) | Voice encoding / decoding method and apparatus | |
| JP2001142499A (en) | Speech encoding device and speech decoding device | |
| Kim et al. | On a Reduction of Pitch Searching Time by Preprocessing in the CELP Vocoder | |
| JP2000029499A (en) | Voice coder and voice encoding and decoding apparatus |