JPH0290813A - Adaptive differential encoder/decoder - Google Patents
Adaptive differential encoder/decoderInfo
- Publication number
- JPH0290813A JPH0290813A JP24105988A JP24105988A JPH0290813A JP H0290813 A JPH0290813 A JP H0290813A JP 24105988 A JP24105988 A JP 24105988A JP 24105988 A JP24105988 A JP 24105988A JP H0290813 A JPH0290813 A JP H0290813A
- Authority
- JP
- Japan
- Prior art keywords
- predictor
- adaptive
- signals
- quantizer
- pole
- 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
- 230000003044 adaptive effect Effects 0.000 title claims abstract description 28
- 238000001228 spectrum Methods 0.000 abstract description 2
- 238000000034 method Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 3
- 230000006978 adaptation Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000005236 sound signal Effects 0.000 description 2
- 230000010354 integration Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000010355 oscillation Effects 0.000 description 1
Landscapes
- Compression, Expansion, Code Conversion, And Decoders (AREA)
Abstract
Description
【発明の詳細な説明】 〔産業上の利用分野〕 本発明は、適応差分PCM符号化方式(以下。[Detailed description of the invention] [Industrial application field] The present invention is based on an adaptive differential PCM encoding method (hereinafter referred to as an adaptive differential PCM encoding method).
ADPCM方式という)による符号器及び復号器に係り
、特にV29等のモデム信号に好適な符号器及び復号器
に関する。The present invention relates to an encoder and decoder based on the ADPCM system (referred to as ADPCM system), and particularly to an encoder and decoder suitable for modem signals such as V29.
ADPCM方式については、アイ・イー・イー・イー、
グローバル テレコミユニケージ■ンズコンファレンス
1984年第23・le1頁から第23・1・4貞(
IEEE GLOBAL 置ECOMUNICATIO
NSCONFERENCE 1984 、23・1・1
〜23・1・4)において論じられている様に、2次適
応極予測器と6次適応零予測器で予測器を形成し、該予
測器の予測係数を入力信号に応じて適応的に変化させる
方式がとられていた。Regarding the ADPCM method, I.E.E.
Global Telecomunications Conference 1984, 23rd, le1 pages to 23rd, 1st, 4th (
IEEE GLOBAL ECOMUNICATIO
NSCONFERENCE 1984, 23.1.1
As discussed in ~23.1.4), a second-order adaptive pole predictor and a sixth-order adaptive zero predictor form a predictor, and the prediction coefficients of the predictor are adaptively calculated according to the input signal. A method of change was adopted.
前記2次適応極予測係数α1.α2は、最大傾斜法によ
り、下記でup −to −kttされる。The second-order adaptive pole prediction coefficient α1. α2 is up-to-ktt below by the maximum slope method.
c’1(k)=(1−2−’) at (k−1)+
(3・2−8) ”)nCP(Al〕zダル(P(A−
1)]
α2 (Al= (1−2−7)α2 (k−1)+2
−’(’(InCMl z弘[P(A−2)’)−f(
al(A−1)]−pgn(P(A)15!!+z[:
P(A−1)])但し、P(At = D Q(kl
+ S gzIA)ここで、DQ(A)は逆量子化残差
信号、Sgz(A)は、零予測器出力である。また1発
振を抑える為lαz(All<0.75 、1cLlf
All<1−”;F”zfAlにその変動範囲を制限し
ている。c'1(k)=(1-2-') at (k-1)+
(3.2-8) ”)nCP(Al]zdal(P(A-
1)] α2 (Al= (1-2-7)α2 (k-1)+2
-'('(InCMl zhiro[P(A-2)')-f(
al(A-1)]-pgn(P(A)15!!+z[:
P(A-1)]) However, P(At = D Q(kl
+ S gzIA) Here, DQ(A) is the dequantized residual signal and Sgz(A) is the zero predictor output. In addition, in order to suppress one oscillation, lαz (All<0.75, 1cLlf
The range of variation is limited to All<1-";F"zfAl.
前記2次適応極予測器は、音声信号に対しては胃い予測
能力を示すが、データ間の相関の小さいV29等のモデ
ム信号に対しては、予測能力が十分でない。その原因と
して、2次極予測器の適応がモデム信号の変化(特に位
相変化)に追従できない事があげられる。この様子を第
3図に示す。Although the secondary adaptive pole predictor exhibits poor prediction ability for voice signals, its prediction ability is not sufficient for modem signals such as V29, which have small correlations between data. The reason for this is that the adaptation of the secondary pole predictor cannot follow changes (particularly phase changes) in the modem signal. This situation is shown in FIG.
本発明の目的は、入力信号の性質に応じた予測器を数種
設け、該予測器を入力信号に応じて適応的に切り替える
事によって、入力信号の性質に応じた予測を行い、良好
な符号化特性を得る事にある。An object of the present invention is to provide several types of predictors according to the characteristics of the input signal, and to adaptively switch the predictors according to the input signal to perform prediction according to the characteristics of the input signal and to obtain a good code. The goal is to obtain the characteristics of
上記目的は、音声の様なデータ間の相関の大きな信号に
対しては適応極予測器を、モデム信号の様な相関の小さ
な信号に対しては、入力信号のスペクトルにマツチした
固定極予測器を設け、ii子化器適応変数の変rJh率
より計算される予測器切替定数α(&)により、適応極
予測器出力Stαと固定極予測器出力Sgfとを入力信
号に応じて適応的に切り替える事により達成される。The above purpose is to use an adaptive pole predictor for signals with high correlation between data such as voice, and a fixed pole predictor that matches the spectrum of the input signal for signals with low correlation such as modem signals. The adaptive pole predictor output Stα and the fixed pole predictor output Sgf are adaptively adjusted according to the input signal by the predictor switching constant α(&) calculated from the variable rJh rate of the childizer adaptive variable ii. This is achieved by switching.
量子化器適応定数y(At(以下スケールファクターと
呼ぶ)は、入力信号のレベルに応じて変動する量子化器
圧伸定数である。音声信号はレベル変動が激しいのに対
して、モデム信号のレベルは、定常状態では、はとんど
変化しない。従ってスケールファクターy (Alも音
声信号入力時は変動が太き(、モデム信号入力時は変動
が小さいので、スケールファクターの変動率を監視する
事で入力信号の種類に応じて予測器を切り替える定数を
発生させる事ができる。The quantizer adaptation constant y (At (hereinafter referred to as scale factor) is a quantizer companding constant that varies depending on the level of the input signal.While audio signals have large level fluctuations, modem signals The level does not change much in a steady state.Therefore, the scale factor y (Al also has a large fluctuation when inputting an audio signal (but the fluctuation is small when inputting a modem signal, so monitor the rate of fluctuation of the scale factor. By doing this, it is possible to generate a constant that switches the predictor depending on the type of input signal.
以下1本発明の一実施例を図面を用いて説明する。第2
図は1本発明による符号器のブロック構成図を示す。該
符号器は、減算器1.適応量子化器2.ステップサイズ
更新器3.適応逆量子化器4、加算器5,10.音声用
適応極予測器6.モデム用固定極予測器7.予測器切替
制御器8.適応零予測器9から成る。An embodiment of the present invention will be described below with reference to the drawings. Second
The figure shows a block diagram of an encoder according to the present invention. The encoder includes subtractors 1. Adaptive quantizer 2. Step size updater 3. adaptive inverse quantizer 4, adders 5, 10 . Adaptive polar predictor for speech6. Fixed pole predictor for modem7. Predictor switching controller 8. It consists of an adaptive zero predictor 9.
入力信号S # (Alは1回路1で予測値3m(kl
との差分をとジ、該差分信号d (A)は、スケーリン
グファクターyfAlによって入力信号のレベルに応じ
てスケーリングされ、4ビツトの量子化器で量子化され
て。Input signal S # (Al has a predicted value of 3 m (kl
The difference signal d(A) is scaled according to the level of the input signal by a scaling factor yfAl, and quantized by a 4-bit quantizer.
童子化値I (jlが出力される。該出力を回路4で逆
量子化し1回路5でSε(A)と加算する事によって再
生信号S?−(klを得る0回路6,7は、 S?−(
klをもとにそれぞれ次入力信号に対する予測値Sgα
(Al 、 5tfik)を出力する0回路8ではスケ
ールファクターy (Alの変化率から極予測器切替定
数αp(jcl&算出し。A doji-ized value I (jl is output. The circuit 6 and 7 obtain the reproduction signal S?-(kl) by dequantizing the output in a circuit 4 and adding it to Sε(A) in a circuit 5. ?-(
The predicted value Sgα for the next input signal is calculated based on kl.
In the circuit 8 that outputs (Al, 5tfik), the scale factor y (calculates the polar predictor switching constant αp(jcl&) from the rate of change of Al.
ep(k) 、 S 1f(4、S gα(A)とから
極子測値Stバk)を算出する。From ep(k), S 1f(4, S gα(A)), the pole measurement value St(k) is calculated.
回路8は回路4の出力逆量子化残差信号D QfA)よ
り。Circuit 8 is derived from the output dequantized residual signal DQfA) of circuit 4.
零予測値3az(A)l出力する。回路10は、 S
ep(k))トS t4k)を加えて予測値Sg (A
)を算出する。The zero predicted value 3az(A)l is output. The circuit 10 is S
ep(k)) and S t4k) to obtain the predicted value Sg (A
) is calculated.
第1図に回路8のブロック構成図を示す。回路8では0
回路3の出力y(klと1サンプル前のyCA−1)と
の差分なとり、該差分’;1y(k−1)で割る事によ
って、y (/c+の変化率を計算する。回路14で該
変化率を数サンプル平均し1回路15で該平均値A、y
f&+が閾値δと比較する。回路16では該比較結果に
基づき、α(kl = (1−2−”)α(A−1)
+F(A)F(/cl = 2”′Z(A y(/c
)〉δ)0 (Ay(kl<δ)
但し、n、αは定数
なるリーク積分な行い、予測器切替定数αF/clを算
出する。回路17では該α(k)を線形結合定数として
。FIG. 1 shows a block diagram of the circuit 8. 0 in circuit 8
Calculate the rate of change of y (/c+) by taking the difference between the output y (kl and yCA-1 of one sample before) of the circuit 3, and dividing the difference by 1y (k-1). In the circuit 14, The rate of change is averaged for several samples, and the average value A, y is obtained in one circuit 15.
f&+ is compared with a threshold δ. Based on the comparison result, the circuit 16 calculates α(kl = (1-2-”) α(A-1)
+F(A)F(/cl = 2”'Z(A y(/c
)>δ)0 (Ay(kl<δ) where n and α are constants, and leakage integration is performed to calculate the predictor switching constant αF/cl. In the circuit 17, α(k) is used as a linear combination constant.
3ga(kl 、 5tf(ん)より極予剃値5p(k
)f37%Al = (1−a(kl) S ef(k
l+α(AIS ea(klにより算出する。3ga (kl, 5tf
) f37%Al = (1-a(kl) S ef(k
Calculated by l+α(AIS ea(kl).
本実施例によれば、入力信号の性質に応じた予測器を選
択する事ができるので予測残差信号が小さ(なり、良好
な符号化が行える。According to this embodiment, since a predictor can be selected according to the characteristics of the input signal, the prediction residual signal is small (ie, good encoding can be performed).
本発明によれば、入力信号に適合した予測器を選択する
事ができるので、予測残差信号を小さくし、良好な符号
化を行える。According to the present invention, it is possible to select a predictor that is suitable for the input signal, so that the prediction residual signal can be reduced and good encoding can be performed.
第1図は本発明の一実施例で第2図の予(P(1器切替
制御部のブロック構成図、第2図は符号器全体の構成図
、第3図はモデム信号入力時の2次適応極予測器の応答
を示す波形図である。
■・・・減算器、 2・・・適応量子化器。
4・・・適応逆量子化器、5.10・・・加算器。
6・・・音声用過応極予測器。
8・・・予測器切替制御部。
9・・・適応零予測器。
12・・・減算器。
14・・・平均値計算器。
16・・・リーク積分器。
11・・・遅延器。
13・・・除算器。
15・・・比較器。FIG. 1 shows an embodiment of the present invention, and FIG. It is a waveform diagram showing the response of the next adaptive pole predictor. ■... Subtractor, 2... Adaptive quantizer. 4... Adaptive inverse quantizer, 5.10... Adder. 6 ... Overstress predictor for speech. 8 ... Predictor switching control section. 9 ... Adaptive zero predictor. 12 ... Subtractor. 14 ... Average value calculator. 16 ... Leak integrator. 11... Delay unit. 13... Divider. 15... Comparator.
Claims (1)
を適応的に量子化する手段と、入力信号の性質に応じた
予測器を選択する手段と、該選択により次信号を予測す
る手段とから成る適応差分符号・復号器において、量子
化器適応定数の変動率を求める手段と、該変動率により
入力信号を識別する手段と、該識別結果に基づいて予測
器を切り替える手段を有する事を特徴とする適応差分符
号・復号器。1. Means for determining the difference between an input signal and a predicted signal, means for adaptively quantizing the difference, means for selecting a predictor according to the properties of the input signal, and means for predicting the next signal based on the selection. an adaptive differential code/decoder comprising means for determining a rate of variation of a quantizer adaptive constant, means for identifying an input signal based on the rate of variation, and means for switching a predictor based on the identification result. An adaptive differential code/decoder featuring:
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP24105988A JPH0290813A (en) | 1988-09-28 | 1988-09-28 | Adaptive differential encoder/decoder |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP24105988A JPH0290813A (en) | 1988-09-28 | 1988-09-28 | Adaptive differential encoder/decoder |
Publications (1)
Publication Number | Publication Date |
---|---|
JPH0290813A true JPH0290813A (en) | 1990-03-30 |
Family
ID=17068696
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
JP24105988A Pending JPH0290813A (en) | 1988-09-28 | 1988-09-28 | Adaptive differential encoder/decoder |
Country Status (1)
Country | Link |
---|---|
JP (1) | JPH0290813A (en) |
-
1988
- 1988-09-28 JP JP24105988A patent/JPH0290813A/en active Pending
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP3423906B2 (en) | Voice operation characteristic detection device and detection method | |
JP3297346B2 (en) | Voice detection device | |
KR870007618A (en) | ADPCM Coder and Decoder | |
EP1598811B1 (en) | Decoding apparatus and method | |
US6643618B2 (en) | Speech decoding unit and speech decoding method | |
US6317063B1 (en) | Inverse quantization using table with reduced size | |
US5687282A (en) | Method and apparatus for determining a masked threshold | |
US5802109A (en) | Speech encoding communication system | |
JP3462066B2 (en) | ADPCM compression device, ADPCM decompression device, and ADPCM compression / decompression device | |
US5696875A (en) | Method and system for compressing a speech signal using nonlinear prediction | |
JPH0290813A (en) | Adaptive differential encoder/decoder | |
WO1998037636A1 (en) | Voice encoding method | |
US6990475B2 (en) | Digital signal processing method, learning method, apparatus thereof and program storage medium | |
JP4442239B2 (en) | Voice speed conversion device and voice speed conversion method | |
US6157906A (en) | Method for detecting speech in a vocoded signal | |
JP4260928B2 (en) | Audio compression apparatus and recording medium | |
JPS62194742A (en) | Adpcm coding system | |
JPS5947903B2 (en) | Digital call audio introduction method | |
JP3747357B2 (en) | Music signal encoding apparatus and decoding apparatus | |
JP2762938B2 (en) | Audio coding device | |
JPH0252280B2 (en) | ||
JP4354038B2 (en) | Digital signal level control device and control method | |
JPH0258430A (en) | Adaptive quantizer | |
JPH0683399A (en) | Voice elimination processing system for speech encoder | |
JP4173218B2 (en) | Audio compression apparatus and recording medium |