JPH0588698A - Code drive lpc speech encoding device - Google Patents

Code drive lpc speech encoding device

Info

Publication number
JPH0588698A
JPH0588698A JP3251303A JP25130391A JPH0588698A JP H0588698 A JPH0588698 A JP H0588698A JP 3251303 A JP3251303 A JP 3251303A JP 25130391 A JP25130391 A JP 25130391A JP H0588698 A JPH0588698 A JP H0588698A
Authority
JP
Japan
Prior art keywords
code
synthesis filter
lpc
codebook
filter coefficient
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.)
Granted
Application number
JP3251303A
Other languages
Japanese (ja)
Other versions
JP3132081B2 (en
Inventor
Koichi Shibagaki
鋼一 柴垣
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
NEC Corp
Original Assignee
NEC Corp
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by NEC Corp filed Critical NEC Corp
Priority to JP03251303A priority Critical patent/JP3132081B2/en
Publication of JPH0588698A publication Critical patent/JPH0588698A/en
Application granted granted Critical
Publication of JP3132081B2 publication Critical patent/JP3132081B2/en
Anticipated expiration legal-status Critical
Expired - Fee Related legal-status Critical Current

Links

Abstract

PURPOSE:To reduce deterioration in sound quality in the case that the amount of encoded information increases. CONSTITUTION:An encoder 8 encodes the composing filter coefficient K from an LPC analyzer 1 and a code A found by a code book searcher 3 according to a speech signal X1 after high frequency components are removed at all times and outputs its result to a terminal 11. A searcher 6, on the other hand, finds a code B from a difference signal DX corresponding to the high frequency components of the speech signal X and outputs it at all times, but a controller 7 outputs the code B to the encoder 8 only when the amount of encoded information is less than a specific value. Therefore, when the amount of encoded information is less than the specific value, the code B is encoded by the encoder 8 together with the composing filter coefficient K and code A, but when the amount of encoded information exceeds the specific value, the code B is not encoded and only the composing filter coefficient K and code A are encoded in this case.

Description

【発明の詳細な説明】Detailed Description of the Invention

【0001】[0001]

【産業上の利用分野】本発明は、コード駆動LPC音声
符号化装置に関するものである。
BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a code driven LPC speech coder.

【0002】[0002]

【従来の技術】従来のコード駆動LPC(CELP)音
声符号化装置では、符号化情報に優先順位を付けていな
いため、符号化情報量が増加し、符号化情報を一部削除
しなければならなくなった場合には、無作為に符号化情
報を削除していた。
2. Description of the Related Art In a conventional code driven LPC (CELP) speech coding apparatus, since the coding information is not prioritized, the coding information amount increases and some coding information must be deleted. When it disappeared, the coding information was deleted at random.

【0003】[0003]

【発明が解決しようとする課題】従って、符号化情報量
が増加した場合には、音質の劣化が生じるという問題が
あった。
Therefore, there is a problem in that the sound quality deteriorates when the amount of encoded information increases.

【0004】本発明の目的は、このような問題を解決
し、符号化情報量が増加した場合にも音質劣化が小さい
コード駆動LPC符号化装置を提供することにある。
An object of the present invention is to solve such a problem and to provide a code driven LPC encoding device which causes little deterioration in sound quality even when the amount of encoded information increases.

【0005】[0005]

【課題を解決するための手段】本発明のコード駆動LP
C音声符号化装置は、入力音声信号をLPC分析して合
成フィルタ係数を求めるLPC分析手段と、前記入力音
声信号の高周波成分を除去するローパスフィルタ手段
と、このローパスフィルタ手段により高周波成分が除去
された音声信号と、前記LPC分析手段が求めた前記合
成フィルタ係数とに従って、所定のコードブックを探索
し、第1のコードを決定する第1のコードブック探索手
段と、このコードブック探索手段により決定された前記
第1のコードと、前記LPC分析手段が求めた前記合成
フィルタ係数とにもとづいて音声信号を合成する合成フ
ィルタ手段と、前記入力音声信号より前記合成フィルタ
手段が合成した前記音声信号を減算し、差信号を出力す
る差分手段と、この差分手段が出力する前記差信号と、
前記LPC分析手段が求めた前記合成フィルタ係数とに
従って、所定のコードブックを探索し、第2のコードを
決定する第2のコードブック探索手段と、前記LPC分
析手段が求めた前記合成フィルタ係数と、前記第1のコ
ードブック探索手段が決定した前記第1のコードと、前
記第2のコードブック探索手段が決定した前記第2のコ
ードとを符号化して出力し、前記第2のコードの符号化
結果は符号化情報量が所定量以下のときのみ出力する符
号化手段とを備えたことを特徴とする。
DISCLOSURE OF THE INVENTION A cord drive LP of the present invention
The C speech coding apparatus has an LPC analysis means for LPC-analyzing an input speech signal to obtain a synthesis filter coefficient, a low-pass filter means for removing a high-frequency component of the input speech signal, and a high-frequency component removed by the low-pass filter means. A first codebook searching means for searching a predetermined codebook and determining a first code according to the speech signal and the synthesis filter coefficient obtained by the LPC analyzing means, and the codebook searching means. Synthesis filter means for synthesizing an audio signal based on the first code and the synthesis filter coefficient obtained by the LPC analysis means, and the audio signal synthesized by the synthesis filter means from the input audio signal. A subtraction means for subtracting and outputting a difference signal, and the difference signal output by the difference means,
Second codebook searching means for searching a predetermined codebook and determining a second code according to the synthesis filter coefficient obtained by the LPC analysis means, and the synthesis filter coefficient obtained by the LPC analysis means. , The first code determined by the first codebook searching means and the second code determined by the second codebook searching means are encoded and output, and the code of the second code The encoding result is provided with an encoding means for outputting only when the encoded information amount is less than or equal to a predetermined amount.

【0006】[0006]

【実施例】次に本発明の実施例について図面を参照して
説明する。図1に本発明によるコード駆動LPC符号化
装置の一例を示す。
Embodiments of the present invention will now be described with reference to the drawings. FIG. 1 shows an example of a code driven LPC encoding device according to the present invention.

【0007】入力端子10に加えられた入力音声信号X
はLPC分析器1と、ローパスフィルタ(LPF)2
と、差分器5とに入力される。LPC分析器1は、音声
信号Xを分析し、合成フィルタ係数Kを求め、コードブ
ック探索器3,6、合成フィルタ4、ならびに符号化器
8に出力する。LPF2は音声信号Xの高周波成分を除
去し、音声信号XLをコードブック探索器3に出力す
る。コードブック探索器3は、LPC分析器1からの合
成フィルタ係数Kより合成フィルタを構成し、コードに
より合成フィルタを駆動して得られる合成音声信号が、
音声信号XLに最も近くなるようなコードを、所定のコ
ードブックを探索して求め、求めたコードAを合成フィ
ルタ4および符号化器8に出力する。
Input voice signal X applied to input terminal 10
Is an LPC analyzer 1 and a low pass filter (LPF) 2
Is input to the differentiator 5. The LPC analyzer 1 analyzes the speech signal X, obtains the synthesis filter coefficient K, and outputs it to the codebook searchers 3 and 6, the synthesis filter 4, and the encoder 8. The LPF 2 removes the high frequency component of the audio signal X and outputs the audio signal XL to the codebook searcher 3. The codebook searcher 3 constitutes a synthesis filter from the synthesis filter coefficient K from the LPC analyzer 1, and the synthesized speech signal obtained by driving the synthesis filter by the code is
A code that is closest to the audio signal XL is obtained by searching a predetermined codebook, and the obtained code A is output to the synthesis filter 4 and the encoder 8.

【0008】合成フィルタ4は、探索器3からのコード
Aにより合成フィルタを駆動して合成音声信号XCを作
成し、差分器5に出力する。差分器5は入力音声信号X
より合成フィルタ4からの合成音声信号XCを減算し、
差信号XDをコードブック探索器6に出力する。コード
ブック探索器6は、コードブック探索器3の場合と同様
に、LPC分析器1からの合成フィルタ係数Kより合成
フィルタを構成し、コードにより合成フィルタを駆動し
て得られる合成音声信号が、差信号XDに最も近くなる
ようなコードを、所定のコードブックを探索して求め、
求めたコードBをON/OFF制御器7に出力する。制
御器7は、符号化器8が出力する符号化情報の情報量が
所定量以下の場合のみ、コードBを符号化器8に供給す
る。符号化器8は、LPC分析器1からの合成フィルタ
係数K、コードブック探索器3からのコードA、ならび
に制御器7からのコードBを符号化し、出力端子11に
出力する。
The synthesizing filter 4 drives the synthesizing filter by the code A from the searcher 3 to generate a synthetic speech signal XC, and outputs it to the difference unit 5. The differencer 5 receives the input voice signal X
Subtract the synthesized speech signal XC from the synthesis filter 4,
The difference signal XD is output to the codebook searcher 6. Similar to the case of the codebook searcher 3, the codebook searcher 6 configures a synthesis filter from the synthesis filter coefficient K from the LPC analyzer 1, and a synthesized speech signal obtained by driving the synthesis filter by the code is A code that comes closest to the difference signal XD is obtained by searching a predetermined codebook,
The obtained code B is output to the ON / OFF controller 7. The controller 7 supplies the code B to the encoder 8 only when the information amount of the encoded information output by the encoder 8 is less than or equal to a predetermined amount. The encoder 8 encodes the synthesis filter coefficient K from the LPC analyzer 1, the code A from the codebook searcher 3, and the code B from the controller 7, and outputs them to the output terminal 11.

【0009】次に動作を説明する。符号化器8は符号化
情報量の多少にかかわらず、LPC分析器1からの合成
フィルタ係数K、およびコードブック探索3が、高周波
成分が除去された音声信号XLにもとづいて求めたコー
ドAを常に符号化し、端子11に出力する。
Next, the operation will be described. Regardless of the amount of encoded information, the encoder 8 outputs the code A obtained from the synthesis filter coefficient K from the LPC analyzer 1 and the codebook search 3 based on the audio signal XL from which the high frequency component has been removed. It is always encoded and output to the terminal 11.

【0010】一方、探索器6は、合成フィルタ係数K、
および音声信号Xの高周波成分に対応する差信号XDに
もとづいてコードBを求めそれを出力するが、制御器7
はそのコードBを、符号化情報量が所定量以下の場合の
み符号化器8に出力する。従って、コードBは符号化情
報量が所定量以下の場合には、合成フィルタ係数Kおよ
びコードAと共に符号化器8により符号化されるが、符
号化情報量が所定量を越える場合には、コードBは符号
化されず、その場合には合成フィルタ係数Kおよびコー
ドAのみが符号化され、出力される。
On the other hand, the searcher 6 has a synthesis filter coefficient K,
And the code B is obtained based on the difference signal XD corresponding to the high frequency component of the audio signal X and is output.
Outputs the code B to the encoder 8 only when the amount of encoded information is less than or equal to a predetermined amount. Therefore, the code B is coded by the encoder 8 together with the synthesis filter coefficient K and the code A when the coded information amount is equal to or less than the predetermined amount, but when the coded information amount exceeds the predetermined amount, Code B is not encoded, in which case only synthesis filter coefficient K and code A are encoded and output.

【0011】すなわち、本実施例のコード駆動LPC音
声符号化装置では、符号化情報量が増加して符号化情報
の一部を削除しなければならなくなった場合、音声信号
の高周波成分のみが削除されるので、音質の劣化は最小
限に抑えられる。
That is, in the code driven LPC speech coding apparatus of this embodiment, when the amount of coded information increases and a part of the coded information has to be deleted, only the high frequency component of the speech signal is deleted. As a result, deterioration of sound quality is minimized.

【0012】[0012]

【発明の効果】以上説明したように本発明のコード駆動
LPC音声符号化装置では、LPC分析手段により合成
フィルタ係数が求められ、そしてローパスフィルタ手段
と第1のコードブック探索手段とにより、合成フィルタ
係数と、高周波成分が除去された音声信号とにもとづい
て第1のコードが決定され、合成フィルタ手段、差分手
段、ならびに第2のコードブック探索手段により、合成
フィルタ係数と、入力音声信号の高周波成分に対応する
信号とにもとづいて第2のコードが決定される。そし
て、符号化情報量が所定量以下の場合には、符号化手段
は第1および第2のコードを共に符号化して出力する
が、符号化情報量が所定量を越えた場合には、符号化手
段は第2のコードの符号化結果は出力しない。すなわ
ち、本発明のコード駆動LPC音声符号化装置では、符
号化情報量が増加して符号化情報の一部を削除しなけれ
ばならなくなった場合、音声信号の高周波成分のみが削
除されることになり、音質の劣化が最小限に抑えられ
る。
As described above, in the code driven LPC speech coder of the present invention, the synthesis filter coefficient is obtained by the LPC analysis means, and the synthesis filter is obtained by the low pass filter means and the first codebook search means. The first code is determined based on the coefficient and the speech signal from which the high frequency component has been removed, and the synthesis filter means, the difference means, and the second codebook search means combine the synthesis filter coefficient and the high frequency of the input speech signal. A second code is determined based on the signal corresponding to the component. Then, when the amount of encoded information is less than or equal to the predetermined amount, the encoding means encodes and outputs both the first and second codes, but when the amount of encoded information exceeds the predetermined amount, the encoding means The encoding means does not output the encoding result of the second code. That is, in the code driven LPC speech coding apparatus of the present invention, when the amount of encoded information increases and a part of the encoded information has to be deleted, only the high frequency component of the speech signal is deleted. And the deterioration of sound quality is minimized.

【図面の簡単な説明】[Brief description of drawings]

【図1】本発明のコード駆動LPC音声符号化装置の一
例を示すブロック図である。
FIG. 1 is a block diagram showing an example of a code driven LPC speech encoding apparatus of the present invention.

【符号の説明】[Explanation of symbols]

1 LPC分析器 2 ローパスフィルタ(LPF) 3,6 コードブック探索器 4 合成フィルタ 5 差分器 7 ON/OFF制御器 8 符号化器 10 入力端子 11 出力端子 DESCRIPTION OF SYMBOLS 1 LPC analyzer 2 Low pass filter (LPF) 3,6 Codebook searcher 4 Synthesis filter 5 Difference device 7 ON / OFF controller 8 Encoder 10 Input terminal 11 Output terminal

Claims (2)

【特許請求の範囲】[Claims] 【請求項1】入力音声信号をLPC分析して合成フィル
タ係数を求めるLPC分析手段と、 前記入力音声信号の高周波成分を除去するローパスフィ
ルタ手段と、 このローパスフィルタ手段により高周波成分が除去され
た音声信号と、前記LPC分析手段が求めた前記合成フ
ィルタ係数とに従って、所定のコードブックを探索し、
第1のコードを決定する第1のコードブック探索手段
と、 このコードブック探索手段により決定された前記第1の
コードと、前記LPC分析手段が求めた前記合成フィル
タ係数とにもとづいて音声信号を合成する合成フィルタ
手段と、 前記入力音声信号より前記合成フィルタ手段が合成した
前記音声信号を減算し、差信号を出力する差分手段と、 この差分手段が出力する前記差信号と、前記LPC分析
手段が求めた前記合成フィルタ係数とに従って、所定の
コードブックを探索し、第2のコードを決定する第2の
コードブック探索手段と、 前記LPC分析手段が求めた前記合成フィルタ係数と、
前記第1のコードブック探索手段が決定した前記第1の
コードと、前記第2のコードブック探索手段が決定した
前記第2のコードとを符号化して出力し、前記第2のコ
ードの符号化結果は符号化情報量が所定量以下のときの
み出力する符号化手段とを備えたことを特徴とするコー
ド駆動LPC音声符号化装置。
1. An LPC analysis means for LPC-analyzing an input voice signal to obtain a synthesis filter coefficient, a low-pass filter means for removing a high-frequency component of the input voice signal, and a voice from which a high-frequency component is removed by the low-pass filter means. Searching a predetermined codebook according to the signal and the synthesis filter coefficient obtained by the LPC analysis means,
A voice signal is generated based on first codebook searching means for determining a first code, the first code determined by the codebook searching means, and the synthesis filter coefficient obtained by the LPC analyzing means. Synthesis filter means for synthesizing, difference means for subtracting the voice signal synthesized by the synthesis filter means from the input voice signal and outputting a difference signal, the difference signal output by the difference means, and the LPC analysis means A second codebook searching means for searching a predetermined codebook and determining a second code according to the synthesis filter coefficient obtained by; and the synthesis filter coefficient obtained by the LPC analysis means,
The first code determined by the first codebook searching means and the second code determined by the second codebook searching means are encoded and output, and the second code is encoded. A code driven LPC speech coding apparatus, comprising: a coding means for outputting the result only when the coded information amount is equal to or less than a predetermined amount.
【請求項2】前記符号化手段は、前記第2のコードブッ
ク探索手段が決定した前記第2のコードを入力とし、符
号化情報量が所定量以下のときのみ入力された前記第2
のコードを出力する制御部と、 前記LPC分析手段が求めた前記合成フィルタ係数と、
前記第1のコードブック探索手段が決定した前記第1の
コードと、前記制御手段が出力する前記第2のコードと
を符号化する符号化部とを備えたことを特徴とする請求
項1記載のコード駆動LPC音声符号化装置。
2. The encoding means receives the second code determined by the second codebook searching means as an input, and inputs the second code only when the encoded information amount is less than or equal to a predetermined amount.
A control unit for outputting the code, and the synthesis filter coefficient obtained by the LPC analysis unit,
The encoding unit that encodes the first code determined by the first codebook search means and the second code output by the control means. Code-driven LPC speech coder.
JP03251303A 1991-09-30 1991-09-30 Code-driven LPC speech encoder Expired - Fee Related JP3132081B2 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP03251303A JP3132081B2 (en) 1991-09-30 1991-09-30 Code-driven LPC speech encoder

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP03251303A JP3132081B2 (en) 1991-09-30 1991-09-30 Code-driven LPC speech encoder

Publications (2)

Publication Number Publication Date
JPH0588698A true JPH0588698A (en) 1993-04-09
JP3132081B2 JP3132081B2 (en) 2001-02-05

Family

ID=17220803

Family Applications (1)

Application Number Title Priority Date Filing Date
JP03251303A Expired - Fee Related JP3132081B2 (en) 1991-09-30 1991-09-30 Code-driven LPC speech encoder

Country Status (1)

Country Link
JP (1) JP3132081B2 (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO1994007239A1 (en) * 1992-09-16 1994-03-31 Fujitsu Limited Speech encoding method and apparatus
US5864650A (en) * 1992-09-16 1999-01-26 Fujitsu Limited Speech encoding method and apparatus using tree-structure delta code book
KR100364578B1 (en) * 2000-06-20 2002-12-12 삼성에버랜드 주식회사 Biological method of stream water-quality improvement using a natural media

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO1994007239A1 (en) * 1992-09-16 1994-03-31 Fujitsu Limited Speech encoding method and apparatus
US5864650A (en) * 1992-09-16 1999-01-26 Fujitsu Limited Speech encoding method and apparatus using tree-structure delta code book
KR100364578B1 (en) * 2000-06-20 2002-12-12 삼성에버랜드 주식회사 Biological method of stream water-quality improvement using a natural media

Also Published As

Publication number Publication date
JP3132081B2 (en) 2001-02-05

Similar Documents

Publication Publication Date Title
JP5301451B2 (en) Spectral value post-processing apparatus and method, and audio signal encoder and decoder
EP1905000B1 (en) Selectively using multiple entropy models in adaptive coding and decoding
JP3346765B2 (en) Audio decoding method and audio decoding device
EP2041745B1 (en) Adaptive encoding and decoding methods and apparatuses
EP1262956A3 (en) Signal encoding method and apparatus
JPH08263098A (en) Acoustic signal coding method, and acoustic signal decoding method
JP2002268696A (en) Sound signal encoding method, method and device for decoding, program, and recording medium
JPH07160294A (en) Sound decoder
US5933802A (en) Speech reproducing system with efficient speech-rate converter
US6061648A (en) Speech coding apparatus and speech decoding apparatus
JPH0341500A (en) Low-delay low bit-rate voice coder
JPH10282997A (en) Speech encoding device and decoding device
JP2007510170A (en) Fast codebook selection method in audio coding
JPH0588698A (en) Code drive lpc speech encoding device
KR950003557B1 (en) Encoding method of voice sample and signal sample
JP2613503B2 (en) Speech excitation signal encoding / decoding method
KR100718487B1 (en) Harmonic noise weighting in digital speech coders
JPH05500573A (en) Digital audio decoder with post filter with reduced spectral distortion
JPH05113799A (en) Code driving linear prediction coding system
JPH0519795A (en) Excitation signal encoding and decoding method for voice
JP2700974B2 (en) Audio coding method
JPH0659699A (en) Voice coding device
JP2898377B2 (en) Code-excited linear prediction encoder and decoder
JP3166697B2 (en) Audio encoding / decoding device and system
US5761633A (en) Method of encoding and decoding speech signals

Legal Events

Date Code Title Description
FPAY Renewal fee payment (event date is renewal date of database)

Free format text: PAYMENT UNTIL: 20071124

Year of fee payment: 7

FPAY Renewal fee payment (event date is renewal date of database)

Free format text: PAYMENT UNTIL: 20081124

Year of fee payment: 8

FPAY Renewal fee payment (event date is renewal date of database)

Free format text: PAYMENT UNTIL: 20081124

Year of fee payment: 8

FPAY Renewal fee payment (event date is renewal date of database)

Free format text: PAYMENT UNTIL: 20091124

Year of fee payment: 9

FPAY Renewal fee payment (event date is renewal date of database)

Free format text: PAYMENT UNTIL: 20091124

Year of fee payment: 9

FPAY Renewal fee payment (event date is renewal date of database)

Free format text: PAYMENT UNTIL: 20101124

Year of fee payment: 10

LAPS Cancellation because of no payment of annual fees