JPS6126080B2 - - Google Patents
- ️Wed Jun 18 1986
Info
-
Publication number
- JPS6126080B2 JPS6126080B2 JP54128516A JP12851679A JPS6126080B2 JP S6126080 B2 JPS6126080 B2 JP S6126080B2 JP 54128516 A JP54128516 A JP 54128516A JP 12851679 A JP12851679 A JP 12851679A JP S6126080 B2 JPS6126080 B2 JP S6126080B2 Authority
- JP
- Japan Prior art keywords
- audio signal
- reference pattern
- pattern memory
- voice
- audio Prior art date
- 1979-10-04 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
- 230000005236 sound signal Effects 0.000 claims description 33
- 238000012360 testing method Methods 0.000 claims description 21
- 238000012935 Averaging Methods 0.000 claims description 7
- 238000000034 method Methods 0.000 claims description 7
- 238000006243 chemical reaction Methods 0.000 claims description 6
- 238000001514 detection method Methods 0.000 claims description 6
- 238000003672 processing method Methods 0.000 claims 1
- 230000006835 compression Effects 0.000 description 7
- 238000007906 compression Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 2
- 238000013102 re-test Methods 0.000 description 1
Landscapes
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Description
【発明の詳細な説明】
本発明は音声機器に於ける音声信号処理方法に
関し、特に音声信号の登録に付随する登録音声信
号のテスト方法に特徴を有する。DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a method for processing audio signals in audio equipment, and is particularly characterized by a method for testing registered audio signals accompanying registration of audio signals.
音声に依つてデータを入力したりする音声機器
が出現している。この種の機器の場合、音声を認
識する際に用いる参照パターンを予め登録してお
き、その参照パターンと音声信号とを比較して音
声の認識を行つている。従つて登録する参照パタ
ーンの正確さはその後の認識時の認識率に大きな
影響を与えるので、通常数回テストを繰り返して
参照パターンの正確さを期している。 2. Description of the Related Art Voice devices that input data by voice have appeared. In the case of this type of device, a reference pattern used when recognizing speech is registered in advance, and speech recognition is performed by comparing the reference pattern and the speech signal. Therefore, since the accuracy of the registered reference pattern has a great influence on the recognition rate during subsequent recognition, tests are usually repeated several times to ensure the accuracy of the reference pattern.
本発明はこの登録音声信号のテストを効率良く
行わしめる方法を提供するもので、以下に詳述す
る。図は本発明方法を実施する際の構成を示した
ブロツク図であつて、1は音声を音声信号に変換
するマイクロフオン、2は該マイクロフオンから
の音声信号をデイジタル信号に変換するA/D変
換回路、3は音声登録モードとテストモードとを
切り換えるスイツチ、4は音声登録回路、5はテ
スト回路、6はこれ等の登録回路4、テスト回路
5から得られる音声信号を一時的に貯えるバツフ
ア回路、7はこのバツフア回路6に貯えられた音
声信号を記憶するに適した状態に圧縮する信号圧
縮メモリ、8はこの信号圧縮メモリ7で圧縮され
た音声信号を記憶する参照パターンメモリ、9は
この参照パターンメモリ8からの音声信号とその
後に圧縮メモリ7から到来する音声信号とを平均
化し、その平均結果を再度参照パターンメモリ8
に書き込む平均化回路、10は参照パターンメモ
リ8からの音声信号とその後に圧縮メモリ7から
得られる音声信号との一致を検出する一致検出回
路で、両信号が不一致の場合は上記テスト回路5
に再度同じ音声に対してのテストを実行する指令
Nを与え、また両信号が一致の場合は上記テスト
回路5に次の音声に就いてのテストを行う指令Y
を与える。 The present invention provides a method for efficiently testing this registered audio signal, and will be described in detail below. The figure is a block diagram showing the configuration when implementing the method of the present invention, in which 1 is a microphone that converts audio into an audio signal, and 2 is an A/D that converts the audio signal from the microphone into a digital signal. a conversion circuit; 3 a switch for switching between the voice registration mode and the test mode; 4 a voice registration circuit; 5 a test circuit; 6 a buffer for temporarily storing voice signals obtained from the registration circuit 4 and the test circuit 5; 7 is a signal compression memory for compressing the audio signal stored in the buffer circuit 6 into a state suitable for storage; 8 is a reference pattern memory for storing the audio signal compressed by the signal compression memory 7; 9 is a reference pattern memory for storing the audio signal compressed by the signal compression memory 7; The audio signal from the reference pattern memory 8 and the audio signal arriving from the compression memory 7 after that are averaged, and the average result is returned to the reference pattern memory 8.
10 is a coincidence detection circuit that detects coincidence between the audio signal from the reference pattern memory 8 and the audio signal obtained from the compression memory 7 after that, and when the two signals do not match, the test circuit 5
A command N is given to test the same sound again, and if both signals match, a command Y is given to the test circuit 5 to test the next sound.
give.
而して音声の登録モードに於てはスイツチ3を
登録回路4側に投入する事に依つてマイクロフオ
ン1からの音声信号はA/D変換回路2でデイジ
タル化された後バツフア回路6、信号圧縮回路7
を経て参照パターンメモリ8に導入記憶される。 In the audio registration mode, by turning on the switch 3 to the registration circuit 4 side, the audio signal from the microphone 1 is digitized by the A/D conversion circuit 2, and then sent to the buffer circuit 6. Compression circuit 7
The pattern is introduced and stored in the reference pattern memory 8 through the process.
先にも説明した如く、一回の音声登録では音声
認識率が低い場合があるので、音声登録後、テス
トモードに移し、参照パターンメモリ8に記憶せ
しめた音声データのテストを行う。 As described above, the speech recognition rate may be low after one-time speech registration, so after the speech registration, the mode is shifted to the test mode and the speech data stored in the reference pattern memory 8 is tested.
スイツチ3をテスト回路5側に倒すとテスト回
路5が作動し、音声信号はA/D変換回路2、バ
ツフア回路6、圧縮回路7を経て平均化回路9並
びに一致検出回路10に導入される。先ず平均化
回路9に於てはこの時到来した音声信号と先に参
図パターンメモリ8に貯えた音声信号との平均化
を行い、再度その平均化結果を参照パターンメモ
リ8に書き込む。また一致検出回路10に於ては
この両信号の一致が検出され、一致が採れた場合
は参照パターンメモリ8に貯えた音声信号が正確
なパターンとして記憶されているとしてテスト回
路5に次の音声についてのテストを実行するよう
にとの指令Yを出し、不一致の場合は参照パター
ンメモリ8に貯えた音声信号が不正確なパターン
として記憶されている事を示しているので、テス
ト回路5に再度同じ音声についてのテストを行え
との指令Nを出す。この再テストは両信号の一致
が採れるまで、即ち参照パターンメモリ8に貯え
た音声信号が正確なパターンに修正されるまで繰
り返される。 When the switch 3 is turned to the test circuit 5 side, the test circuit 5 is activated, and the audio signal is introduced into the averaging circuit 9 and the coincidence detection circuit 10 via the A/D conversion circuit 2, the buffer circuit 6, and the compression circuit 7. First, the averaging circuit 9 averages the audio signal that has arrived at this time with the audio signal previously stored in the reference pattern memory 8, and writes the averaging result into the reference pattern memory 8 again. In addition, the coincidence detection circuit 10 detects a coincidence between these two signals, and if a coincidence is found, it is assumed that the audio signal stored in the reference pattern memory 8 is stored as an accurate pattern, and the next audio signal is sent to the test circuit 5. If they do not match, this indicates that the audio signal stored in the reference pattern memory 8 is stored as an inaccurate pattern, so the test circuit 5 is issued a command Y to perform a test on the Issues command N to conduct a test on the same voice. This retest is repeated until both signals match, that is, until the audio signal stored in the reference pattern memory 8 is corrected to the correct pattern.
本発明は以上の説明から明らかな如く、参照パ
ターンメモリに貯えた音声信号と、その後に得ら
れる音声信号との平均化を行つてその結果を再度
参照パターンメモリに書き込むと共にこの両信号
の一致を検出していてその一致が採れるまでテス
トモードを繰り返すので、参照パターンメモリに
記憶せしめた音声信号が非常に正確なものとな
り、音声認識率を高いものとする事が出来る。 As is clear from the above description, the present invention averages the audio signal stored in the reference pattern memory and the audio signal obtained thereafter, writes the result again to the reference pattern memory, and checks whether the two signals match. Since the test mode is repeated until a match is found during detection, the speech signal stored in the reference pattern memory becomes very accurate and the speech recognition rate can be made high.
図は本発明方法を実施する際の構成を示すブロ
ツク図であつて、4は登録回路、5はテスト回
路、8は参照パターンメモリ、9は平均化回路、
10は一致検出回路、を夫々示している。
The figure is a block diagram showing the configuration when implementing the method of the present invention, in which 4 is a registration circuit, 5 is a test circuit, 8 is a reference pattern memory, 9 is an averaging circuit,
Reference numeral 10 indicates a coincidence detection circuit.
Claims (1)
【特許請求の範囲】
1 音声を音声信号に変換するマイクロフオン
と、該マイクロフオンからの音声信号をA/D変
換するA/D変換回路と、該変換回路でデイジタ
ル化された音声信号を参照パターンメモリに貯え
る音声登録手段と、上記A/D変換回路からのデ
イジタル化音声信号と上記参照パターンメモリに
貯えた音声信号とを比較検査する為に、先に参照
パターンメモリに貯えられた音声信号とその後に
得られる音声信号との平均化を行つてその平均化
結果を再度参照パターンメモリに書き直す平均化
回路、並びに参照パターンメモリからの音声信号
とその後に得られる音声信号との一致を検出する
一致検出回路を備えたテスト手段とを用い、
音声登録モードに於ては上記登録手段に依つて
参照パターンメモリに順次音声信号を登録し、
テストモードに於ては上記テスト手段に依つて
先に参照パターンメモリに貯えた音声信号とその
後に得られる音声信号との平均化を行いその平均
化出力を新たに参照パターンメモリに書き込むと
共に上上記一致検出回路でこの両音声信号の一致
が採れるまでこの平均化処理を繰り返す事を特徴
とした音声機器に於ける音声信号処理方法。[Claims] 1. Refers to a microphone that converts audio into an audio signal, an A/D conversion circuit that converts the audio signal from the microphone into A/D, and an audio signal digitized by the conversion circuit. A voice registration means stored in the pattern memory and a voice signal previously stored in the reference pattern memory in order to compare and inspect the digitized voice signal from the A/D conversion circuit and the voice signal stored in the reference pattern memory. an averaging circuit which performs averaging with the subsequently obtained audio signal and rewrites the averaged result in the reference pattern memory, and detects a match between the audio signal from the reference pattern memory and the subsequently obtained audio signal. In the voice registration mode, the voice signals are sequentially registered in the reference pattern memory by the above-mentioned registration means, and in the test mode, the voice signals are registered in the reference pattern memory first by the above-mentioned test means in the voice registration mode. The audio signal stored in the reference pattern memory and the audio signal obtained thereafter are averaged, and the averaged output is newly written to the reference pattern memory, and this process is continued until the above coincidence detection circuit finds a match between the two audio signals. An audio signal processing method in audio equipment characterized by repeating averaging processing.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP12851679A JPS5651799A (en) | 1979-10-04 | 1979-10-04 | Sound signal handling system in acoustic device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP12851679A JPS5651799A (en) | 1979-10-04 | 1979-10-04 | Sound signal handling system in acoustic device |
Publications (2)
Publication Number | Publication Date |
---|---|
JPS5651799A JPS5651799A (en) | 1981-05-09 |
JPS6126080B2 true JPS6126080B2 (en) | 1986-06-18 |
Family
ID=14986665
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
JP12851679A Granted JPS5651799A (en) | 1979-10-04 | 1979-10-04 | Sound signal handling system in acoustic device |
Country Status (1)
Country | Link |
---|---|
JP (1) | JPS5651799A (en) |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPS5870293A (en) * | 1981-10-22 | 1983-04-26 | 日産自動車株式会社 | Voice recognition equipment for vehicle |
JPS5888797A (en) * | 1981-11-20 | 1983-05-26 | 三洋電機株式会社 | Voice recognition equipment |
JPS58136098A (en) * | 1982-02-08 | 1983-08-12 | 株式会社京三製作所 | Recognition pattern learning system |
JPS5934598A (en) * | 1982-08-20 | 1984-02-24 | 大日本印刷株式会社 | Voice input unit for printing |
-
1979
- 1979-10-04 JP JP12851679A patent/JPS5651799A/en active Granted
Also Published As
Publication number | Publication date |
---|---|
JPS5651799A (en) | 1981-05-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP3263392B2 (en) | 2002-03-04 | Text processing unit |
JPS5862699A (en) | 1983-04-14 | Voice recognition equipment |
CN113259832B (en) | 2023-04-28 | Microphone array detection method and device, electronic equipment and storage medium |
EP0302614A1 (en) | 1989-02-08 | Speech recognition device |
JPS6126080B2 (en) | 1986-06-18 | |
JPS6312312B2 (en) | 1988-03-18 | |
JPS63179398A (en) | 1988-07-23 | Voice recognition |
JPS5812426A (en) | 1983-01-24 | Testing device for analog-to-digital converter |
JPS6135494A (en) | 1986-02-19 | Voice recognition processor |
CN112331233B (en) | 2024-07-23 | Auditory signal identification method, device, equipment and storage medium |
JPS6332399B2 (en) | 1988-06-29 | |
JPS63278100A (en) | 1988-11-15 | Voice recognition equipment |
JP3063855B2 (en) | 2000-07-12 | Finding the minimum value of matching distance value in speech recognition |
JPS59119284A (en) | 1984-07-10 | Defect analyzing device of logical circuit |
JPS58205199A (en) | 1983-11-30 | Voice recognition equipment |
JPS62134557A (en) | 1987-06-17 | Analogue signal output apparatus for ultrasonic flaw detector of multiplexer system |
JPH07260827A (en) | 1995-10-13 | Apparatus for measuring waveform |
JPH0262879B2 (en) | 1990-12-26 | |
JPS61250699A (en) | 1986-11-07 | Voice recognition equipment |
JPS60205600A (en) | 1985-10-17 | Voice recognition equipment |
JPS62102291A (en) | 1987-05-12 | Automatic diagnosing apparatus for voice input/output unit |
JPH10135820A (en) | 1998-05-22 | Method for inspecting counter operation and serial access memory |
JPH06214594A (en) | 1994-08-05 | Voice recognition method |
JPH0459637B2 (en) | 1992-09-22 | |
JPH0421879B2 (en) | 1992-04-14 |