JP2006308993A - Device for recording and reproducing speech signal - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To automatically divide a news program and a language teaching program into files at soundless portions and to store them into a storage medium in a radio receiver etc with a built-in AM/FM tuner. <P>SOLUTION: A speech signal level comparison portion 12 compares an input speech signal level with a prescribed level to detect a soundless portion. A control section 16 divides the speech signal into a file at the soundless portion and stores the file into the storage medium 22 when the soundless portion lasts for a certain time and the speech portion right before it lasts for a certain time. Soundless data or index data may forcibly be added to the head of the file and the head of the file may be faded in. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an audio signal recording / reproducing apparatus, and more particularly to an apparatus for detecting and dividing a silent part of an input audio signal.

  2. Description of the Related Art Conventionally, a technique is known in which a silent portion of an input audio signal is detected, divided into a plurality of files, and stored in a storage medium. For example, in Patent Document 1 shown below, a voice signal level is compared with a predetermined level to determine a silent part of the voice signal, and the voice signal is divided into a plurality of call files by the silent part to record data. It is disclosed that it records in a part.

  Patent Document 2 discloses a technique for automatically detecting the start and end of audio data and performing a fade-in process or a fade-out process.

JP 2001-285461 A JP 2001-228887 A

  However, in the configuration in which the silent part of the input audio signal is simply detected and the detected silent part is divided into a plurality of files, for example, in a news program, although there are originally a series of news articles, There was a problem such as division recording due to the presence of a silent part on the way. Whether or not it is a silence part can be determined based on whether or not the silence state lasts for a certain period of time, but even if it is a silence part that lasts for a certain period of time, there may be cases where it should not be divided into different files from the context. Yes, if it is divided in such a case, there will be a problem that the audio is interrupted midway during file playback, or the file is disclosed from the middle of the sentence when the file is selected, and the user needs to edit or process the file separately. Will occur.

  Furthermore, when each file is played back with the fade-in process at the beginning, there is no problem as long as it is a file that records music data, but in news articles, etc., it becomes difficult to hear the head sound due to the fade-in process. Problems arise.

  An object of the present invention is to automatically divide an input audio signal into a plurality of files and enable user convenience during playback, that is, random access, and the audio signal is not interrupted even if fade-in processing is performed. The object is to provide an apparatus capable of reproduction.

  The present invention includes a silence detection means for detecting a silence part of an input audio signal, a first determination means for judging whether or not the silence part has continued for a first predetermined time, and the silence part is a first part. When it is determined that the sound part has continued for a certain period of time or longer, second determination means for determining whether or not the sound part immediately before the silent part has continued for a second certain period of time or more, If the input audio signal is determined to have continued for a certain period of time, the input audio signal is divided into files by the silent section and stored in a storage medium, operation means operable by the user, and the operation means Reproduction means for reproducing the input audio signal stored in the storage medium in units of files in response to the operation signal.

  In the present invention, the silent part is not simply detected and divided into files, but the silent part is continuously present for a certain time or more, and the sound part immediately before the silent part is continuously present for a certain time or longer. Or the input audio signal is divided into files and stored in a storage medium. That is, when the silent part continues for a certain period of time and the immediately preceding voice part exists continuously for a certain period of time, the file is divided in the middle of a sentence with a simple configuration by dividing the file.

  In the present invention, by adding silence data or audio index data that continues for a predetermined time to the head of each divided file, the influence on the original audio data can be suppressed even if fade-in processing is performed during file reproduction.

  According to the present invention, it is possible to automatically divide an input audio signal into a plurality of files and to make user's convenience during reproduction, that is, random access possible. Even if fade-in processing is performed, the audio signal can be reproduced without interruption.

  Hereinafter, an embodiment of the present invention will be described with reference to the drawings, taking a radio receiver as an example.

  FIG. 1 is a block diagram showing the configuration of an audio signal recording / reproducing apparatus according to this embodiment. In the present embodiment, the audio signal recording / reproducing apparatus functions as a radio receiver incorporating an AM / FM tuner.

  The AM / FM tuner 10 receives AM or FM radio broadcast and outputs the received audio signal to the audio signal level comparison unit 12 and the audio processing unit 14.

  The audio signal level comparison unit 12 compares the input audio signal from the AM / FM tuner 10 with a predetermined level and outputs the comparison result to the control unit 16. In the present embodiment, the audio signal level comparison unit 12 determines that the input audio signal level is equal to or higher than a predetermined level, and the audio signal level is lower than the predetermined level. The audio signal level comparison unit 12 may output a signal or flag to that effect to the control unit 16 only when the level of the input audio signal is lower than a predetermined level. A specific signal or flag may be output when the input audio signal level is equal to or higher than a predetermined level.

  The audio processing unit 14 converts the input audio signal into a digital signal, further compresses the audio signal in an audio data format such as WAV or MP3, and outputs the compressed signal to the control unit 16. Further, the audio processing unit 14 expands the audio signal supplied from the control unit 16, specifically, the audio signal read from the storage medium 22 described later, outputs it to the audio output unit 24, and reproduces it as an audio signal. .

  The control unit 16 inputs the comparison result from the audio signal level comparison unit 12 and the compressed digital audio data from the audio processing unit 14, and controls recording and reproduction of the compressed audio data according to the comparison result in the audio signal level comparison unit 12. The data is stored in the storage medium 22 via the unit 20. The control unit 16 detects a silent part of the input audio signal and determines whether or not the silent part continues for a certain time T0 or more. When the silent part continues for a certain time T0 or more, it is determined that there is a significant silent part that can be divided into file units, and the input audio signal is divided into files and stored in the storage medium 22. However, even if the silent part continues for a certain time T0 or longer, the control unit 16 does not continue when the duration of the voice part existing immediately before the silent part has not continued for a certain time T1 or longer. Do not execute file division processing. Therefore, in this case, the silent part, the immediately preceding voice part, and the immediately following voice part are collectively stored as one file. That is, the control unit 16 determines whether or not there is a silent part continuously for a certain period of time and whether or not there is an audio part that exists continuously for a certain period of time immediately before the silent part. Two determinations are executed, and whether to divide the file is determined according to the two determination results.

  The storage medium 22 includes a known semiconductor memory such as a USB memory or a flash memory, and stores the input audio signal divided into a plurality of files by the control unit 16 in units of files. The storage format is also known, for example, has a file management unit and a data storage unit, the identification data ID and file attributes of each file divided into the file management unit are stored, and divided into files for each file in the data storage unit Input voice data is stored.

  The operation unit 18 is an operation unit that can be operated by the user, and includes an arbitrary form such as a touch switch, an operation button, or an operation lever. The user operates the operation unit 18 to input a reproduction instruction for a desired file. In response to an operation signal from the operation unit 18, the control unit 16 reads the corresponding file from the storage medium 22 and supplies it to the audio processing unit 14. The audio processing unit 14 decompresses the corresponding file and outputs it to the audio output unit 24. 2. Description of the Related Art An audio player that has a built-in hard disk or flash memory, stores audio data input from a portable medium such as a network or CD, and reproduces and outputs music data in response to a user operation is known. In these audio players, the user can play music data by operating the play button, and can search for the next song or the immediately preceding song by operating the fast-forward button or the backward button. In the present embodiment, news programs and language programs on AM / FM radio are automatically divided at the silent part and stored in the storage medium 22 as a plurality of files. For example, in the case of a news program, when an announcer reads a plurality of articles, the announcer reads it with a predetermined interval between each article. Therefore, in this embodiment, each article of the news program is automatically divided into files and stored in the storage medium 22, and the user can find the beginning of an arbitrary song by operating the fast-forward button or the backward button on the audio player. It is possible to reproduce desired news articles sequentially by the same operation as performing the above.

  FIG. 2 shows a conceptual configuration of file division of the input audio signal in the present embodiment. As shown in FIG. 2A, it is assumed that the input voice signal is composed of a voice part (sound part) and a silent part. When there is a silent part 102 following the voice part 100, and the silent part 102 continues for a certain time T0 or longer, as shown in FIG. The file is divided and stored in the storage medium 22. That is, the sound unit 100 and the silence data for a certain time T0 in the silence unit 102 are divided into one file 1 and stored. A part of the silent part 102 existing at the end of the file 1 may be removed and stored in the storage medium 22. When the sound part 104 and the soundless part 106 exist after the soundless part 102, the sound part 104 is similarly divided and stored in the storage medium 22 when the soundless part 106 continues for a predetermined time T0 or more. In FIG. 2B, a file including the audio part 100 is indicated as file 1, and a file including the audio part 104 is indicated as file 2. The index number that functions as the identification data ID of each file is obtained by incrementing the counter each time a new voice part is detected.

  When the sound part 108 and the soundless part 110 exist after the soundless part 106 and the soundless part 110 continues for a certain time T0 or more, the sound part 108 immediately before the soundless part 110 does not continue for a certain time T1 or more. Therefore, the silence part 110 is determined not to be a significant silence part, and the file is not divided. The sound part 112 and the sound part 114 exist after the soundless part 110, the soundless part 114 continues for a predetermined time T0 or more, and the total sound part immediately before the soundless part 114 (in this case, the sound part 108 and the sound part 112). The file is divided into files 3 when the interval of the above is equal to or longer than the predetermined time T1. If the silence part 110 divides the file, a file including the voice part 108 is formed. However, in the news article, the voice part 108 is too short and may not be formed as a meaningful phrase or sentence. It is possible to divide the file into meaningful files by ensuring the duration of the copy or the minimum unit of the file to be divided. The minimum duration T1 of the voice unit may be input from the operation unit 18 by the user and stored in the memory of the control unit 16, or may be stored in advance as a default value in the memory of the control unit 16. The minimum duration T1 is, for example, 30 seconds. The fixed time T0 for determining the significance of the silent part is set to 2 seconds, for example.

  FIG. 3 shows a process flowchart of file division in the present embodiment. First, it is determined whether or not the input audio signal level is equal to or higher than a predetermined level (S101). The predetermined level may be stored in advance as a default value in the memory of the audio signal recording / reproducing apparatus, or may be set by the user via the operation unit 18. In an FM broadcast news program or the like, BGM may exist in the background of an audio news article. When the predetermined level is set low, there is a possibility that the news article cannot be divided into a plurality of files due to the presence of BGM. In such a case, the user can set a predetermined level according to the BGM level. Regardless of the presence of BGM, silence can be detected and news articles can be divided into files. A predetermined level for each radio station or each radio program may be stored in the memory of the control unit 16, and a predetermined level corresponding to the radio station or radio program to be stored in the storage medium 22 may be read from the memory and set. .

  When the silent part is detected with the audio signal level being lower than the predetermined level, it is determined whether or not the detected silent part is continuously present for a certain time T0 or more (S102). If the silent part does not continue for a certain time T0 or more, the division process is not performed because it is not suitable for dividing the file. On the other hand, when the silent part continues for a certain period of time T0 or more, it is further determined whether or not the voice part immediately before the silent part (the total voice part not divided) continues for a certain period of time T1 or more ( S103). For this determination, for example, a timer may be started simultaneously with the beginning of the audio unit, and the time obtained by stopping the timer when it is detected in S101 that the audio signal level is lower than a predetermined level may be compared with T1. Even if the silent part continues for a certain time T0 or more, if the immediately preceding voice part does not continue for a certain time T1 or more, the file division is not performed. Only when the immediately preceding audio part continues for a certain time T1 or more, the file is divided by the silent part detected in S101 (S104). The process of S104 can also be referred to as a process of ensuring the minimum time or the minimum data length of the file to be divided. At the time of file division, the counter value of the file counter is incremented by 1 (S105) and registered in the file management unit of the storage medium 22.

  As described above, the input audio signal is automatically divided into a plurality of files and stored in the storage medium 22. For example, when 10 news articles are read out in a news program, a total of 10 files are divided and stored. An index number is added to each file, and the user can play and output an arbitrary file at random by operating the operation unit 18.

  When the audio signal recording / playback device has a function of playing music data and has a function of playing back the top of each song as in the prior art, the top of the news article file is faded. There may be a situation where the sound is interrupted due to the in-process and it is difficult to hear the voice.

  Therefore, as shown in FIG. 4, the control unit 16 automatically divides the input audio signal into a plurality of files by the silence unit and forcibly adds silence data that continues for a predetermined time to the head of each divided file. May be stored in the storage medium 22. In FIG. 4B, the input audio signal is divided into a plurality of files 1, 2, 3 as in FIG. 2, after which the silence data 200 is added to the file 1, the silence data 202 is added to the file 2, Silence data 204 is added to the file 3. By adding silent data that continues for a predetermined time to the beginning of each file 1 to 3, as shown in FIG. 5, the beginning of each file is gradually increased from 0% output to 100% output. Even if the process is executed, the fade-in process is completed in the silent data portion forcibly added to the head of each file. As a result, the original audio part is not fade-in processed and is reproduced and output from the state of 100% output. The Therefore, even if the fade-in process is performed, the original sound is reproduced without interruption. The duration or data length of the silence data 200, 202, 204 is set to be substantially the same as the time required for the fade-in process.

  In FIG. 4, silence data 200, 202, 204 is added to the beginning of each automatically divided file 1, 2, 3, but other data (including dummy data) is added instead of silence data. It is also possible to suppress sound interruption due to the fade-in process.

  For example, as shown in FIG. 6, audio index data may be added to the head of each automatically divided file. This audio index data is obtained by adding the identification data ID of each file set in the process of S105 of FIG. 3 as audio data. In FIG. 6B, audio data “index 1” is added. When the file to which the audio index data is added is reproduced, the original audio data is reproduced after the audio “index 1”. As a result of the fade-in process being performed in the audio index portion, the portion of “index 1” is reproduced by fade-in, and the original audio portion is subsequently output as audio. With the above processing, even in an apparatus configuration in which the beginning of a file is faded in, it is possible to reliably eliminate the situation where it is difficult to hear the beginning voice.

  In this embodiment, a USB flash audio player can also be used as the storage medium 22. In this case, in FIG. 1, the control unit 16 and the USB flash audio player are connected via a USB terminal. Various types of USB flash audio players are commercially available, but they often have no fast forward / rewind playback function and only have a forward / backward button for the previous / next song. Even in such a USB flash audio player, in the present embodiment, the contents of a radio program can be automatically divided into files and downloaded to the USB flash audio player, so that there is an advantage that pseudo fast forward / rewind playback can be performed. .

1 is a configuration block diagram of an audio signal recording / reproducing apparatus according to an embodiment. It is file division explanatory drawing of the input audio | voice signal in embodiment. It is a file division processing flowchart of an embodiment. It is file division explanatory drawing in other embodiment. It is fade-in process explanatory drawing in a silence part. It is fade-in explanatory drawing in an audio | voice index part.

Explanation of symbols

  10 AM / FM tuner, 12 audio signal level comparison unit, 14 audio processing unit, 16 control unit, 18 operation unit, 20 recording / playback control unit, 22 storage medium, 24 audio output unit.

Claims (4)

Silence detection means for detecting a silence portion of the input audio signal;
First determination means for determining whether or not the silent portion continues for a first predetermined time or more;
Second determination means for determining whether or not the sound part immediately before the silent part continues for a second predetermined time or more when it is determined that the silent part continues for a first fixed time or longer; ,
A dividing unit configured to divide the input voice signal into a file by the silent part and store the file in a storage medium when it is determined that the voice part continues for a second predetermined time or more;
User-operable operation means,
In response to an operation signal from the operation means, a reproduction means for reproducing the input audio signal stored in the storage medium in units of the file;
An audio signal recording / reproducing apparatus comprising: The apparatus of claim 1.
The audio signal recording / reproducing apparatus characterized in that the dividing means appends silence data that continues for a predetermined time to the beginning of the file and stores it in the storage medium. The apparatus of claim 1.
The audio signal recording / reproducing apparatus characterized in that the dividing means adds audio index data that continues for a predetermined time to the beginning of the file and stores it in the storage medium. The apparatus according to any one of claims 2 and 3,
The audio signal recording / reproducing apparatus, wherein the predetermined time is set to be substantially the same as a fade-in time when the file is reproduced by the reproducing means.

Images