JP2002287800A - Speech signal processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To clearly hearably reproduce only necessary speech signal within a short time by a simple operation. SOLUTION: A speech signal processor fetches coded speech data in block units by a data fetching part 15 according to the request. A speech detection control deciding part 18 decides whether the coded speech data of a fetched block is a coded data of a speech signal or a coded data of a non-speech signal, and when the decision result is the coded data of the non-speech signal, the deciding part scraps the coded data and requests the data fetching part 15 to fetch data the next block, and when the decision result the coded data of the speech data, the deciding part outputs the coded data and requests the data fetching part 15 to fetch the next data. By processing the coded data outputted from the deciding part 18 by decoding, non-speech signal data blocks among the coded speech data can be scrapped, and only speech signal data blocks can be decoded and reproduced.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to an audio signal processing device for recording or reproducing encoded audio data.

[0002]

2. Description of the Related Art With the development of signal processing technology, a digital signal coded by PCM (Pulse Code Modulation) is used to handle an audio signal by CELP (Code Excited Linear).
Prediction) and other methods are used for communication, voice recording, and the like, which are handled by compressing (encoding) / expanding (decoding) a voice.

For example, in a communication system such as a mobile phone,
When encoding a voice signal, it is determined whether or not the signal is a voice signal or a non-voice signal (voice detection control). For the non-voice signal, control is performed such that data is transmitted only intermittently, thereby reducing transmission power. On the receiving side (decoding side), power saving communication is realized by replacing the missing encoded data with pseudo noise (Comfort Noise) and reproducing. Such a speech coding system is, for example, an ETSI (European Telec)
AMR (Adaptive Multi-Rate) standardized by the Communications Standards Institute.

In recent years, devices that apply the above-described audio encoding technology to record / reproduce audio-encoded data on a randomly accessible medium such as a flash memory have been increasing. Such audio data reproducing devices often require a method of recording for a long time without discriminating whether the audio information is necessary or not, and then re-listening the content again later in a short time. In this case, it is realized by the user repeating the fast-forward operation and the reproduction operation, or by performing a special reproduction such as a double-speed reproduction to enable a quick listening.

[0005]

However, when the user tries to reproduce only the necessary audio data in a short time, the operation is cumbersome in the repetition of the fast-forward operation and the reproduction operation as described above, and the necessary audio information is not obtained. There is a possibility of missing in fast forward. In the method using double-speed playback, the operation is simple, but since audio information is played back at a speed different from the original playback speed, it is very difficult to hear, and it is not possible to express subtle nuances such as intonation. There's a problem.

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and an object of the present invention is to enable a user to reproduce only necessary audio data in a short time with a simple operation, and to listen to the reproduction. An object of the present invention is to provide an audio signal processing device which can easily express subtle nuances.

[0007]

According to a first aspect of the present invention, there is provided an audio signal processing apparatus for reproducing audio encoded data which is encoded in units of blocks together with audio detection control. Data fetching means for fetching voice coded data in units of blocks in response to a request, and for voice coded data of a block fetched by this means, determine whether the data is coded data of a voice signal or coded data of a non-voice signal, If the determination result is encoded data of a non-voice signal, the encoded data is discarded, and a request for data acquisition of the next block is issued to the data acquisition means,
When the result of the determination is encoded data of an audio signal, a decision control means for outputting the encoded data and requesting the data acquisition means to acquire the next data; and an encoded data output from the decision control means. And decoding means for decoding the non-speech signal data block of the speech coded data and decoding and reproducing only the speech signal data block.

Therefore, according to the first aspect, when audio encoded data accompanied by audio detection control is reproduced, non-audio signal data blocks are skipped, and only audio signal data blocks are decoded. Therefore, it is possible to reproduce in a short time and at a speed that is easy to hear.

According to a second aspect of the present invention, there is provided an audio signal processing apparatus for encoding and recording an audio signal, wherein the input signal is divided into blocks, and audio detection control means for determining whether the signal is an audio signal or a non-audio signal for each block; A speech encoding unit that encodes a signal into speech encoded data for each block; and among the speech encoded data obtained by the speech encoding unit, the speech detection control unit determines that the signal is a non-speech signal. Audio encoded data recording means for discarding the audio encoded data of the block and recording only the audio encoded data determined to be the audio signal, so that the data of the non-audio signal among the audio encoded data is provided. The block is discarded, and the data block of the audio signal is recorded.

Therefore, according to the second invention, it is possible to save the capacity of the recording medium, and to reproduce a clear voice in a short time without using the reproduction method described in the first invention. Becomes

A third invention provides an audio reproduction processing unit for reproducing audio-encoded data which is audio-encoded in block units with audio detection control, and an audio recording processing unit for encoding and recording an audio signal. In the audio signal processing device, the audio reproduction processing unit captures the audio coded data in units of blocks in response to a request, and performs the audio detection on the audio coded data of the block fetched by the means. Determining whether the data is coded data of a voice signal or coded data of a non-voice signal based on the information, and if the determination result is coded data of a non-voice signal, discards the coded data to remove the coded data; Requesting data acquisition of the next block, and if the determination result is encoded data of an audio signal, output the encoded data to output the data The determination means for requesting the next means to take in the next data, and the decoding processing means for decoding the encoded data output from the determination control means. The data block is discarded, only the data block of the audio signal is decoded and reproduced, and the audio recording processing unit blocks the input signal, and audio detection control means for determining whether the audio signal or the non-audio signal for each block, The audio encoding means for encoding the input signal into audio encoded data for each block, among the audio encoded data obtained by the audio encoding means,
A voice coded data recording unit for discarding voice coded data of a block determined as a non-voice signal by the voice detection control unit and recording voice coded data determined as a voice signal; The data block of the non-voice signal among the encoded data is discarded, and only the data block of the voice signal is recorded.

Therefore, according to the third aspect of the present invention, when the audio coded data with the audio detection control is reproduced while having the effects of the first and second aspects, the data block of the non-audio signal is reproduced. Is skipped, and only the data block of the audio signal is decoded, so that it is possible to reproduce the data in a short time and at an easy-to-recognize speed, to save the capacity of the recording medium, and at the time of the reproduction as described in the first invention. Even without such a reproducing method, clear voice reproduction can be performed in a short time.

According to a fourth aspect of the present invention, in the configuration of the first or third aspect, the voice-encoded data that is voice-coded in block units with the voice detection control includes a voice signal for each block. Speech detection information for judging whether the block is a non-speech signal is added, and the determination control means determines whether the block is coded data of a speech signal or coded data of a non-speech signal based on the speech detection information added for each block. Is determined.

According to the fourth aspect, it is not necessary for the judgment control means to newly judge whether the audio signal or the non-audio signal is for each block, and the configuration can be simplified.

According to a fifth aspect of the present invention, in the configuration of the first or third aspect, the audio signal is kept for a predetermined period (for example, a speech coding frame cycle period) during the discarding process of the non-speech signal by the determination control means. When the determination of the encoded data is not obtained, a pseudo noise signal is output or the received frame signal is decoded and output as it is.

According to the fifth aspect of the present invention, the predetermined time or more
When the encoded data of the audio signal is not taken in, the silence pseudo signal is reproduced, so that there is no possibility of being erroneously recognized as a failure.

[0017]

Embodiments of the present invention will be described below in detail with reference to the drawings.

FIG. 1 is a block diagram showing an audio reproduction processing configuration as an embodiment of an audio signal processing device according to the present invention.

This audio signal processing apparatus uses an audio encoding system with audio detection control stored in the apparatus, for example, an AM
An apparatus that decodes and reproduces audio encoded data encoded by R (Adaptive Multi-Rate), and is particularly characterized by having a fast-listening reproduction function according to the present invention.

In FIG. 1, a voice coded data storage unit 1
In 1, data digitally recorded by an audio recording processing device different from the present device or encoded voice data downloaded by a communication device different from the present device is stored. The key input unit 12 senses a user's operation such as reproduction (normal reproduction, fast-listening reproduction), stop, and the like, and the sensed information is notified to the main control unit 13. The main control unit 13 controls each processing unit according to the information notified from the key input unit 12. Here, if the notification information is a reproduction operation, the voice coded data stored in the voice coded data storage unit 11 is extracted in units of one frame in response to a data request signal from the voice coded data acquisition unit 15, The data is sent to the stored data reproduction buffer 14. At this time, if the notification information from the key input unit 12 is a normal playback operation, the fast-listening control signal is deactivated ("0").
1 ").

The stored data reproduction buffer 14 operates as a ring buffer, and includes a write pointer register (wp) 141 and a read pointer register (rp).
142 and a reproduction buffer memory 143. The value stored in the write pointer register 141 is
The write start address in the reproduction buffer memory 143 when the main control unit 13 writes the encoded audio data is shown.
Therefore, the main control unit 14 writes data from the address indicated by the write pointer register 141 when writing the audio coded data, and updates the stored value of the write pointer register 141 by the written data size. On the other hand, the read pointer register 142 indicates a reading start address in the reproduction buffer memory 143 when the audio encoded data acquisition unit 15 acquires the audio encoded data.

Upon receiving the frame readout notification signal transmitted in the voice coded frame period from the voice detection control determining section 18, the voice coded data capturing section 15 performs voice coding for one frame from the stored data reproduction buffer 14. The data is read and stored in the audio encoding frame buffer 16, and a frame write notification signal is sent to the audio detection control determination unit 18. In addition, the amount of accumulated data in the accumulated data reproduction buffer 14 is monitored, and when the accumulated data amount falls below the threshold value, the data request signal is activated to urge the main control unit 13 to write the encoded audio data. When reading the audio coded frame, the maximum frame size is first read without updating the value stored in the write pointer register 142 of the stored data reproduction buffer 14, and the process waits until the next frame read notification signal. Unless there is control of data re-reading at the time of early listening reproduction, which will be described later, the audio coded data capturing unit 15 basically repeats the above processing for each frame period of the audio coding system.

When the voice detection control determining unit 18 recognizes that the voice coded frame has been written into the voice coded frame buffer 16 by the data write notification signal,
With reference to the frame, it is determined whether the frame is a frame in which audio information is encoded or a frame in which non-audio information is encoded. The determination is made by referring to the frame type information in the audio encoded frame and the frame size table 17.

The frame size table 17 stores information in which the frame type is associated with the frame type together with the identification of voice / non-voice. For example, when the audio coding method is AMR, the frame type differs for each coding rate, and has the relationship shown in FIG. By referring to this table 17, the voice detection control determination unit 18 recognizes the size of the frame written in the voice coded frame buffer 16 together with the voice / non-voice discrimination, and determines the size of the stored data reproduction buffer by the frame size. 14
Of the read pointer register 142 is updated.

Here, when the fast listening control signal from the main control unit 13 is inactive ("0"), that is, in normal reproduction, the voice detection control determining unit 18 performs voice decoding on the voice coded frame within the frame period. The data is sent to the processing unit 20, and the completion of frame reading is notified to the audio coded data capturing unit 15 by the frame reading notification signal. The pseudo-noise is generated and output from the audio decoding processing unit 20 during normal reproduction of the non-audio information frame.

On the other hand, when the fast-listening control signal is active ("
1 "), that is, in the case of fast-listening playback, the operation differs depending on whether the audio coded frame is audio information / non-speech information. In the case of audio information, the same operation as in normal playback is performed, In this case, the non-speech determination notification signal is made active ("1"), which becomes an input to the AND circuit 19 together with the early-hearing control signal, and is output to the speech coded data acquisition unit 15 as a data reread signal. That is, when both the early-hearing control signal and the non-voice determination notification signal are active ("1"), the data reread signal becomes active ("1").

When receiving the active data rereading signal, the voice coded data capturing section 15 captures the next frame from the stored data reproduction buffer 14 without waiting for the frame period, and sends it to the voice detection control determining section 18 again. . This control is repeated until the voice detection control determination unit 18 detects a frame of voice information. as a result,
The frame of the non-voice information is discarded and skipped, and only the frame of the voice information is sent to the voice decoding processing unit 20.

The audio decoding processing section 20 decodes the input audio encoded frame into a PCM signal, and the decoded signal is converted into an analog audio signal by a D / A (digital / analog) converter 21, and is output by a speaker 22. The voice is output.

FIGS. 3A and 3B show waveforms of an audio signal decoded by the above reproduction control. FIG. 3A shows an audio encoded data stream, and FIG. 3B shows an audio waveform at the time of normal reproduction of the stream of FIG. , (C) show audio waveforms when the stream of (a) is played back at high speed. As shown in FIG. 3A, a period T during which a speech coded frame of speech information continues.
1, T3, a voice coded data stream in which a period T2 in which a voice coded frame of non-voice information continues is
As shown in FIG. 3B, all frames of the decoded audio waveform during normal reproduction are reproduced. On the other hand, as shown in FIG. 3 (c), the decoded audio waveform at the time of the fast listening reproduction is skipped in the audio encoded frame period T2 of the non-audio information and is not decoded. That is, the pseudo noise waveform corresponding to T2 is omitted from the normal reproduction audio waveform, and the audio information waveforms corresponding to T1 and T3 are reproduced in a continuous form. From these waveforms, it can be seen that only significant audio information can be reproduced in a short time.

Therefore, according to the audio signal processing device having the above-described configuration, when the operation of designating the fast listening reproduction is performed, the non-audio information is skipped by the reproduction processing of the encoded audio data accompanied by the audio detection control. For this reason, only significant audio information can be reproduced in a short time by a simple operation for the user, and since it is reproduced at a constant speed that is easy to hear, subtle nuances can be expressed.

In the above embodiment, when the frame duration T2 of the non-voice information is long and the period required for frame skipping exceeds one frame, that is, within the processing time of the buffer cycle of the voice detection control processing unit 18, When the frame of the audio information cannot be detected, the audio decoding processing unit 20 outputs the pseudo noise. At that time, the Isle outputs the received frame as it is and decodes and reproduces it. Thus, an effect of preventing the user from erroneously recognizing a failure can be obtained.

FIG. 4 is a block diagram showing an audio recording processing configuration as another embodiment of the audio signal processing apparatus according to the present invention.

This audio signal processing apparatus is characterized in that it records audio data encoded by an audio encoding system with audio detection control, for example, AMR, and has a short-time recording function according to the present invention. Things.

In FIG. 4, the key input unit 31 detects an operation such as recording (normal recording, short-time recording) and stop from the user, and notifies the main control unit 32 of the information. The main control unit 32 controls each processing unit according to the notified information. In the case of a recording operation, the stored data recording buffer 34 for performing FIFO processing on the output data of the audio encoded data recording unit 33 is polled, the audio encoded data written in the buffer 34 is taken out, and the audio encoded data accumulation unit 35 To record. At this time, if the recording operation is normal recording, the short-time recording control signal is deactivated ("0"),
For short-time recording, the short-time recording control signal is activated ("1"). Hereinafter, the recording operation will be described along the flow of audio data.

The analog audio signal input from the microphone 36 is converted to an A / D (analog / digital) converter 3
The signal is converted into a PCM signal at 7 and input to the voice detection control determination section 38. The voice detection control determination unit 38 determines whether the information is voice information or non-voice information by analyzing the PCM signal (for details on voice detection control determination, refer to VAD control of AMR). Processing unit 39
Is repeated for each frame period. When the short-time recording control signal from the main control unit 32 is active ("1") and the input signal is determined to be non-voice information, the non-voice determination notification signal is activated ("1"). To This signal is input to the AND circuit 40 together with the short-time recording control signal, and is both active ("
1 "), it is output as a data recording control signal to the voice coded data recording unit 33. That is, when both the short-time recording control signal and the non-voice determination notification signal become active (" 1 "), the data recording control signal is output. The signal also becomes active ("1").

The audio encoding processing section 39 encodes the input PCM data by a predetermined audio encoding method, for example, AMR, outputs the encoded audio encoded frame to the audio encoded frame buffer 41, and outputs a frame write notification. A signal is sent to the voice coded data recording unit 33 to notify completion of frame writing.

Upon receiving the frame write notification signal, the voice coded data recording unit 33 reads the voice coded frame from the voice coded frame buffer 41, and stores the frame type information and the frame size table 4 in the frame.
2, the frame size of the speech coded frame is recognized. The frame size table 42
As in the case of FIG. 2, information in which the frame type is associated with the identification of voice / non-voice with respect to the frame type is stored.

Here, when the data recording control signal is inactive ("0"), that is, when normal recording is performed, the audio encoded data recording unit 33 writes the encoded audio frame by the frame size into the accumulated data recording buffer 34. , End the processing of the frame. On the other hand, when the data recording control signal is active (“1”), short-time recording control is performed, and the audio encoded data stored in the audio encoded frame buffer 41 is non-audio information. It is not output to the storage data recording buffer 34. As a result, the frame is discarded, and the encoded voice data storage 35
Will not be recorded.

Therefore, according to the audio signal processing apparatus having the above-described configuration, at the time of the short-time recording control, the audio coded data of only the audio information is recorded, and the non-audio information is discarded. Also, at the time of reproduction, even if the reproduction method as described in the embodiment of FIG.
Clear voice reproduction is possible in a short time with normal reproduction.

In the above embodiment, the case where the present invention is applied to a single reproduction unit and a single recording unit has been described. However, it is also possible to apply the present invention to a recording processing unit and a reproduction processing unit of a recording / reproducing apparatus. According to this configuration, when the capacity of the recording medium has room, the encoded data of the audio signal is recorded as it is by the normal recording process, and the non-audio information is discarded at the time of reproduction, and the early listening process is performed. If there is no room in the capacity of the medium, a usage method such as performing short-time recording control for discarding non-voice information during voice recording can be realized.

Further, in the above embodiment, the fast listening playback is performed at the same speed as the normal playback. However, the present invention is not limited to this, and the speed may be changed as needed. Of course.

[0042]

As described above, according to the present invention, non-speech information is skipped when speech coded data accompanied by speech detection control judgment is reproduced, so that only speech information can be reproduced in a short time. It is possible to reproduce at a speed that is easy to hear. Also, only the audio information is recorded during the short-time recording control, and the non-audio information is discarded, so that the capacity of the recording medium can be saved, and at the time of reproduction, normal reproduction can be performed in a short time without using a special reproduction method. Sound reproduction is possible.

Therefore, only necessary voice data can be reproduced in a short time by a simple operation for the user,
In addition, it is possible to provide an audio signal processing device which is easy to hear in the reproduction and can express subtle nuances.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an audio reproduction processing configuration as an embodiment of an audio signal processing device according to the present invention.

FIG. 2 is an exemplary view showing information contents of a frame size table according to the embodiment;

FIG. 3 is an exemplary waveform chart showing an audio signal waveform decoded by the reproduction control of the embodiment.

FIG. 4 is a block diagram showing an audio recording processing configuration as an embodiment of the audio signal processing device according to the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 11 ... Voice coded data storage part 12 ... Key input part 13 ... Main control part 14 ... Stored data reproduction buffer 141 ... Write pointer register (wp) 142 ... Read pointer register (rp) 143 ... Reproduction buffer memory 15 ... Voice coding Data import unit 16 ... Sound encoding frame buffer 17 ... Frame size table 18 ... Sound detection control determination unit 19 ... AND circuit 20 ... Sound decoding processing unit 21 ... D / A (digital / analog) converter 22 ... Speaker 31 ... Key Input unit 32 Main control unit 33 Voice encoded data recording unit 34 Stored data recording buffer 35 Voice encoded data storage unit 36 Microphone 37 A / D (analog / digital) converter 38 Voice detection control determination Unit 39: voice coding processing unit 40: AND circuit 41: voice code Frame buffer 42: Frame size table

Claims (5)

[Claims] 1. An audio signal processing apparatus for reproducing encoded audio data encoded in units of blocks together with audio detection control, comprising: a data capturing unit for acquiring the encoded audio data in units of blocks in response to a request; It is determined whether the coded data of the block fetched by this means is coded data of a voice signal or coded data of a non-voice signal, and if the determination result is coded data of a non-voice signal, the code is determined. Discards the coded data and requests the data fetching means to fetch data of the next block. If the determination result is coded data of an audio signal, the coded data is output and the next Determination control means for requesting data capture, and decoding processing means for decoding encoded data output from the determination control means; An audio signal processing device. 2. An audio signal processing apparatus for encoding and recording an audio signal, wherein: an input signal is divided into blocks, and audio detection control means for determining whether an audio signal or a non-audio signal is provided for each block; Voice coding means for coding into voice coded data; and voice coded data of a block determined to be a non-voice signal by the voice detection control means among voice coded data obtained by the voice coded means. And a voice coded data recording unit for recording voice coded data determined to be a voice signal. 3. An audio signal processing device comprising: an audio reproduction processing unit that reproduces audio encoded data that is audio-encoded in block units with audio detection control; and an audio recording processing unit that encodes and records an audio signal. In the above, the audio reproduction processing unit may include: a data fetching unit that fetches the coded audio data in units of blocks in response to a request; and for the coded audio data of the block fetched by the unit, the coded data of the audio signal or the non-audio It is determined whether the data is coded data of a signal, and if the result of the determination is coded data of a non-voice signal, the coded data is discarded and a request is made to the data fetching means to fetch data of the next block. If the result is encoded data of an audio signal, the encoded data is output and a decision system for requesting the data acquisition means to acquire the next data is issued. Control means; and decoding processing means for decoding encoded data output from the determination control means, wherein the audio recording processing section blocks the input signal, and performs an audio signal or a non-audio signal for each block. Voice detection control means for determining whether or not the input signal is coded into voice coded data for each block; and voice detection among voice coded data obtained by the voice coding means. Voice-coded data recording means for discarding voice-coded data of a block determined to be a non-voice signal by the control means and recording voice-coded data determined to be a voice signal; Signal processing device. 4. Speech coded data that is coded in units of blocks with the voice detection control is provided with voice detection information for determining whether a voice signal or a non-voice signal is provided for each block, 4. The voice according to claim 1, wherein the determination control means determines whether the block is coded data of a voice signal or coded data of a non-voice signal based on voice detection information added in block units. Signal processing device. 5. The decoding processing means outputs or receives a pseudo-noise signal when the determination control means cannot determine the encoded data of the audio signal for a certain period or more during the discard processing of the non-audio signal. 4. The audio signal processing device according to claim 1, wherein the frame signal is decoded and output as it is.