JPH07169291A - Voice recording device and voice reproducing device - Google Patents

Abstract

PURPOSE:To continuously record without degrading voice quality and to efficiently use a memory section by discriminating whether voice information of each frame is voiced or silent, storing the discriminated result information in a memory section, and recording voice information in a silent section again when capacity of the memory section reaches the maximum storing capacity. CONSTITUTION:Digital voice information by a microphone 1 and an A/D converter 3 is encoded by a encoding/decoding section 4, and frame voice information is stored in a voice memory section 9 and recorded. In this case, it is discriminated by a system control section 5 whether frame voice information is voiced or silent, the discriminated result 1 or 0 is recorded. And when capacity of the memory section 9 reaches the maximum storing capacity, frame voice information succeeding to a part of the memory section 9 in which silent discriminating information 0 is recorded is recorded again, the last address value recorded again is also stored. Thereby, continuous recording can be performed without degrading voice quality and the memory can be efficiently used. Furthermore, reproduction is performed in the same way.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice recording device and a voice reproducing device.

[0002]

2. Description of the Related Art As one of audio recording devices, there is a digital recorder that compresses and encodes an audio signal input from a microphone and records the encoded data in a semiconductor memory. JP-A-2-162954 discloses that in such a digital decoder, in order to reduce the occupation ratio of the semiconductor memory, the audio data is temporarily extracted after the audio data storage amount of the semiconductor memory reaches the maximum storage capacity. A method of recording again after performing a predetermined audio compression process is disclosed.

[0003]

However, the above-mentioned Japanese Patent Laid-Open Publication No. Hei 2-162954 discloses recording once in order to recompress recorded data when the semiconductor memory reaches the maximum storage capacity. Because it interrupts the operation,
It has the drawback that continuous recording is not possible and the overall recording time is long. Further, since the recorded audio data is extracted and the compression processing is performed again, there is a problem that the sound quality when reproduced is deteriorated.

The audio recording apparatus and the audio reproducing apparatus of the present invention have been made by paying attention to such a problem, and an object thereof is to enable continuous recording without deteriorating the sound quality, and An object of the present invention is to provide an audio recording device and an audio reproducing device that can efficiently use a semiconductor memory.

[0005]

In order to achieve the above object, the present invention converts an input voice signal into a digital signal, encodes it in frame units of a voice analysis and synthesis system, and stores it in a semiconductor memory. An audio recording device for recording, wherein a determination unit determines whether the audio information of each frame is related to voice or silence, and the determined audio information of each frame is related to voice. If there is one, the first discriminating information recording means for recording the second discriminating information, and if the capacity of the semiconductor memory reaches the maximum storage capacity, it returns to a predetermined position, and the second discriminating information. The audio information is re-recorded in the frame marked with, and the discrimination information is retained as the second re-recording means, and the address value storage means for storing the address value of the last re-recorded frame.

Further, according to the present invention, the audio data is composed in units of frames, and information indicating whether the recorded audio data is voiced or silenced and a silence frame of the audio data once recorded are provided. A voice reproducing device for reproducing voice information from a semiconductor memory in which a final address when voice data is re-recorded is recorded. When the address is detected, only the voiced frames are sequentially reproduced for the part related to the re-recording, and when the final address of the semiconductor memory is reached, the audio information re-recorded in the silent frame is sequentially processed until the final address related to the re-recording. And a reproduction control means for reproducing.

[0007]

That is, in the audio recording apparatus of the present invention, when the audio signal is recorded in the semiconductor memory in frame units of the audio analysis and synthesis system, the audio information of each frame is related to the presence or absence of sound. If the audio information of each frame that has been discriminated is voiced, the first discriminant information is recorded, and if it is voiceless, the second discrimination information is recorded. Then, when the capacity of the semiconductor memory reaches the maximum storage capacity, the semiconductor memory returns to a predetermined position, the discrimination information is held as it is while the voice information is re-recorded in the frame to which the second discrimination information is added, and The address value of the last recorded frame is stored.

In the audio reproducing apparatus of the present invention, the audio data is composed in units of frames, the recorded audio data indicates whether the recorded audio data is voiced or silent, and the audio data once recorded. When reproducing the audio information from the semiconductor memory in which the final address is recorded when the audio data is re-recorded in the silent frame, the presence or absence of the final address information related to the re-recording is detected, and the final address related to the re-recording is detected. In this case, only voiced frames are sequentially reproduced for the portion related to re-recording, and when the final address of the semiconductor memory is reached, the audio information re-recorded in the silent frame is sequentially reproduced up to the final address for re-recording.

[0009]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described in detail below with reference to the drawings. FIG. 1 is a block diagram of a digital recorder to which the present invention is applied. In the figure, the microphone 1 is coded via a microphone amplifier 2 and an A / D converter 3.
It is connected to the first terminal of the decoding unit 4. The speaker 12 is connected to the second terminal of the encoding / decoding unit 4 via the power amplifier 11 and the D / A converter 10.

The third terminal of the encoding / decoding unit 4 is connected to the audio memory unit 9 via the memory control unit 7. On the other hand, the system control unit 5 is connected to the operation input unit 6 having operation buttons for recording, playing, and stopping, the encoding / decoding unit 4, and the memory control unit 7 in order to control each unit. It is connected to the voice memory unit 9 via the address counter 8.

In the above structure, the input voice signal is converted into an analog electric signal by the microphone 1 during recording,
It is amplified by the microphone amplifier 2 and input to the A / D converter 3. The A / D converter 3 converts an analog electric signal into a digital signal. The encoding / decoding unit 4 encodes the digital signal in frame units of a voice analysis / synthesis system. The encoded digital signal is stored in the audio memory unit 9 via the memory control unit 7. This voice memory unit 9
In addition to the data output from the memory control unit 7, the index data used for finding the beginning of the recorded contents and the rewriting final address are stored in the memory. The address counter 8 performs a counting operation in accordance with the address data given from the system control unit 5, and addresses the voice memory unit 9.

Next, during reproduction, the digital data stored in the audio memory unit 9 is read out and supplied to the encoding / decoding unit 4 via the memory control unit 7. The encoding / decoding unit 4 decodes the read digital data. The decoded digital data is the D / A converter 1
At 0, it is converted back to an analog audio signal and the power amplifier
It is amplified by 1 and output from the speaker 12.

The system control unit 5 controls the operations of the encoding / decoding circuit 4, the memory control unit 7, the audio memory unit 9, and the address counter 8 during the above-described recording and reproducing operations. FIG. 2 is a diagram showing a recording configuration of the voice memory unit 9.
That is, the memory space is composed of the rewriting final frame address portion (hereinafter referred to as the RWEAD portion) 9a and the index portion 9
It is roughly divided into b and the voice data section 9c. Here, a head address is assigned to the RWEAD unit 9a and a value of 0 is written. From this,
The voice data portion 9c and the silent frame included in the voice data portion 9c have addresses 0 and subsequent addresses. In this embodiment, when the voice memory unit 9 reaches the maximum storage capacity, the voice memory unit 9 returns to a predetermined address value, and new voice data is rewritten from there to the silent frame portion, but at that time, the recording stop button is pressed. Then, the last written address value of the silent frame portion is written in the RWEAD portion 9a. In this case, R
As described above, the address written in the WEAD section 9a is the address after address 0. On the contrary, when the voice memory unit 9 does not reach the maximum storage capacity and new voice data is not rewritten in the silent frame portion, 0 is still written in the RWEAD unit 9a. Therefore, RWEA
It is possible to judge whether or not the voice memory unit 9 has reached the maximum storage capacity based on the value of the D unit 9a.

The index section 9b includes a voice data section 9
The operation start position information and the operation end position information are recorded for each of the plurality of voice message files stored in c.

Next, the operation of the digital recorder according to this embodiment will be described in detail in two cases of the recording mode and the reproduction mode. First, the operation in the recording mode will be described with reference to the flowchart of FIG. First, the operation input unit 9
Wait until the record button is pressed (step S
1, S2). When the record button is pressed, the RWEAD section 9a
Is determined to be the end address, and if the value is the end address, the process ends (step S3). If it is not the end address, it is determined whether or not the value of the RWEAD unit 9a is 0 (step S4). If it is 0, it is determined that the voice memory 9 has not reached the maximum storage capacity as described above, and the memory The control unit 7 obtains the recording start position in the voice data unit 9c of the voice memory unit 9, stores the start address indicating the start position in the index unit 9b as the operation start information, and sets it in the address counter 8 (step). S6).

After that, the audio data from the A / D converter 3 is encoded by the encoding / decoding unit 4 for one frame (step S7), and whether the encoded frame is voiced or silent. A determination is made (step S8). When it is determined that the voice is present, VS which is a flag for determining whether the voice is present or not.
To 1 (step S10), and when it is determined that there is no sound, V
S is set to 0 (step S9). Next, it is determined whether or not the value of the address counter 8 is the end address of the memory (step S11), and if it is not the end address, the encoded data and the voice / silent discrimination flag VS are set in the address counter 8. Store in the indicated address value (step S
12), the value of the address counter 8 is updated (step S13). Then, it is detected whether or not the recording stop button is pressed (step S14), and if pressed, 0, which is information indicating that the memory has not reached the storage amount, is set to RW.
It is stored in the EAD unit 9a (step S15), the end position is determined, and the recording process is exited (step S16).
If not pressed, the process jumps to step S7.

In step S11, if the value of the address counter 8 is the end address of the memory, that is, if the memory reaches the maximum storage capacity, the index section 9b.
The start address as the operation start position information stored in step S1 is set in the address counter 8 (step S1).
7). After that, rewriting is executed from this start address, but the rewritten part is only the silent part of the voice message file currently stored. However, if the value set in the address counter 8 is used as the start address of the voice data section 9c, it can be stored in the silent section of all the voice message files, and the longest recording is possible.

Then, the voice / non-voice determination flag of the voice data portion 9c indicated by the current address counter 8 is detected (step S18). If the discrimination flag VS is 0, that is, no voice, the encoding / decoding portion is detected. 4, the voice data for one frame is encoded (step S19), and the encoded data is stored again in the voice data section 9c indicated by the current address counter 8 (step S20).

Here, the voice data to be stored is to be re-stored for both voiced and silent voices. However, if it is desired to store only voiced voices, the determination of voiced voices / voiceless voices is also performed at the time of encoding in step S19. It suffices to store only the voice in step 20. Next, it is detected whether or not the recording stop button is pressed (step S21), and if it is pressed, the current value of the address counter 8 is read by the RWEAD unit 9
It is stored in a (step S22), the end position is determined, and the recording process is exited (step S23).

On the other hand, if not pressed, step 24
It jumps to and updates the current value of the address counter 8, and then determines whether or not this value is the end address (step S25). If the value of the address counter 8 is the end address, the process jumps to step 26 and ends. If it is not the end address, the process jumps to step 18.

In step 4, the RWEAD unit 9
When a is not 0, that is, when the memory reaches the maximum storage capacity and rewriting is performed in a silent frame, RWEAD
The contents of the section 9a are copied to the address counter 8 (step S5) and the process jumps to step S24.

FIG. 3 (a) shows a voice waveform input from the microphone 1 and includes frame intervals t ₀ t ₁ and t _n-1 t _n.
Is a silent part. FIG. 3B is a diagram showing a state in which such a voice waveform is recorded in the voice memory unit 9, and shows a state before the voice memory unit 9 reaches the maximum storage capacity. In the figure, since the first frame and the n-th frame corresponding to the frame intervals t ₀ t ₁ and t _n-1 t _n are silent parts, the discrimination flag VS is set to 0, and the frame interval t ₁ t
_Since the second frame corresponding to 2 has sound, the discrimination flag VS
Is set to 1.

FIG. 4B shows the frame interval t ₀ t shown in FIG. 4A after the audio memory unit 9 reaches the maximum storage capacity.
It is a figure which shows the state which re-stored the voiced information which belongs to ₁ and t _n-1 t _n in the audio | voice memory part 9. In this case, the discrimination flag VS is kept in the same state even after being re-recorded.

Next, the operation in the reproduction mode will be described in detail with reference to the flowchart of FIG. First, the state is kept waiting until the play button is pressed (steps ST1, ST
2). When the play button is pressed, the audio data section 9
The read position of c is obtained from the operation position information of the index portion 9b, and this is set as the start address in the address counter 8 (step ST3). Then, whether or not the rewriting is performed, that is, the RWEAD unit 9a
It is detected whether the value of 0 is 0 (step ST4), 0
In the case of, as described above, it is determined that the rewriting is not performed, the encoded data is read from the address of the audio data unit 9 currently indicated by the address counter 8, and is decoded by the encoding / decoding unit 4, The decrypted data is passed to the D / A converter 10 for reproduction (step ST
5). Next, it is detected whether or not the current value of the address counter 8 is the end address (step ST6), and if it is the end address, the reproducing process is exited and the process ends (step ST8). If it is not the end address, the current address counter 8 is updated (step ST7) and the process jumps to step ST5.

In step ST4, the RWEAD unit 9a
If the value is not 0, it is determined that rewriting has been performed, so the process jumps to step ST9 to compare the current value of the address register 8 and the value of the RWEAD unit 9a.
If the condition of step ST9 is not satisfied, the area from the reproduction start address to the address of the RWEAD section 9a has been rewritten, so it is determined whether the determination flag VS is 0 (step ST10). If the determination flag VS is not 0 here, it is determined that the voiced portion is present, so the audio data stored in this area is read and reproduced. Further, when VS is 0, it is regarded as a silent portion, so the address counter 8 is updated (step ST
11), and returns to step ST9.

On the other hand, when the value of the address counter 8 satisfies the condition of step 9, there is no rewritten audio data, so that the process jumps to step ST12 and normal reproduction is performed.

Next, the address counter 8 is updated to determine whether or not the current value of the address counter 8 is the end address (steps ST13 and ST14). If it is not the end address, the process jumps to step ST9. If it is the end address, the address counter 8 uses the reproduction operation position information of the index portion 9b as the start address.
(Step ST15). Then, while updating the address counter 8, the voice data stored in the silent portion (VS = 0) is read out and reproduced (loop of step ST16, step ST18, step ST19, step ST21). Here, if the determination flag VS is not 0 in step ST18, the address counter 8 is updated and the process jumps to step ST16 (step S16).
T20).

The above-described processing is repeated until the value of the address counter 8 matches the value of the RWEAD section 9a, and when they match, the reproduction processing is exited (step ST17). As described above, the audio recording apparatus according to the first configuration of the present embodiment converts the input audio signal into a digital signal, performs the encoding on a frame-by-frame basis of the audio analysis / synthesis system, and converts it into a semiconductor memory (audio memory). Which is recorded in the section 9) and determines whether the audio information of each frame is related to voice or silence (system control section 5).
If the audio information of each of the determined frames is related to voice, the first determination information (VS = 1) is recorded, and if it is related to silence, the second determination information (VS = 0) is recorded. When the capacity of the information recording means (voice data section 9c) and the semiconductor memory reaches the maximum storage capacity, a predetermined position, for example, the start address of the currently stored voice message file or the start address position of the voice data section 9c is reached. Returning, the voice information is re-recorded in the frame to which the second discrimination information is added, and the discrimination information (system control unit 5) for holding the discrimination information in the second state and the address of the last frame re-recorded. Address value storage means (RWEAD) for storing a value
Section 9a).

Further, in the second structure of the present embodiment, in the above-mentioned first structure, the re-recording means includes means for re-recording only voiced sound. In addition, in the third configuration of the present embodiment, in the first configuration or the second configuration described above, when the frames to which the second determination information is added continue for a predetermined number or more, the re-recording unit changes the frame to the frame. It includes means for rerecording audio information. Therefore, the audio information is only for consecutive silent frames,
Alternatively, the continuous silent frames are preferentially re-recorded.

Further, in the audio reproducing apparatus according to the fourth structure of the present embodiment, the audio data is composed in units of frames, and the information indicating whether the recorded audio data is voiced or silent. The final address information relating to the re-recording is to reproduce the voice information from the semiconductor memory (the voice memory unit 9) in which the final address when the voice data is re-recorded in the silent frame of the recorded voice data is recorded. If a detection means (system control unit 5) for detecting the presence or absence of the re-recording and a final address related to the re-recording are detected, only voiced frames are sequentially reproduced for the portion related to the re-recording, and the final address of the semiconductor memory is reached. Then, a reproduction control unit (system control unit 5) is provided which sequentially reproduces the voice information re-recorded in the silent frame up to the final address related to the re-recording.

Further, in the fifth configuration of the present embodiment, in the fourth configuration described above, between the final address relating to re-recording and the final address of the memory, the reproduction control means includes the voiced frame and the silent frame. It includes means for sequentially reproducing.

Therefore, according to the first configuration, even if the semiconductor memory reaches the maximum storage capacity, new voice data can be rewritten in the silent portion, so that continuous recording can be performed without degrading the sound quality. It is possible to use the memory efficiently.

Further, according to the second configuration, since only voiced speech is re-recorded, the memory can be effectively used. Further, according to the third configuration, as compared with the case where silent frames are present separately, the connection of each frame becomes smoother, so that the sound quality at the time of reproduction is improved.

Further, according to the above-mentioned fourth structure, natural reproduction according to the recording order becomes possible. Further, according to the fifth configuration, it is possible to reproduce with high quality in the portion that is not related to re-recording.

[0035]

As described above in detail, according to the voice recording apparatus of the present invention, even if the semiconductor memory reaches the maximum storage capacity, new voice data can be rewritten in the silent portion, so that the sound quality is not deteriorated. Can be continuously recorded, and the memory can be used efficiently. Further, according to the audio reproducing device of the present invention, natural reproduction according to the recording order is possible.

[Brief description of drawings]

FIG. 1 is a block diagram of a digital recorder to which the present invention is applied.

FIG. 2 is a diagram showing a recording configuration of an audio memory unit.

FIG. 3A shows a voice waveform input from a microphone, and FIG. 3B shows a state in which such a voice waveform is recorded in the voice memory unit before the voice memory unit reaches the maximum storage capacity. FIG.

FIG. 4A is a diagram showing a voice waveform when a voice waveform rerecorded in a silent portion after the voice memory unit reaches the maximum storage capacity is added to the voice waveform shown in FIG. Yes,
(B) shows that after the voice memory unit reaches the maximum storage capacity,
It is a figure which shows the state which recorded the audio waveform in the silent part.

FIG. 5 is a flowchart showing an operation of the digital recorder according to the present embodiment in a recording mode.

FIG. 6 is a flowchart showing an operation in a reproduction mode of the digital recorder according to the present embodiment.

[Explanation of symbols]

1 ... Microphone, 2 ... Microphone amplifier, 3 ... A / D converter, 4
... encoding / decoding section, 5 ... memory control section, 6 ... operation input section, 7 ... memory control section, 8 ... address counter, 9 ... voice memory section.

Claims (5)

[Claims] 1. An input voice signal is converted into a digital signal and encoded in frame units of a voice analysis and synthesis system,
An audio recording device for recording in a semiconductor memory, wherein the audio information of each frame has a determination means for determining whether the audio information is related to voice or silence, and the audio information of each determined frame is present. Discrimination information recording means for recording the first discrimination information for sounds and second discrimination information for sounds, and returning to a predetermined position when the capacity of the semiconductor memory reaches the maximum storage capacity, and second Re-recording means for re-recording the voice information in the frame to which the discrimination information is attached, holding the discrimination information in the second state, and address value storing means for storing the address value of the last re-recorded frame, An audio recording device comprising: 2. The voice recording apparatus according to claim 1, wherein the re-recording means re-records only voiced sound. 3. The re-recording means re-records the audio information in the frame when the frame to which the second discrimination information is added continues for a predetermined number or more.
Or the voice recording device according to 2. 4. The audio data is composed on a frame-by-frame basis, and information indicating whether the recorded audio data is voiced or silenced, and the audio data is reproduced in a silent frame of the audio data once recorded. A voice reproducing device for reproducing voice information from a semiconductor memory in which a final address when recorded is recorded, the detecting means detecting presence / absence of final address information relating to re-recording, and a final address relating to re-recording being detected. In the case of the re-recording, only the voiced frames are sequentially played back, and when the final address of the semiconductor memory is reached, the audio information re-recorded in the silent frames is sequentially played back to the final address related to the re-recording. And a means for reproducing the sound. 5. The audio according to claim 4, wherein the reproduction control means sequentially reproduces a voiced frame and a silent frame between the final address related to the re-recording and the final address of the semiconductor memory. Playback device.