JPH08137492A - Speech time base converter - Google Patents

Abstract

(57) [Summary] [Purpose] When playing back an audio signal at a speed lower than the speed at which it was recorded, the audio time that can be automatically converted to a speed that is easy to listen to while minimizing the loss of data while maintaining the pitch at the time of recording. An axis conversion device is provided. [Structure] The sound reproduced at M times speed from the recording / reproducing unit 101 is converted into a digital signal at a cycle T / M by an A / D converter 102. This signal is subjected to voice / non-voice determination by the voice / non-voice determination unit 103, and only the voice / voice portion is accumulated in the buffer memory 105 by the write control unit 104. The read control unit 106 reads data from the buffer memory 105 at a cycle T, and the time axis control unit 109 performs a time axis changing process of the data. The data remaining amount monitoring unit 107 sequentially measures the data remaining amount from the current writing / reading position in the buffer memory, and based on this result, the adaptive speed control unit 108
Then, the time base conversion speed ratio is determined based on a predetermined rule according to the remaining amount of data.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice capable of arbitrarily extending the length of the time axis of the voice, which is required when the voice is played back at a low speed on a video tape recorder (VTR) or the like. The present invention relates to a time axis conversion device.

[0002]

2. Description of the Related Art Conventionally, there has been an audio time base conversion device for reproducing an audio signal at a speed different from that at the time of recording. For example, in a tape recorder, the playback speed can be increased or decreased by adjusting the tape running speed. However, as the playback speed changes, the pitch changes at the same time, making it difficult to hear the content. Therefore, there has been proposed an audio time base conversion device capable of changing the reproduction speed without changing the pitch.

A conventional speech time base converter will be described below with reference to the drawings. FIG. 8 is a block diagram showing the configuration of a conventional audio time base converter. In FIG. 8, 1 is a recording / reproducing unit for recording and reproducing an acoustic signal, 2
A converts the reproduced analog signal to a digital signal
/ D converter, 3 is a buffer memory for accumulating digital data, 4 is a D / A converter, 5 is a write controller for controlling writing of data into the buffer memory, and 6 is for reading data from the buffer memory. It is a read control unit for controlling.

The operation of the speech time base conversion apparatus configured as described above will be described below. Here, a description will be given of an audio time axis conversion device that restores a pitch to a state at the time of recording and reproduces the audio signal when the audio signal is reproduced at a recording speed or less on a recording medium.

First, the recording / reproducing unit 1 reproduces an acoustic signal at a recording speed. Here, the recording / reproducing unit is, for example, VT.
R, tape recorder, etc. Next, the acoustic signal reproduced from the recording / reproducing unit 1 is converted into a digital signal by the A / D converter 2 at a sampling period T that satisfies the sampling theorem. These digital signals that have been A / D converted,
The write control unit 5 sequentially operates the buffer memory 3 at a cycle T.
Accumulate in. Next, the digital signals accumulated in the buffer memory 3 are grouped into sections corresponding to a fixed time, and the sections are repeatedly read at least twice or more continuously. Finally, the D / A converter 4 converts the digital signal read by the read control unit 6 into an analog signal at a sampling cycle T. By this series of processes, the voice time base conversion can be realized without changing the pitch.

FIG. 9 shows an example of processing when the time-axis conversion speed ratio is 0.5. (A) shows the data at the time of recording, and (b) shows the data stored in the buffer memory. The data string of (c) is reproduced by repeating each block of (b) twice at the sampling cycle T. The data string has the same pitch as the data string of (a) and the length is 2 in the time axis direction. Is doubled.

[0007]

In the above-mentioned conventional example,
Although the pitch is the same as that at the time of recording, the speed of voice depends only on the reproduction speed from the recording medium, and only uniform reproduction speed control can be performed. Furthermore, the continuity of the voice is lost between the repeated blocks, and the quality of the reproduced sound deteriorates, making it difficult to hear. Further, when converting to a speed slower than the speed of reproduction from the recording medium, there is a possibility that the buffer memory for accumulating the input signal becomes insufficient. In that case, some data must be discarded in order to reduce the amount of data on the buffer memory, resulting in the loss of information.

The present invention is intended to solve the above-mentioned problems, and when reading from a recording medium at a reproduction speed including a recording speed and lower than that, the information loss is suppressed as much as possible and the speed is automatically made easy to listen to. An object of the present invention is to provide a voice time axis conversion device that can be changed.

[0009]

In order to solve the above problems, an audio time axis converter according to claim 1 of the present invention comprises a recording / reproducing unit for reading out an audio signal from a recording medium at a reproduction speed equal to or lower than a recording speed, An A / D converter for converting the reproduced analog signal into a digital signal, a sound / silence judgment unit for judging a sound part and a silence part in the input signal, a buffer memory for accumulating the input signal, A write control unit that controls writing of data to the buffer memory and its address, a read control unit that controls reading of data from the buffer memory and its address, and the current write address and read address Data remaining amount monitoring unit that measures the remaining amount of data in the , A time axis control section for expanding the audio time axis according to the time axis conversion speed ratio determined by the adaptive speed control section, and D / for converting the data after time axis conversion into an analog signal. And an A converter.

An audio time base conversion device according to a second aspect of the present invention comprises an adaptive speed control unit which determines a time base conversion speed ratio in proportion to the remaining amount of data.

According to another aspect of the present invention, there is provided an audio time base conversion device including an adaptive speed control unit for determining a time base conversion speed ratio based on a conversion rule that nonlinearly corresponds to the remaining amount of data.

According to a fourth aspect of the present invention, there is provided an audio time axis conversion apparatus in which sections A and B which include the same number of consecutive sample values and are adjacent to each other in the order of A and B and a weighting coefficient which increases monotonically are section A.
The crossfade processing section X obtained by adding the value obtained by multiplying the sample value of
Output from the adaptive speed control unit by alternately outputting the sections D and C while adjusting the length of the section D connected in the order of A-X-B and the length of the unprocessed section C that directly outputs the data. The time axis control unit adjusts the time axis according to the time axis conversion speed ratio.

According to the fifth aspect and the sixth aspect of the present invention, when the remaining amount of data is less than a certain value, all the data are accumulated in the buffer memory, and when the remaining amount is more than the certain value, only the voiced section is stored. A write control unit for accumulating in the buffer memory is provided.

[0014]

According to the structure of claim 1, only the section which is determined to be voiced as a result of the voiced / non-voiced determination is accumulated in the buffer memory, and the remaining amount of data is measured from the write and read address positions. Depending on the remaining amount, the speed can be automatically adjusted according to the amount of silence, such as when the remaining amount is high, the speed is as close as possible to the playback speed, and as the remaining amount decreases, the speed gradually decreases. It is possible to reproduce the sound part as late as possible. Even when playing at the same speed as the recording speed, if you control the speed after time axis conversion to be lower than the recording speed according to the amount of silence, you can hear fast-speaking voices or foreign language sources originally. It can also be made easy.

According to the second aspect of the present invention, since the remaining amount of data and the time-axis conversion speed ratio change in proportion to each other, it is possible to prevent a rapid change in the speed of the voice and make the voice easy to hear.

According to the third aspect of the present invention, if a non-linear conversion rule that stays at a low speed for a long time is set,
Due to the effect of suppressing the time-axis conversion speed ratio to be low, it is possible to stably reproduce the sound on the low speed side, and it is possible to realize an audio time-axis conversion device that is very easy to hear.

According to the fourth aspect of the invention, the time axis conversion can be performed at an arbitrary time axis conversion speed ratio by adjusting the lengths of the crossfade processing section length and the non-processing section length. Furthermore, if the section length for crossfade processing is set to a fixed length, it is possible to obtain an arbitrary time-axis conversion speed ratio simply by adjusting the unprocessed section length that directly outputs the data accumulated in the buffer memory. It is possible to realize a voice time base conversion device that can convert to an arbitrary speed with a very simple configuration.

According to the fifth and sixth aspects, at the time when the remaining amount of data is small, all the data is stored in the buffer memory regardless of the sound / non-sound determination result of the input signal, so that the data to be read out. It is possible to obtain a natural reproduced sound with little discomfort, since the reproduced sound does not become silent (mute state) in the middle.

[0019]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of a speech time base conversion apparatus according to the first embodiment of the present invention. In FIG. 1, 101 is a recording / reproducing unit for recording and reproducing an acoustic signal, 102 is an A / D converter for converting an analog signal reproduced by the recording / reproducing unit 1 into a digital signal,
Reference numeral 103 denotes a sound / sound determination unit that determines whether the converted digital signal sequence is a voiced portion or a silence portion, and 104 is a buffer memory 105 for writing the signal sequence determined to be the voiced portion.
A writing control unit for controlling writing to the upper side and its address, 106 is a reading control unit for controlling reading of data stored in the buffer memory 105 and its address, and 107 is a present writing address and reading on the buffer memory 105. A data remaining amount monitoring unit that measures the data remaining amount from the positional relationship of addresses, 108 is an adaptive speed control unit that determines the time axis conversion speed ratio from the data remaining amount obtained by the data remaining amount monitoring unit, and 109 is an adaptive speed control A time axis control unit that performs time axis conversion processing according to the time axis conversion speed ratio determined by the unit, and 110 is a D / A converter that converts time axis converted digital data into an analog signal.

The operation of the speech time base conversion apparatus configured as described above will be described in detail below with reference to FIG.

First, from the recording / reproducing unit 101, M at the time of recording
The acoustic signal is read at a speed of (≦ 1) times. Hereinafter, the time-axis conversion speed ratio will represent the relative speed to the recording speed. Here, assuming that the sampling period at the time of recording in the recording / reproducing unit 101 is T, the acoustic signal reproduced by the recording / reproducing unit 101 at M times speed is successively digital signal sequence at the sampling period T / M by the A / D converter 102. Is converted to. The voiced / unvoiced determination unit 103 determines whether the sample value sequence is a voiced part or a silent part.

The determination as to whether there is a sound part or a silent part is made as follows, for example. If the sample value sequence of the digital signal is Si,
When (Numerical formula 1) is satisfied for the N sample value series, it is determined that the sample value sequence is voiced, and when it is not satisfied, it is silent. Here, Pth is a predetermined threshold value for determining whether or not there is sound.

[0023]

[Equation 1]

Here, a pointer (hereinafter, referred to as an address indicating an address at which data is to be stored next on the buffer memory 105)
If it is determined that there is sound in the above (Equation 1), the write control unit 10
4, the sample value series is sequentially stored at the address position of the buffer memory 105 indicated by the write pointer Pw, and Pw is incremented each time. On the contrary, when it is determined that there is no sound, the write control unit 104 stops the data writing operation to the buffer memory 105.
As a result, only the sound part data is stored in the buffer memory 105.

It should be noted that here, it is determined that the sound is present when the expression (1) is satisfied, and the sound is not satisfied when the expression (1) is satisfied. However, it is determined that there is silence before or after the section satisfying the expression (1). A short sample value sequence may be included in the voiced section.

The read control unit 106 sequentially reads the data in the buffer memory 105 at a cycle T and sends it to the time axis control unit 109. Here, a pointer indicating the address of the data on the buffer memory 105 to be read next (hereinafter,
Let Pr be the write pointer), the data remaining amount monitoring unit 107 determines the buffer memory 105 based on the positional relationship between the write pointer Pw and the read pointer Pr.
The remaining amount of the above unprocessed data is sequentially measured.

FIG. 2 is an explanatory diagram showing a method of measuring the remaining amount of data, and FIG. 2 (a) shows the positional relationship between two pointers.
Two cases of (b) can be considered. In FIG. 2, the start address of the buffer memory is a ₀ , and the end address is a
_{If n-1} (however, a _n-1 > a ₀ ), the remaining data amount Z that has not been read is shown by the shaded portion in FIG. 2 and can be calculated by the following equation.

[0028]

[Equation 2]

This is equivalent to treating the buffer memory 105 as a so-called cyclic memory. Here, in order to read and output data from the memory, the write pointer Pw must precede the read pointer Pr on the cyclic memory, so Pw and Pr
If they overlap (Pw = Pr), the read control unit 106 stops the read operation and the read pointer Pr
Maintains the address value at that time. Where Pw and P
There are two possible overlapping states of r: one in which Pr catches up with Pw from the state in FIG. 2A, and one in which Pw catches up with Pr from the state in FIG. 2B. The former is a case where the amount of unprocessed data becomes 0, and the latter is a case where the amount of unprocessed data is equal to the capacity of the buffer memory 105. In the former case, unprocessed data is completely lost, but in the latter case, unprocessed data fills the buffer memory 105 and remains. However, also in the latter case, all of these unprocessed data are discarded and the remaining amount of data is reset to zero.

The adaptive speed control unit 108 sets the time-axis conversion speed ratio based on the value of the data remaining amount Z obtained by the data remaining amount monitoring unit 107. That is, when the remaining amount of data is large, the speed after the time axis conversion is set to a speed as close as possible to the reproduction speed, and when the remaining amount of data is small, the speed is as slow as possible so that the read pointer Pr cannot catch up with the write pointer Pw. Set the time axis conversion speed ratio to.

Hereinafter, the operation of the adaptive speed control unit 108 will be described in the case where the recording / reproducing unit 101 reproduces at the same speed as the recording speed. Here, the maximum value of the speed after the time axis conversion is the same as the recording speed, and the minimum value is V _o . FIG. 3 shows the relationship between the remaining amount of data and the time-axis conversion speed ratio for the remaining amount, which is the rule for setting the time-axis conversion speed ratio. FIG. 3A shows a rule for linearly associating the remaining amount of data with the time-axis conversion speed ratio. In this case, the time-axis conversion speed ratio V can be calculated by the following equation.

[0032]

(Equation 3)

FIG. 3 (b) shows an example of a rule for non-linearly associating the remaining amount of data with the time-axis conversion speed ratio.
Here, if it is made to correspond to a quadratic curve, the time axis conversion speed ratio V can be calculated by the following equation.

[0034]

[Equation 4]

In the case of FIG. 3 (a), the time-axis conversion speed ratio can be changed smoothly when the remaining amount of data is increased or decreased, while in the case of FIG. 3 (b), a certain amount of data is stored in the buffer memory. It is characterized in that it can be stabilized at the slowest speed until it is stored in 105.
FIG. 3C is an example of the case where the non-linear correspondence is defined in a stepwise manner, and the time-axis conversion speed ratio V can be calculated by the following equation.

[0036]

(Equation 5)

According to the rule shown in FIG. 3C, it is possible to realize almost the same control as the rule of FIG. 3B with a smaller amount of calculation and a circuit scale.

As described above, by determining the time-axis conversion speed ratio based on the correspondence rule of FIG. 3, even for a signal reproduced at the same speed as the recording speed, for an input signal having a silent portion to a certain extent or more. , It is possible to set the speed of the voiced part to a speed slower than the playback speed, and when the signal without the silent part continues, the maximum time-axis conversion speed ratio (equal to the recording speed here) Speed, that is,
Since 1) is set, there is no loss of information.

Although the maximum value of the time-axis conversion speed ratio is 1, that is, the recording speed, the same rule can be set even if the maximum value is smaller than 1 (for example, 0.8). However, in this case, since the time axis conversion speed ratio of less than 0.8 is set at the maximum even for the signal reproduced at the recording speed when the sound continues, the buffer memory becomes full and the data is written. In some cases, it may be necessary to discard some of the data without being able to. This corresponds to the case where Pw catches up with Pr in FIG. 2B, for example. In this case, by resetting the remaining data amount to 0 as described above, the buffer memory accumulated so far is The data for the capacity will be discarded.

For example, if the capacity of the buffer memory is 256K
If the bit and sampling frequency are 10 kHz and one sample is treated as 8-bit data, 32 K points (about 3.2 seconds) of voiced data is discarded.

By thus setting the maximum value of the speed after the time axis conversion to be low, a part of the data is lost depending on the amount of silence, but the entire data is reproduced at a slower and easy-to-listen speed. It becomes possible to do.

The value of the time axis conversion speed ratio V determined by the adaptive speed control unit 108 is sent to the time axis control unit 109, and the time axis conversion unit 109 converts the time axis according to the time axis conversion speed ratio V. Processing is performed.

FIG. 4 is a block diagram showing a detailed structure of the time axis control unit 109. In FIG. 4, reference numeral 401 is a control circuit that performs overall control, 402 is a switching circuit that switches between cross-fade processing or non-processing that performs weighted addition according to a command from the control circuit, 403 is a latch circuit that temporarily holds data, Reference numeral 404 is a crossfade circuit that performs weighted addition processing, and the other blocks are the same as the blocks of the same name in FIG. Less than,
The operation of the time axis control unit 109 will be described with reference to FIG.

First, the control circuit 401 determines the crossfade section length K and the unprocessed section length S in order to realize the time-axis conversion speed ratio V. Here, the crossfade section length K is a fixed value, but it is of course possible to change the value of K according to the time-axis conversion speed ratio V.

FIG. 5 is a schematic diagram for explaining the time axis conversion processing. FIG. 5 (a) is data before processing, and FIG. 5 (b).
Indicates data after processing. Further, the portion of the length K sandwiched between the data string A and the data string B in FIG. 5B is a data string obtained by crossfading the data string A and the data string B.

Here, in order to realize the time-axis conversion speed ratio V, the total length of the unprocessed data strings A, B and C (2
1 / V of (K + S) is the data string length (3K +
S may be set so as to be S). Therefore, the unprocessed section length S is calculated by the following equation.

[0047]

(Equation 6)

Now, assuming that the read pointer Pr points to the beginning of the data string A in FIG. 5A, the crossfade processing will be described. The crossfade process is composed of three processes.

First, the first process will be described. FIG. 6 is a flowchart showing a part of the crossfade processing. First, the control circuit 401 switches the switching circuit 402 to the non-processing side while referring to the time-axis conversion speed ratio V (S11). Next, the read control unit 106 is given a pointer P.
An instruction is issued to read the data indicated by r (S12).
Then, the read data is directly input to the D / A converter 110 without any processing (S13). Finally, the read pointer Pr is incremented (S14). Then, the above operation is repeated until all the data strings A are processed (S15).

Next, the second processing will be described. First,
The control circuit 401 commands the read control unit 106 so that the read pointer Pr points to the beginning of the data string A. Next, the control circuit 401 switches the switching circuit 402 before the crossfade process, and instructs the read control unit 106 to read the data indicated by the read pointer Pr. The read data is input to and held in the latch circuit 403. Next, the control circuit 401 commands the read control unit 106 to read the data indicated by the (Pr + K) address of the K sample destination. The read data is directly input to the crossfade circuit 404. Weighted addition is performed using this data and the two data held in the latch circuit 403 immediately before. Here, the data sequence A in FIG. 5A is d ₀ , d ₁ , ..., D _K-1 , and the data sequence B is d _K , d _{K + 1} , ..., d _2K-1 . . Also, the weighting function that increases monotonically is w ₁ (t) (where 0 ≦ w ₁ (t) ≦ 1, t
= 0,1, ..., K-1), the function that decreases monotonically is w ₂ (t) = 1-w ₁ (t)
Then, the value Ct after weighted addition is obtained by the following equation.

[0051]

(Equation 7)

After that, the read pointer Pr is incremented. The control circuit 401 continuously performs the same process as the above K times. After all the cross-fading processing of the data sequence A and the data sequence B of FIG. 5A is completed, the read pointer P
It is set to the value of r + K.

Next, the third processing will be described. At the time when the second process is completed, the read pointer Pr points to the beginning of the data string B, so that the same process as the process for the data string A in the first process is performed on the data string B as it is. Specifically, the control circuit 401 switches the switching circuit 402 to the non-processing side, and the read control unit 10
6 is instructed to read the data indicated by the pointer Pr. The read data is directly processed by the D / A converter 1 without any processing.
Input to 10. Finally, the read pointer Pr is incremented. Until all data string B is processed,
The above operation is repeated.

When the cross-fade process is completed as a result of these three processes, the control circuit 401 switches the switching circuit 402 to the non-processing side, and the data of the number corresponding to the length S determined in (Equation 6) is buffered. It is read from the memory 105 and directly input to the D / A converter 110.

After that, the time axis conversion at the time axis conversion speed ratio V can be realized by repeating alternately outputting the cross-faded data of 3K in length and the unprocessed data of length S. When the time-axis conversion speed ratio set by the adaptive speed control unit 108 is changed at a certain time point, the non-processed section length is changed by (Equation 6), and the same process as above is continued thereafter, so that Change the time axis conversion speed ratio.

The data string thus time-axis converted is finally converted into an analog signal by the D / A converter 110 in the cycle T. As a result, it is possible to obtain an audio signal in which the pitch is the same as that at the time of recording and the speed is adaptively switched at the reproduction speed M or less.

As described above, according to the first embodiment described above, the voice / sound determination unit 103, the data remaining amount monitoring unit 107 for measuring the data remaining amount from the positional relationship between the write pointer and the read pointer, and the data An adaptive speed control unit 108 that determines the time-axis conversion speed ratio according to the remaining amount is provided, and the speed after time-axis conversion is gradually increased when the data remaining amount is large, and gradually when the data remaining amount is small. By controlling so as to be slow, it is possible to listen to a low-speed reproduced audio signal at a slow speed equal to or lower than the reproduction speed according to the amount of silence contained therein, and with almost no loss of information. It will be possible.

Further, by providing the time axis control unit 109 capable of converting the time axis to a desired speed by adjusting the crossfade section length and the unprocessed section length, high quality time axis conversion can be performed. realizable. Furthermore, if the crossfade section length is fixed to a preset value, the time axis conversion can be achieved at any speed by changing only the unprocessed section length, and a voice time axis conversion device is realized with a very simple configuration. can do.

Particularly, in a recording / reproducing section accompanied by an image such as a VTR, for example, the image is reproduced at 1 × speed and only the sound is reproduced as 1
Since it can be played back at a slow speed of double speed or less, it has a great effect of improving the ease of listening.

Next, the second embodiment of the voice time base conversion device of the present invention
An example will be described. The basic configuration of the audio time axis converter of the present embodiment is the same as the block diagram shown in FIG. 1, and only the operation of the write controller 104 is different.

FIG. 7 is a flow chart showing the operation of the write controller of this embodiment. Hereinafter, only the operation of the write control unit will be described with reference to FIG. 7. The write control unit 104 sequentially takes in the value of the remaining data amount Z measured by the remaining data amount monitoring unit 107 (S1) and compares it with a preset threshold value Zth (S2). If Z is larger than Zth, there is a large amount of remaining data.
Based on the result of the sound / sound determination unit 103, it is determined whether or not the current input data is sound / sound (S3), and only when there is sound, the data is written in the buffer memory 105 (S4) and the write pointer Pw is incremented. (S5). S
If Z is smaller than Zth in 2, the remaining amount of data is small, so that the data is written to the buffer memory 105 regardless of the sound / sound determination result, and the write pointer P
Increment w.

In the series of processes described above, in the case of a signal containing a lot of silence, control is performed so that the read pointer Pr does not catch up with the write pointer Pw in FIG. 2A, that is, the remaining amount of data does not become zero. ing.

As described above, according to the above-described second embodiment, when the remaining amount of data is less than the predetermined fixed amount, the write control unit for accumulating all the data in the buffer memory is provided, so that the voice sound is generated. When a signal with few sections is reproduced, it is possible to prevent the reproduction sound from being interrupted (mute state) without the data remaining amount becoming 0, and to realize a sound time axis conversion device capable of natural reproduction with less discomfort. be able to.

[0064]

As described above, in the present invention, the recording speed M
Only the voiced part is stored in the buffer memory by the voiced / non-voiced determination from the audio signal reproduced at (≦ 1) times, and the remaining amount of data stored in the buffer memory is measured from the write and read address positions. Depending on the result, it is possible to automatically adjust the speed after time axis conversion according to the amount of silence so that it becomes as slow as possible when the remaining amount is low and gradually approaches the playback speed as the remaining amount increases. As a result, during low speed playback, the pitch is the same as during recording,
Moreover, it is possible to provide an audio time axis conversion device capable of reproducing audio at a speed slower than the reproduction speed and easy to listen to. As a result, the sound reproduced at the recording speed can be further slowed down, and it is possible to easily listen to fast-paced words and foreign language sources.

By adjusting the crossfade section length and the unprocessed section length to provide a time axis control unit for converting the time axis to a desired speed, time axis conversion processing at an arbitrary speed can be realized. Moreover, if the crossfade length is fixed, the time axis conversion can be realized at any speed simply by adjusting the length of the unprocessed section, so the time axis converted high quality playback sound can be achieved with a simple circuit configuration. Can be provided.

Further, when the remaining amount of data is small, all the data is stored in the buffer memory regardless of whether the sound is present or not, so that the data to be output when the signal with a small number of voiced sections is reproduced. It is possible to provide a sound time axis conversion device that outputs a natural reproduced sound with little discomfort, without being muted during reproduction.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an audio time base converter according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram of a method for measuring a remaining memory amount according to the present embodiment.

FIG. 3 is an explanatory diagram of a speed setting method of an adaptive speed control unit according to the present embodiment.

FIG. 4 is a circuit diagram of a time axis control unit according to the present embodiment.

FIG. 5 is a principle diagram showing an operation of a time axis control unit according to the present embodiment.

FIG. 6 is a flowchart showing a part of the operation of the time axis control unit according to the present embodiment.

FIG. 7 is a flowchart showing the operation of the write control unit according to the second embodiment of the present invention.

FIG. 8 is a block diagram of a conventional audio time base converter.

FIG. 9 is a principle diagram showing a data processing method of a conventional audio time base converter.

[Explanation of symbols]

 Reference numeral 101 recording / reproducing unit 102 A / D converter 103 sound / silence determination unit 104 writing control unit 105 buffer memory 106 reading control unit 107 remaining data amount monitoring unit 108 adaptive speed control unit 109 time axis control unit 110 D / A converter

 ─────────────────────────────────────────────────── ─── Continuation of front page (72) Inventor Koji Watanabe 1006 Kadoma, Kadoma City, Osaka Prefecture Matsushita Electric Industrial Co., Ltd. (72) Norikazu Ueno, 1006 Kadoma, Kadoma City, Osaka Matsushita Electric Industrial Co., Ltd.

Claims (6)

[Claims] 1. A recording / reproducing unit for reading an acoustic signal stored in a recording medium at a speed of M (≦ 1) times as high as that at the time of recording, and an analog signal read by the recording / reproducing unit is converted into a digital signal. An A / D converter, a sound / silence judging section for judging a sound part and a silent part of an input signal, a buffer memory for accumulating a digital signal output from the A / D converter, and the sound part. The write control unit that controls the write address so that only the data in the section that is determined to be sound by the silence determination unit is written in the buffer memory, the read control unit that reads the data in the buffer memory, and the last in the write control unit. By monitoring the positional relationship between the address on the buffer memory written in the buffer memory and the address on the buffer memory last read by the read controller. A data remaining amount monitoring unit that measures the amount of unprocessed data in the buffer memory, and a speed after time-axis conversion with respect to the recording speed according to a predetermined rule according to the remaining amount of data obtained from the data remaining amount monitoring unit An adaptive speed control unit that determines the ratio of the time, a time axis control unit that extends the time axis of the voice according to the time axis conversion speed ratio determined by the adaptive speed control unit, and a process obtained by the time axis control unit. An audio time base converter comprising a D / A converter for converting data into an analog signal. 2. The adaptive speed control unit sets the maximum value and the minimum value of the converted speed equal to or less than the reproduction speed M from the recording medium, and determines the time-axis conversion speed ratio in proportion to the remaining amount of data. The audio time base conversion device according to claim 1, wherein 3. The adaptive speed control unit sets the maximum value and the minimum value of the converted speed to be equal to or less than the reproduction speed M from the recording medium, and sets the time based on the conversion rule nonlinearly associated with the remaining amount of data. The audio time axis converter according to claim 1, wherein the axis conversion speed ratio is determined. 4. The time axis control unit includes sections A and B that include the same number of consecutive sample values and are adjacent to each other in the order of A and B.
And a value obtained by multiplying the sample value in the section A by a monotonically increasing weight coefficient and a value obtained by multiplying the sample value in the section B by a monotonically decreasing weight coefficient, a crossfade processing section X By adjusting the length of the section D connected in the order of A-X-B and the length of the unprocessed section C for directly outputting the data, and outputting the sections D and C alternately. The audio time axis converter according to any one of claims 1, 2 and 3, wherein the time axis is adjusted in accordance with the time axis conversion speed ratio output from. 5. When the data remaining amount monitoring unit determines that the remaining amount of data is equal to or less than a certain amount, the writing control unit writes all the data regardless of the determination result of the sound / silence determination unit. When the remaining amount of data is accumulated in the buffer memory and the remaining amount of data in the data remaining amount monitoring unit is equal to or more than a certain amount, the write address is controlled so that only the data which is determined to be voiced by the voice / non-voice determination unit is stored in the buffer memory. The audio time base conversion device according to claim 1, characterized in that. 6. When the data remaining amount monitoring unit determines that the remaining amount of data is equal to or less than a certain amount, the writing control unit writes all the data regardless of the determination result of the sound / silence determination unit. When the remaining amount of data is accumulated in the buffer memory and the remaining amount of data in the data remaining amount monitoring unit is equal to or more than a certain amount, the write address is controlled so that only the data which is determined to be voiced by the voice / non-voice determination unit is stored in the buffer memory. The audio time base conversion device according to claim 4, wherein.