JP3081469B2 - Speech speed converter - Google Patents

Abstract

PURPOSE: To reduce a processing load, to reduce deviation between a video and an audio and to prevent increase in memory capacity by speech speed conversion-processing an input audio signal with a speech speed conversion- processing means. CONSTITUTION: Since a ring memory 7 becomes easy to become an overflow just before state the more in a program whose vocalizing speed is faster, a compression rate is decided so that an audio reproducing speed becomes close to a double speed. On the contrary, the compression rate is decided so that the audio reproducing speed becomes close to a one time speed the more in the program whose vocalizing speed is later. Thus, the audio reproducing speed becomes the speed of the double speed or below and according to the original vocalizing speed, and a natural reproducing audio is obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a speech speed converter for changing the speech speed of an audio signal, and for example, to listen to the audio of a laser disk, a VTR, a TV, a TV telephone, or a TV conference system accompanied by video at a high or low speed. Audio reproduction device, radio with a hearing aid that converts audio signals into slow, easy-to-hear audio, telephones, hearing aids, or tape recorders, stereo systems, CD players, audio guidance systems that perform early or late audio listening The present invention relates to a speech speed conversion device used for an English language learner or the like that converts English voice spoken at native speed into slow and easy-to-hear voice.

[0002]

2. Description of the Related Art As a conventional technique for converting a speech speed, there is an analog time axis expansion / compression technique. However,
In the speech speed conversion method using the analog time-base expansion / compression technology, since simple speech waveform thinning or repeated insertion of speech waveforms are merely performed, the sound connection becomes discontinuous, resulting in poor sound quality. There is a problem of getting worse.

[0003] As a time axis expansion / compression technique for speech that provides good sound quality, there is a technique of detecting a pitch cycle of speech by digital signal processing and thinning or inserting a pitch portion in units of the detected pitch cycle. However, in the speech speed conversion method using the digital time-base expansion / compression technique, the audio signal is compressed or expanded at a uniform compression / expansion rate regardless of the silent section and the audio section of the audio signal. When a VTR is played back at double speed, when an English language learner plays English voice, or the like, there is a problem that the playback speed of the voice section becomes too fast and the voice cannot be heard.

[0004]

In order to solve the above problem, a speech speed conversion method for identifying a silent section of a voice signal and a voice section, deleting the silent section, and extending the voice section in units of a pitch period has been proposed. It has already been developed (Ref. A (hereinafter referred to as
The first conventional method): IETF SP92-56, HC
92-33 (1992-09) Title "One Method for Absorbing Time Elongation Associated with Speech Speed Conversion" Published by The Institute of Electronics, Information and Communication Engineers, Reference B (hereinafter referred to as "second conventional method"): IEICE Tech. -150 (1993-0)
3) Title "Evaluation of speech rate conversion method by hearing impaired"
Published by The Institute of Electronics, Information and Communication Engineers). According to this method, the reproduction speed of the voice section can be reduced, and the voice can be easily heard. However, this method has the following problems.

[0005] In the first conventional method, a high processing load is required due to a large processing load, and power consumption is increased. In the second conventional method, the difference between video and audio becomes too large to make it difficult to grasp the contents, and the capacity of the memory for storing the audio signal becomes enormous, resulting in high cost.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a speech speed conversion device which can reduce the processing load, reduce the difference between video and audio, and does not require a huge memory capacity for storing audio signals. .

It is another object of the present invention to reduce the voice reproduction speed for a voice in a voice section with respect to a set reproduction speed magnification while minimizing the lack of a voice in a voice section of an input signal. It is an object of the present invention to provide a speech speed conversion device capable of performing the conversion.

[0008]

According to a first aspect of the present invention, there is provided a first speech speed conversion device for performing speech speed conversion processing of an input voice signal, a ring memory in which an output of the speech speed conversion processing device is written, and Means for reading out data from the ring memory at a constant speed, wherein the speech speed conversion processing means is configured such that when the input voice signal is a voice section and the ring memory is not in a state immediately before overflow, the set reproduction speed magnification is set to n and 1 / n or more. Means for performing compression / expansion processing on an input audio signal at a compression rate determined according to a program type set by an operator.

The second speech speed conversion device according to the present invention outputs A / D conversion means for sampling an input analog audio signal at a sampling frequency corresponding to a set reproduction speed magnification, and outputs the A / D conversion means. A speech memory, to which a required number of speech signals are inputted into the frame memory, speech rate conversion processing means for performing speech rate conversion processing on those speech signals, and speech rate conversion processing means. A ring memory to which an output is written, a reading means for reading data from the ring memory based on a read signal having a frequency equal to a sampling frequency at the time of 1 × speed reproduction,
And a storage amount calculation means for calculating the storage amount of the ring memory based on the write signal and the read signal of the ring memory, and the speech speed conversion processing means converts the required number of audio signals input to the frame memory into According to the section discriminating means for discriminating whether the corresponding input voice is a voice section or a silence section, and the compression and decompression processing or deletion of the required number of voice signals according to the output of the section discriminating means and the output of the accumulation amount calculating means. Signal processing means for performing processing, wherein when the input voice is a voice section and the ring memory is not in a state immediately before an overflow, the compression rate is 1 / n or more, where n is a set reproduction speed magnification, It is characterized by including means for performing a compression / expansion process at a compression ratio determined according to a program type set by an operator.

In a third speech speed converter according to the present invention, a required number of audio signals are input to a frame memory in which an input digital audio signal is written at a speed corresponding to a set reproduction speed magnification. Speech rate conversion processing means for performing speech rate conversion processing on these voice signals, a ring memory in which the output of the speech rate conversion processing means is written, reading means for reading data from the ring memory at a constant speed, and a ring. A storage amount calculating unit for calculating the storage amount of the ring memory based on the write signal and the read signal of the memory is provided, and the speech speed conversion processing unit corresponds to the required number of audio signals input to the frame memory. In accordance with the section discriminating means for discriminating whether the input voice is a voice section or a silent section, and the output of the section discriminating means and the output of the accumulation amount calculating means, A signal processing means for performing compression and expansion processing or deletion processing with respect to the required number of audio signals,
When the input voice is in the voice section and the ring memory is not in the state immediately before the overflow, the signal processing means has a compression ratio of 1 / n or more, where n is the set reproduction speed magnification, and the program type set by the operator is It is characterized in that it includes means for performing compression / decompression processing at a compression ratio determined accordingly.

The above-mentioned ring memory has a ring structure (ring
structure). The ring structure refers to a structure in which the pointer of the last item of the chained list points to the first item.

The signal processing means of the second or third speech speed conversion device includes, for example, based on the output of the section discriminating means and the output of the storage amount calculating means, (1) the input voice is a voice section and A first mode in which the ring memory is not in the state immediately before the overflow; (2) a second mode in which the input voice is in the voice section and the ring memory is in the state immediately before the overflow; and (3) a continuous section in which the input voice is the silent section and the silent section. Is less than a predetermined silence deletion start point determination value,
Third, the ring memory is not in the state immediately before overflow
(4) a fourth mode in which the input voice is a silent section, the duration of the silent section is less than a predetermined silent deletion start point determination value, and the ring memory is in a state immediately before overflow; A fifth mode in which the duration of the silent section is longer than or equal to a predetermined silent deletion start point discrimination value and the ring memory is not in a state immediately before an underflow; and (6) the input sound is a silent section and a silent section. Is greater than or equal to a predetermined silence deletion start point determination value,
And the ring memory is in the state immediately before underflow.
A mode discriminating unit for discriminating which of the modes, when the mode is determined to be the first mode or the third mode, the set reproduction speed magnification is n and the compression ratio is 1 / n or more; Processing means for performing compression / expansion processing at a compression ratio determined according to the program type set by the user, when it is determined that the second mode or the fourth mode has been selected,
A second processing unit that deletes an audio signal until the storage amount of the ring memory becomes a state immediately before an underflow; a third processing unit that deletes an audio signal in a silent section when the fifth mode is determined; and a sixth processing unit. A fourth processing means is provided for performing compression / expansion processing at a compression ratio of 1 / n ± α (where α is a value of 0 or more and 1 or less), where n is a set reproduction speed magnification when the mode is determined. Things are used.

[0013] The silence deletion start point discrimination value may be adjusted according to the storage amount of the ring memory.

According to a fourth aspect of the present invention, there is provided a speech rate conversion device for converting an input voice signal into a speech rate, a ring memory in which the output of the speech rate conversion means is written, and a constant speed from the ring memory. Means for reading data, wherein the speech rate conversion processing means has a compression rate of 1 / n or more, where n is a set reproduction speed magnification, when the input audio signal is in a voice section and the ring memory is not in a state immediately before overflow. Means for performing compression / expansion processing on an input audio signal at a compression rate determined according to a program type set by an operator and a storage amount of a ring memory.

A fifth speech speed conversion device according to the present invention outputs A / D conversion means for sampling an input analog audio signal at a sampling frequency corresponding to a set reproduction speed magnification, and outputs the A / D conversion means. A speech memory, to which a required number of speech signals are inputted into the frame memory, speech rate conversion processing means for performing speech rate conversion processing on those speech signals, and speech rate conversion processing means. A ring memory to which an output is written, a reading means for reading data from the ring memory based on a read signal having a frequency equal to a sampling frequency at the time of 1 × speed reproduction,
And a storage amount calculation means for calculating the storage amount of the ring memory based on the write signal and the read signal of the ring memory, and the speech speed conversion processing means converts the required number of audio signals input to the frame memory into According to the section discriminating means for discriminating whether the corresponding input voice is a voice section or a silence section, and the compression and decompression processing or deletion of the required number of voice signals according to the output of the section discriminating means and the output of the accumulation amount calculating means. Signal processing means for performing processing, wherein when the input voice is a voice section and the ring memory is not in a state immediately before an overflow, the compression rate is 1 / n or more, where n is a set reproduction speed magnification, It includes a means for performing compression / expansion processing at a compression rate determined according to the program type set by the operator and the amount of storage in the ring memory. To.

In a sixth speech speed converter according to the present invention, a required number of audio signals are input to a frame memory in which an input digital audio signal is written at a speed corresponding to a set reproduction speed magnification. Speech rate conversion processing means for performing speech rate conversion processing on these voice signals, a ring memory in which the output of the speech rate conversion processing means is written, reading means for reading data from the ring memory at a constant speed, and a ring. A storage amount calculating unit for calculating the storage amount of the ring memory based on the write signal and the read signal of the memory is provided, and the speech speed conversion processing unit corresponds to the required number of audio signals input to the frame memory. In accordance with the section discriminating means for discriminating whether the input voice is a voice section or a silent section, and the output of the section discriminating means and the output of the accumulation amount calculating means, A signal processing means for performing compression and expansion processing or deletion processing with respect to the required number of audio signals,
The signal processing means, when the input voice is a voice section and the ring memory is not in a state immediately before overflow, the set reproduction speed magnification is n and the compression rate is 1 / n or more, and the program type and It is characterized in that it includes means for performing compression / expansion processing at a compression ratio determined according to the storage amount of the ring memory.

The signal processing means of the fifth or sixth speech speed conversion device includes, for example, based on the output of the section discriminating means and the output of the accumulation amount calculating means, (1) the input voice is a voice section and A first mode in which the ring memory is not in the state immediately before the overflow; (2) a second mode in which the input voice is in the voice section and the ring memory is in the state immediately before the overflow; and (3) a continuous section in which the input voice is the silent section and the silent section. Is less than a predetermined silence deletion start point determination value,
Third, the ring memory is not in the state immediately before overflow
(4) a fourth mode in which the input voice is a silent section, the duration of the silent section is less than a predetermined silent deletion start point determination value, and the ring memory is in a state immediately before overflow; A fifth mode in which the duration of the silent section is longer than or equal to a predetermined silent deletion start point discrimination value and the ring memory is not in a state immediately before an underflow; and (6) the input sound is a silent section and a silent section. Is greater than or equal to a predetermined silence deletion start point determination value,
And the ring memory is in the state immediately before underflow.
A mode discriminating unit for discriminating which of the modes, when the mode is determined to be the first mode or the third mode, the set reproduction speed magnification is n and the compression ratio is 1 / n or more; Processing means for performing compression / expansion processing at a compression rate determined according to the program type and the storage amount of the ring memory set by the user, the second mode or the fourth mode
A second processing unit that deletes an audio signal until the storage amount of the ring memory becomes a state immediately before an underflow when the mode is determined; and a second processing unit that deletes an audio signal in a silent section when the mode is determined. (3) processing means, and when the mode is determined to be the sixth mode, the compression / expansion processing is performed at a compression ratio of 1 / n ± α (α is a value of 0 or more and 1 or less), where n is a set reproduction speed magnification. The one provided with the fourth processing means is used.

The silence deletion start point discrimination value may be adjusted according to the storage amount of the ring memory.

A seventh speech speed conversion device according to the present invention outputs A / D conversion means for sampling an input analog audio signal at a sampling frequency corresponding to a set reproduction speed magnification, and outputs the A / D conversion means. A speech memory, to which a required number of speech signals are inputted into the frame memory, speech rate conversion processing means for performing speech rate conversion processing on those speech signals, and speech rate conversion processing means. A ring memory to which an output is written, a reading means for reading data from the ring memory based on a read signal having a frequency equal to a sampling frequency at the time of 1 × speed reproduction,
And a storage amount calculation means for calculating the storage amount of the ring memory based on the write signal and the read signal of the ring memory, and the speech speed conversion processing means converts the required number of audio signals input to the frame memory into According to the section discriminating means for discriminating whether the corresponding input voice is a voice section or a silence section, and the compression and decompression processing or deletion of the required number of voice signals according to the output of the section discriminating means and the output of the accumulation amount calculating means. Signal processing means for performing processing, wherein when the input voice is a voice section and the ring memory is not in a state immediately before overflow, and when the compression rate fixed mode is selected, the set reproduction speed magnification is set to n. Perform compression / expansion processing at a compression rate of 1 / n or more and determined according to the program type set by the operator; When the reduction rate variation mode is selected, the compression rate is 1 / n or more, where n is the set reproduction speed magnification, and the compression rate is determined according to the program type set by the operator and the storage amount of the ring memory. And means for performing compression / decompression processing at a rate.

In an eighth speech speed conversion apparatus according to the present invention, a required number of audio signals are input to a frame memory in which an input digital audio signal is written at a speed corresponding to a set reproduction speed magnification. Speech rate conversion processing means for performing speech rate conversion processing on these voice signals, a ring memory in which the output of the speech rate conversion processing means is written, reading means for reading data from the ring memory at a constant speed, and a ring. A storage amount calculating unit for calculating the storage amount of the ring memory based on the write signal and the read signal of the memory is provided, and the speech speed conversion processing unit corresponds to the required number of audio signals input to the frame memory. In accordance with the section discriminating means for discriminating whether the input voice is a voice section or a silent section, and the output of the section discriminating means and the output of the accumulation amount calculating means, A signal processing means for performing compression and expansion processing or deletion processing with respect to the required number of audio signals,
The signal processing means, when the input voice is a voice section and the ring memory is not in a state immediately before overflow,
When the fixed compression rate mode is selected, the compression / expansion process is performed at a compression rate of 1 / n or more, where n is the set reproduction speed magnification, and is determined according to the program type set by the operator. When the compression rate fluctuation mode is selected, the set reproduction speed magnification is set to n and 1 / n
It is characterized in that it includes a means for performing the compression / expansion processing at the above-mentioned compression rate and at a compression rate determined according to the program type and the storage amount of the ring memory set by the operator.

As the signal processing means of the seventh or eighth speech speed conversion device, for example, based on the output of the section discriminating means and the output of the accumulation amount calculating means, (1) the input voice is a voice section and A first mode in which the ring memory is not in the state immediately before the overflow; (2) a second mode in which the input voice is in the voice section and the ring memory is in the state immediately before the overflow; and (3) a continuous section in which the input voice is the silent section and the silent section. Is less than a predetermined silence deletion start point determination value,
Third, the ring memory is not in the state immediately before overflow
(4) a fourth mode in which the input voice is a silent section, the duration of the silent section is less than a predetermined silent deletion start point determination value, and the ring memory is in a state immediately before overflow; A fifth mode in which the duration of the silent section is longer than or equal to a predetermined silent deletion start point discrimination value and the ring memory is not in a state immediately before an underflow; and (6) the input sound is a silent section and a silent section. Is greater than or equal to a predetermined silence deletion start point determination value,
And the ring memory is in the state immediately before underflow.
A mode discriminating unit for discriminating which of the modes, when the first mode or the third mode is discriminated, when the fixed compression ratio mode is selected,
The compression / expansion process is performed at a compression ratio of 1 / n or more, where n is the set reproduction speed magnification, and is determined according to the program type set by the operator, and the compression ratio fluctuation mode is selected. In some cases, the compression / expansion process is performed at a compression rate of 1 / n or more, where n is a set reproduction speed magnification, and is determined according to the program type set by the operator and the storage amount of the ring memory. When the processing mode is determined to be the second mode or the fourth mode, the second processing means for deleting the audio signal until the storage amount of the ring memory becomes the state immediately before the underflow. When the processing mode is determined to be the fifth mode, A third processing means for deleting the audio signal in the silent section, and a compression rate 1 / n ± α (where α
Is a value greater than or equal to 0 and less than or equal to 1).

The silence deletion start point discrimination value may be adjusted according to the storage amount of the ring memory.

In the second, third, fifth, sixth, seventh or eighth speech speed converting apparatus, the section discriminating means calculates, for example, a power average value of a required number of audio signals input to a frame memory. And means for determining whether the input voice is a voice section or a silent section based on the calculated power average value and a given threshold value. The threshold value may be adjusted according to the storage amount of the ring memory.

The section discriminating means includes, for example, means for calculating a cumulative power value of a required number of audio signals input to the frame memory, and based on the calculated cumulative power value and a given threshold value. In addition, a device provided with a discriminating means for discriminating whether an input voice is a voice section or a silent section is used. The threshold value may be adjusted according to the storage amount of the ring memory.

As the section discriminating means, for example, means for calculating an average value of the amplitude of a required number of audio signals input to the frame memory, and based on the calculated average value of the amplitude and a given threshold value In addition, a device provided with a discriminating means for discriminating whether an input voice is a voice section or a silent section is used. The threshold value may be adjusted according to the storage amount of the ring memory.

As the section discriminating means, for example, a means for calculating a cumulative amplitude value of a required number of audio signals input to the frame memory, and a section based on the calculated cumulative amplitude value and a given threshold value In addition, a device provided with a discriminating means for discriminating whether an input voice is a voice section or a silent section is used. The threshold value may be adjusted according to the storage amount of the ring memory.

The section discriminating means includes, for example, detecting means for detecting the periodicity of a required number of audio signals input to the frame memory, and based on the detected cycle.
The one provided with a discriminating means for discriminating whether the input voice is a voice section or a silent section is used.

The section discriminating means may be, for example, a predetermined one of a predetermined number of audio signals input to the frame memory.
Or calculating means for calculating a power spectrum for a plurality of frequency bands, and determining means for determining whether the input voice is a voice section or a silent section based on the calculated power spectrum and a given threshold value. Things are used. The threshold value may be adjusted according to the storage amount of the ring memory.

[0029]

In the first speech speed conversion device according to the present invention, the input voice signal is subjected to speech speed conversion processing by speech speed conversion processing means. The output of the speech speed conversion processing means is written to the ring memory. Data written to the ring memory is read at a constant speed. When the input voice signal is a voice section and the ring memory is not in a state immediately before overflow, the set reproduction speed magnification is set to n.
The compression / expansion process is performed on the input audio signal at a compression rate of 1 / n or more, which is determined according to the program type set by the operator.

In the second speech speed converter according to the present invention,
The input analog audio signal is sampled by the A / D converter at a sampling frequency corresponding to the set reproduction speed magnification. The audio signal output from the A / D converter is input to the frame memory. Each time a required number of voice signals are input to the frame memory, the voice speed conversion processing means performs voice speed conversion processing on those voice signals. The output of the speech speed conversion processing means is written to the ring memory. The data written in the ring memory is read based on a read signal having a frequency equal to the sampling frequency at the time of 1 × speed reproduction. The storage amount of the ring memory is calculated by the storage amount calculation means based on the write signal and the read signal of the ring memory.

In the speech speed conversion processing means, the input voice corresponding to the required number of voice signals input to the frame memory is determined by the section determining means as a voice section or a silent section. Then, compression / expansion processing or deletion processing is performed on the required number of audio signals in accordance with the output of the section determination means and the output of the accumulation amount calculation means. In the signal processing means, when the input voice is a voice section and the ring memory is not in the state immediately before the overflow, the compression rate is 1 / n or more, where n is the set reproduction speed magnification, and the compression rate depends on the program type set by the operator. The compression / expansion processing is performed at the compression ratio determined in this manner.

In the third speech speed converter according to the present invention,
The input digital audio signal is written to the frame memory at a speed corresponding to the set reproduction speed magnification. Each time the required number of audio signals are input to the frame memory,
The voice speed conversion processing means performs voice speed conversion processing on those voice signals. The output of the speech speed conversion processing means is
Written to the ring memory. Data written to the ring memory is read at a constant speed based on a read signal. The storage amount of the ring memory is calculated by the storage amount calculation means based on the write signal and the read signal of the ring memory.

In the speech speed conversion processing means, the input voice corresponding to the required number of voice signals input to the frame memory is determined by the section determining means as a voice section or a silent section. Then, compression / expansion processing or deletion processing is performed on the required number of audio signals in accordance with the output of the section determination means and the output of the accumulation amount calculation means. In the signal processing means, when the input voice is a voice section and the ring memory is not in the state immediately before the overflow, the compression rate is 1 / n or more, where n is the set reproduction speed magnification, and the compression rate depends on the program type set by the operator. The compression / expansion processing is performed at the compression ratio determined in this manner.

In the fourth speech speed converter according to the present invention,
The input voice signal is subjected to speech speed conversion processing by speech speed conversion processing means. The output of the speech speed conversion processing means is written to the ring memory. The data written to the ring memory
Read at a constant speed. In the voice speed conversion processing means, when the input voice signal is a voice section and the ring memory is not in a state immediately before overflow, the compression rate is 1 / n or more, where n is the set reproduction speed magnification, and is set by the operator. A compression / expansion process is performed on the input audio signal at a compression rate determined according to the type of program and the storage amount of the ring memory.

In the fifth speech speed converter according to the present invention,
The input analog audio signal is sampled by the A / D converter at a sampling frequency corresponding to the set reproduction speed magnification. The audio signal output from the A / D converter is input to the frame memory. Each time a required number of voice signals are input to the frame memory, the voice speed conversion processing means performs voice speed conversion processing on those voice signals. The output of the speech speed conversion processing means is written to the ring memory. The data written in the ring memory is read based on a read signal having a frequency equal to the sampling frequency at the time of 1 × speed reproduction. The storage amount of the ring memory is calculated by the storage amount calculation means based on the write signal and the read signal of the ring memory.

In the speech speed conversion processing means, the input voice corresponding to the required number of voice signals input to the frame memory is discriminated by the section discriminating means as a voice section or a silent section. Then, compression / expansion processing or deletion processing is performed on the required number of audio signals in accordance with the output of the section determination means and the output of the accumulation amount calculation means. In the signal processing means, when the input voice is in the voice section and the ring memory is not in the state immediately before the overflow, the compression rate is not less than 1 / n with the set reproduction speed magnification being n, and the program type and the ring set by the operator are set. The compression / decompression processing is performed at a compression ratio determined according to the amount of storage in the memory.

In the sixth speech speed converter according to the present invention,
The input digital audio signal is written to the frame memory at a speed corresponding to the set reproduction speed magnification. Each time the required number of audio signals are input to the frame memory,
The voice speed conversion processing means performs voice speed conversion processing on those voice signals. The output of the speech speed conversion processing means is
Written to the ring memory. Data written to the ring memory is read at a constant speed based on a read signal. The storage amount of the ring memory is calculated by the storage amount calculation means based on the write signal and the read signal of the ring memory.

In the voice speed conversion processing means, the input voice corresponding to the required number of voice signals input to the frame memory is determined by the section determining means as a voice section or a silent section. Then, compression / expansion processing or deletion processing is performed on the required number of audio signals in accordance with the output of the section determination means and the output of the accumulation amount calculation means. In the signal processing means, when the input voice is in the voice section and the ring memory is not in the state immediately before the overflow, the compression rate is not less than 1 / n with the set reproduction speed magnification being n, and the program type and the ring set by the operator are set. The compression / decompression processing is performed at a compression ratio determined according to the amount of storage in the memory.

In the seventh speech speed converter according to the present invention,
The input analog audio signal is sampled by the A / D converter at a sampling frequency corresponding to the set reproduction speed magnification. The audio signal output from the A / D converter is input to the frame memory. Each time a required number of voice signals are input to the frame memory, the voice speed conversion processing means performs voice speed conversion processing on those voice signals. The output of the speech speed conversion processing means is written to the ring memory. The data written in the ring memory is read based on a read signal having a frequency equal to the sampling frequency at the time of 1 × speed reproduction. The storage amount of the ring memory is calculated by the storage amount calculation means based on the write signal and the read signal of the ring memory.

In the speech speed conversion processing means, the input voice for the required number of voice signals input to the frame memory is determined by the section determining means as a voice section or a silent section. Then, compression / expansion processing or deletion processing is performed on the required number of audio signals in accordance with the output of the section determination means and the output of the accumulation amount calculation means. In the signal processing means, when the input speech is in the speech section and the ring memory is not in the state immediately before the overflow, and the fixed compression rate mode is selected, the compression rate is not less than 1 / n, where n is the set reproduction speed magnification. The compression / expansion process is performed at the compression rate determined according to the program type set by the operator, and when the compression rate variation mode is selected, the compression rate of 1 / n or more is set with the set reproduction speed magnification as n. The compression / expansion processing is performed at a compression rate determined according to the program type and the storage amount of the ring memory set by the operator.

In the eighth speech speed converter according to the present invention,
The input digital audio signal is written to the frame memory at a speed corresponding to the set reproduction speed magnification. Each time the required number of audio signals are input to the frame memory,
The voice speed conversion processing means performs voice speed conversion processing on those voice signals. The output of the speech speed conversion processing means is
Written to the ring memory. Data written to the ring memory is read at a constant speed based on a read signal. The storage amount of the ring memory is calculated by the storage amount calculation means based on the write signal and the read signal of the ring memory.

In the speech speed conversion processing means, the input voice corresponding to the required number of voice signals input to the frame memory is determined by the section determining means as a voice section or a silent section. Then, compression / expansion processing or deletion processing is performed on the required number of audio signals in accordance with the output of the section determination means and the output of the accumulation amount calculation means. In the signal processing means, when the input speech is in the speech section and the ring memory is not in the state immediately before the overflow, and the fixed compression rate mode is selected, the compression rate is not less than 1 / n, where n is the set reproduction speed magnification. The compression / expansion process is performed at the compression rate determined according to the program type set by the operator, and when the compression rate variation mode is selected, the compression rate of 1 / n or more is set with the set reproduction speed magnification as n. The compression / expansion processing is performed at a compression rate determined according to the program type and the storage amount of the ring memory set by the operator.

[0043]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment in which the present invention is applied to a VTR will be described below with reference to the drawings.

FIG. 1 shows the overall structure of the speech speed conversion device.

The input audio signal is amplified by the ALC amplifier 1 and then sent to the A / D converter 2 where it is converted into, for example, a 12-bit digital signal. The standard sampling frequency of the A / D converter 2 is, for example, 8 KHz. At the time of 2 × speed reproduction, the sampling frequency fs of the A / D converter 2
AD is 16 KHz.

The output of the A / D converter 2 is a DSP (Digital
l Signal Processor) 4 and also to the level detector 3. The level detection unit 3 includes the A / D conversion unit 2
When the A / D-converted data reaches the maximum value of the conversion range, the ALC (automatic level control) signal is output to A
Output to LC amplifier 1. Thereby, the ALC amplifier 1
Is controlled so that the input signal of the A / D converter 2 does not exceed the maximum range. That is, when the playback tape speed of the VTR changes, the input signal level of the ALC amplifier 1 also changes. Therefore, by automatically adjusting the amplifier gain based on the output of the level detector 3, the A / D
The input signal of the converter 2 is prevented from exceeding the maximum range.

The DSP 4 includes a frame memory 5 having a capacity capable of storing two frames of audio signals, and a speech speed conversion unit 6 for performing a speech speed conversion process on a frame basis for the audio signals stored in the frame memory 5. ing. Here, it is assumed that one frame is composed of 200 pieces of sampling data.

One of the first half area and the second half area in the frame memory 5 is processed by the speech speed converter 6 for one frame of the audio signal stored in one of the areas, and at the same time, the A The signal from the / D converter 2 is accumulated. When the signal for one frame is accumulated in the other area, the data in the area is processed by the speech speed conversion unit 6 and the data already processed is stored. The signal from the A / D conversion unit 2 is stored in the one of the areas.

The data output from the voice speed converter 6 is written to the ring memory 7 based on a write clock. The data written in the ring memory 7 is read based on a read clock. The signal read from the ring memory 7 is converted into an analog signal by the D / A converter 8, then amplified by the amplifier 10, and output as an audio output signal.

The sampling frequency fs of the D / A converter 8
DA is 8 KHz. The frequency of the read clock of the ring memory 7 is also 8 KHz. As the ring memory 7, a memory of 21845 × 12 bits, that is, a memory of 21845 words is used. Therefore, the maximum time during which data can be stored in the ring memory 7 (the maximum delay time of the output time with respect to the input signal) is 21
845 x 1/8000 = 2.73 seconds.

The write clock for the ring memory 7 is input to an up-counting input terminal (UP) of the up-down counter 9. A read clock for the ring memory 7 is input to a down-counting input terminal (DOWN) of the up-down counter 9. The up / down counter 9 calculates the difference between the total number of input write clocks and the total number of input read clocks (the ring memory 7).
Count) and count the count value to 15 bits
As a digital signal. The output of the up / down counter 9 is sent to the speech speed converter 6.

FIG. 2 shows a detailed configuration of the speech speed conversion unit 6.

The audio signal read from the frame memory 5 is sent to the power calculator 11, and the average power value P of the audio signal for one frame is calculated. This average power value P represents the amplitude of each audio signal in one sampled frame as i ₀ , i ₁ , ... i _{N- 1} (However, N = 200)
Then, it is obtained by the following Expression 1.

[0054]

(Equation 1)

The average power value P obtained by the power calculator 11 is sent to the comparator 12. The threshold value Th is sent from the threshold value memory 13 to the comparison unit 12, and the average power value P is equal to or larger than the threshold value Th (P ≧ Th) or the average power value P is smaller than the threshold value Th. (P <Th) is determined. The comparison unit 12 outputs a signal indicating that the current frame is a voice section when the average power value P is equal to or larger than the threshold Th (P ≧ Th). Is a signal indicating that is a silent section.

The threshold value Th is, for example, 2 when the number of quantization bits of the A / D converter 2 is 12 bits.
Set to ¹² . It should be noted that the threshold value Th is set as follows.
May be changed. That is, as shown by a dotted line in FIG. 2, a power steady state detection and threshold value updating unit 14 is provided. The power steady state detection and threshold update unit 14 calculates the average power value P from the power calculation unit 11 as
It is determined whether or not it is constant over a predetermined number of frames (for example, 40 frames). If it is constant (steady state), a value twice as large as the average power value P at that time is written into the threshold value memory 13, The threshold Th is updated. However, the maximum value of the updated threshold is a predetermined value,
For example, it is limited to 2 ¹⁴ . By doing so, it is possible to treat the constantly occurring noise as a silent section.

The voice section and the silent section of the input signal may be determined based on the power accumulated value Pa of the voice signal of each frame and the given threshold value as shown in the following equation (2). .

[0058]

(Equation 2)

The output of the comparing section 12 is sent to the conditional branching section 15. The output of the ring memory accumulated amount state determination unit 16 is input to the conditional branch unit 15. Also, conditional branching unit 1
5 is supplied with an audio signal from the frame memory 5 via the power calculator 11. Furthermore, conditional branching unit 1
5, a pause continuation length setting memory 17 is connected. In the pause continuation length setting memory 17, a pause continuation length Tdel (a silence deletion start point discrimination value) for determining a deletion start point of a silent section is set.

Based on the amount of storage sent from the up / down counter 9, the ring memory storage amount state determination unit 16 determines that the state of the ring memory 7 has become the state immediately before overflow and that the state of the ring memory 7 is under. Detects the state immediately before the flow.

That is, the overflow detection data memory 18 stores overflow detection data Tmax, and the underflow detection data memory 19 stores underflow detection data Tmin.
The overflow detection data Tmax is set to, for example, a value 21645 smaller than the total number of words (TOTAL) 21845 of the ring memory 7 by 200. The underflow detection data Tmin is set to, for example, 200.

The accumulated amount sent from the up / down counter 9 is equal to the overflow detection data Tmax.
At this point, the ring memory storage amount state determination unit 16 outputs a detection signal immediately before overflow. When the accumulated amount sent from the up / down counter 9 becomes equal to or less than the underflow detection data Tmin, the ring memory accumulated amount state determination unit 16 outputs a detection signal immediately before the underflow. The conditional branch unit 15 determines that the ring memory 7 is in the state immediately before the overflow when the immediately before overflow detection signal is input, and the ring memory 7 is in the state immediately before the underflow when the immediately before underflow detection signal is input. Is determined.

The condition branching unit 15 determines whether the signal is a speech section or a silent section sent from the comparing unit 12, a detection signal regarding the ring memory state sent from the ring memory storage amount state determining unit 16, and a pause continuation. Length setting memory 1
Based on the pause duration Tdel set to 7, the following six cases are classified. And
In response, the multiplexer 20 is controlled to send the audio signal to a predetermined processing unit.

(1) First Case (case 1) When it is determined that the input signal is in the voice section and the ring memory 7 is not in the state immediately before the overflow, the first case occurs.

In this case, the audio signal is sent to the pitch compression / expansion means 23 via the multiplexer 20.
The pitch compression / expansion means 23 performs variable speech control (VSC). Assuming that the reproduction speed magnification of the VTR is n, the compression ratio of the input signal is 1 unit.
The decompression and compression processing is performed at a compression ratio α of / n or more. Compression rate α
Is determined by the compression / expansion rate adjusting means 42. As the decompression compression method used here, for example, an overlap addition method (Pointer Interval Contr
ol Overlap and Add: PICOLA), TDHS (Tim
e Domain Harmonic Scaling) method. The signal subjected to the expansion / compression processing by the pitch expansion / compression means 23 is sent to the ring memory 7 via the demultiplexer 27, and is written to the ring memory 7 according to a write clock.

During double-speed playback of a VTR, the A / D
The sampling frequency fsAD of the converter 2 is 16 KHZ, and the sampling frequency fsDA of the D / A converter 8 is 8 KHZ. Therefore, the pitch is restored and output.

In conventional general time-axis expansion compression, compression is performed at a compression rate of 1/2 at the time of double speed reproduction. In other words, two pitch periods are thinned out to one pitch period. For this reason, the output sound is twice as fast as the standard sound speed. In other words, in the normal reproduction of the double speed reproduction, the output audio is twice the standard audio speed. However, the pitch remains unchanged.

On the other hand, in the pitch expansion / compression means 23 provided in the speech speed conversion section 6 of FIG. 2, the compression rate α is determined by the mode and ring memory set by the user using the operation section (not shown). 7 is determined by the compression / decompression rate adjusting means 42 based on the change in the accumulation amount of the data No. 7. However, the compression ratio α is a value of 以上 or more.

Here, the user operates the operation unit (not shown).
There are two types of methods for setting the mode by using.

First, a program setting mode (program type) immediately before starting recording (recording) or during recording (recording).
A recording medium, such as video tape, audio tape,
The user sets and stores it in a DAT or the like or a setting program memory (not shown).

At the time of reproduction, the program setting mode (program type) stored in the recording medium or the set program memory is read, and the audio reproduction speed is controlled using this mode signal (see FIG. 2).

Second, the user specifies the program setting mode immediately before the start of reproduction or at the time of reproduction, and controls the audio reproduction speed using this mode signal (see FIG. 2).

The type of mode set by the operation unit sets a program setting mode for selecting a program and whether the compression rate α is fixed or varied for the program set in the program setting mode. There is a fixed fluctuation setting mode.

The following table shows examples of programs set by the program setting mode during double-speed playback of a VTR, and the audio playback speed (compression) for each program when the fixed mode is set for each program. (Rate) and an example of a fluctuation range of the audio reproduction speed (compression rate) for each program when the fluctuation mode is set for each program.

[0075]

[Table 1]

The audio playback speed range for the fixed mode and the audio playback speed range for the variable mode for each program are as follows:
It is set based on the following concept. That is, the utterance speed differs depending on the program content. For example, in drama, news, F1 relay and shogi programs,
The utterance speed is the fastest in F1 broadcasting, news, drama,
The utterance speed decreases in the order of the shogi game. Such a difference in utterance speed is caused by the number of mora per unit time. Mora refers to the relative length of a sound that is a unit such as stress or intonation in a prosodic sound, and one mora corresponds to the length of one syllable including a short vowel.

The average value of the number of mora per unit time of each program is as follows, although it varies depending on the speaker.

F1 relay: 12 mora / s News: 8 mora / s Drama: 5 mora / s Shogi game: 3 mora / s

When the fixed mode is set, the compression rate of the set program with respect to the audio reproduction speed in the fixed mode is determined as the compression rate α. For example,
When a news program is set and the fixed mode is set, the compression rate α is determined to be a compression rate for 1.4 times speed, for example, 0.714. As described above, a program having a higher utterance speed has a lower compression ratio (higher audio reproduction speed), and thus has the following advantages.

That is, the higher the utterance speed of a program, the more likely it is that the ring memory 7 will be in the state immediately before the overflow. Therefore, the compression ratio is determined so that the sound reproduction speed is close to twice the speed. Conversely, the compression rate is determined such that the lower the utterance speed of the program, the closer the sound reproduction speed to the normal speed.
Therefore, the sound reproduction speed is equal to or lower than the double speed and is a speed corresponding to the original utterance speed, and a more natural reproduced sound can be obtained.

When the variable mode is set, the compression rate α is determined as follows in the range of the compression rate with respect to the audio reproduction speed range in the variable mode for the set program. The compression / expansion rate adjusting means 42 sets the compression rate so that the compression rate increases as the storage amount of the ring memory 7 decreases, that is, the audio reproduction speed decreases, and the compression rate increases as the storage amount of the ring memory 7 increases. The compression ratio α is determined so that is smaller, that is, the audio reproduction speed is higher.

That is, when it is determined that the case corresponds to the first case (case 1), the compression / decompression rate adjusting means 42
Obtains the storage amount of the ring memory 7 from the ring memory storage amount state determination unit 16. And the obtained ring memory 7
Is divided by the sampling frequency of the D / A converter 8 to calculate the accumulation time Tm. The compression rate α is determined based on the calculated accumulation time Tm.

More specifically, the storage amount of the ring memory 7 obtained from the ring memory storage amount state determination unit 16 is a sampling frequency of the D / A conversion unit 8 of 80.
By dividing by 00, the accumulation time Tm is obtained. Then, the compression rate α for the storage time Tm is obtained based on the data of the compression rate for the storage time created in advance for each program.

The following table shows an example of the data of the compression rate α with respect to the accumulation time Tm for the F1 relay program at the time of double-speed reproduction of the VTR. In this table, V
Indicates the audio playback speed corresponding to the compression ratio.

[0085]

[Table 2]

As can be seen from this table, the ring memory 7
As the accumulation time Tm becomes shorter, the compression ratio α becomes larger, and the sound reproduction speed becomes slower. Conversely, as the storage time Tm of the ring memory 7 increases, the compression ratio α decreases, and the audio reproduction speed increases. Therefore, when the variable mode is set, in addition to the above-described advantages when the fixed mode is set, there is an advantage that the missing portion of the voice in the voice section of the input signal can be reduced as much as possible.

In the above-described method, the missing portion of the voice is reduced as much as possible. However, in the F1 relay and the fast-talking news, the elderly person may not be able to hear it. In such a case, the number of missing portions of the sound may be increased, for example, the sound reproduction speed range with respect to the storage time may be set to 1.0 to 1.3 times, and the sound may be made slow. In this way, although the number of missing voices increases, the speed of the voice to be played back is reduced, and the elderly people can easily hear the voices.

If the compression ratio α is not less than 圧 縮, for example, not shown in Table 1 above, but for convenience of explanation, if it is determined to be 2/3, the three pitch period is thinned out to two pitch periods. I will Therefore, the output sound is 3/2 times the standard sound speed. In this case also, the pitch remains unchanged. As described above, when compression is performed at a compression ratio of 2/3, the compression ratio is 1 /
As compared with the case of 2, the signal is expanded by 2 / 3-1 / 2 = 1/6. This extension is the amount of storage in the ring memory 7.

A method of compressing an input signal at a compression ratio of 2/3 using PICOLA will be briefly described with reference to FIG. First, a pitch period is extracted from an input signal.
Let the extracted pitch period be Tp. The waveform A is weighted linearly from 1 to 0 (weight function K1), and a waveform A 'is created. A weight (weight function K2) from 0 to 1 is assigned to the waveform B,
A waveform B 'is created.

Then, the waveforms A 'and B' are added to generate a waveform A '* B' having a length Tp.
These weights are added to maintain continuity at connection points before and after the waveform A '* B'. Next, the pointer is moved by 3Tp, which is a length determined based on the compression ratio, and the same operation is performed. Thus, two waveforms A ′ * B ′ and C are obtained from the three waveforms A, B, and C. In this way, the signal for three pitch periods becomes 2
The signal is compressed to a pitch period signal.

As a decompression method by the pitch decompression / compression means 23, as shown in FIGS. 17 (a) and (b), decompression processing is performed in units of a predetermined fixed frame length Ts without extracting a pitch. You may do so. The fixed frame length Ts is set to, for example, a length of 200 input data. FIG. 17 shows an example in which 3Ts is changed to 2Ts.

In the method shown in FIG. 17A, of the waveforms A, B, and C having the fixed frame length Ts, the waveform A is given a weight (weight function K1) that goes linearly from 1 to 0. Thus, a waveform A ″ is created.
Weight (weight function K2) toward the waveform B "
Is created.

Then, the waveforms A ″ and B ″ are added to generate a waveform A ″ * B ″ having a length Ts.
These weights are added to maintain continuity at connection points before and after the waveform A "* B". And the next waveform C
Is output as is. As a result, two waveforms A "* B" and C are obtained from the three waveforms A, B, and C. In this way, the signal for 3Ts is compressed into a signal for 2Ts.

In the method shown in FIG. 17B, the waveform A among the waveforms A to C having the fixed frame length Ts has a weight (weight function K3) in which, for example, 20 data from the top linearly goes from 0 to 1. To obtain a waveform A ″.
A weight B (weight function K4) that linearly goes from 1 to 0 is applied to the first to 200th input data to obtain a waveform B ". Then, the waveform C is deleted. The following three waveforms D to F
, The same processing is performed. In this way,
A signal consisting of three waveforms AC (or DF) is 2
Compressed into a signal consisting of two waveforms A "and B" (or D "and E"). That is, the signal for 3Ts is 2
It is compressed to a signal of Ts.

When the decompression and compression processing in units of the fixed frame length is used, the sound quality is reduced, but the processing amount is reduced as compared with the case where the decompression and compression processing for each pitch period is used.

When this speech speed converter is applied to an English language learning device (at the time of 1 × speed reproduction), the sampling frequency fsAD of the A / D converter 2 is 8 KHZ,
The sampling frequency fsDA of the / A conversion unit 8 is 8 KHZ
It is. In this case, the compression rate α is determined to be a value of 1 or more by the compression / expansion rate adjusting means 42. The compression ratio α is
For example, if it is determined to be 1.5, the audio signal is expanded by the pitch compression / expansion means 23 so that two pitch periods become three pitch periods. That is, when the voice section is 1.5
It is doubled. Therefore, in this case, the signal is expanded by 3 / 2−1 = 1/2 with respect to the normal reproduction of the 1 × speed reproduction.
Is the accumulated amount.

(2) Second case (case 2) When it is determined that the input signal is in the voice section and the ring memory 7 is in the state immediately before the overflow, the second case
It becomes a case.

In this case, the audio signal is sent to the input signal deleting section 21 via the multiplexer 20, and the audio signal is deleted. Specifically, the writing operation to the ring memory 7 is stopped until the count value of the up / down counter 9 becomes the underflow detection data Tmin, that is, until the ring memory 7 is in the state immediately before the underflow.

When the ring memory 7 is in the state immediately before underflow, 200 or less, for example, 100 muffling signals (signals having a value “0”) are output from the muffling insertion section 22, and the muffling signals are output from the demultiplexer 27. Is sent to the ring memory 7 via the. The reason why the mute signal is written in the ring memory 7 is to prevent a click sound from being generated at a joint of audio signals due to audio deletion.

(3) Third Case (case 3) It is determined that the input signal is a silent section, the duration of the silent section is less than the set pause duration Tdel, and the ring memory 7 is not in the state immediately before the overflow. When this is done, it becomes the third case.

In this case, the same processing as in the first case is performed.

(4) Fourth Case (case 4) It is determined that the input signal is a silent section, the duration of the silent section is less than the set pause duration Tdel, and the ring memory 7 is in a state immediately before overflow. When it is done, it becomes the fourth case.

In this case, the same processing as in the second case is performed.

(5) Fifth Case (case 5) If the input signal is a silent section, the duration of the silent section is longer than the set pause duration Tdel, and the ring memory 7 is not in the state immediately before the underflow. When it is determined, it is the fifth case.

In this case, the audio signal is sent to the input signal deleting section 25 via the multiplexer 20, and the audio signal is deleted. Specifically, the writing operation to the ring memory 7 is stopped. However, in order to prevent the start portion of the voice section (unvoiced section) from being lost or to prevent a click sound from being generated at a joint due to deletion of the voice, the waveform synthesis / insertion unit 26 performs the waveform synthesis / insertion. Processing is performed.

The waveform synthesizing and inserting process performed by the waveform synthesizing and inserting unit 26 will be described with reference to FIGS. In the method according to FIG. 4A, the waveform combining and inserting unit 26
A memory 31 and a second memory 32 are provided. At the start of the input signal deletion process by the input signal deletion unit 25, a predetermined length T equal to or less than one frame length from the deletion start point.
s, for example, an input signal for one frame is stored in the first memory 31.
Are sequentially stored in address order. Next, the first memory 31
Is multiplied with the content A of the first memory 31 by a function K1 that linearly changes from 1 to 0 as the address of the first memory 31 increases. Then, the multiplication result A ′ is written into the first memory 31 again.

An input signal of a predetermined length Ts immediately before the end point of the input signal deletion section by the input signal deletion section 25 is
The data is sequentially stored in the second memory 32 in the order of addresses. next,
As the address of the second memory 32 increases, 0 to 1
Is multiplied by the content B of the second memory 32. Then, the multiplication result B ′ is written into the second memory 32 again. Thereafter, the first memory 31
Is added to the content B 'of the second memory 32 to obtain data A' * B 'having a predetermined length Ts.
Then, the obtained data A ′ * B ′ for the predetermined length Ts is sent to the ring memory 7 via the demultiplexer 27 and written into the ring memory 7.

In the method shown in FIG. 4B, input signals of a predetermined length Ts, for example, one frame, which are equal to or less than one frame length, are sequentially stored in the first memory 31 in order of address from the deletion start point. Next, a function K3 having a slope that changes linearly from 1 to 0 at the rear end is stored in the content A of the first memory 31.
Is multiplied by Then, the multiplication result A ′ is again the first
The data is written to the memory 31.

The input signal of a predetermined length Ts immediately before the end point of the input signal deletion section by the input signal deletion section 25 is
The data is sequentially stored in the second memory 32 in the order of addresses. next,
The content B of the second memory 32 is multiplied by a function K4 having a slope linearly changing from 0 to 1 at the front end. Then, the multiplication result B ′ is written into the second memory 32 again. Thereafter, the content A 'of the first memory 31 and the content B' of the second memory 32 are joined to obtain 2Ts worth of data A '+ B'. And the obtained 2Ts
The minute data A ′ + B ′ is sent to the ring memory 7 via the demultiplexer 27 and written to the ring memory 7. FIG. 4B shows an example in which Ts is the length of one frame, but data having half the length of one frame may be used as Ts.

When the input signal deleting section 25 repeatedly performs the process of deleting the audio signal in the silent section, the ring memory 7 may be in a state immediately before the underflow. In this case, the input signal for the predetermined length Ts is stored in the second memory 32 from the time when the ring memory 7 enters the state immediately before the underflow. Then, based on the data stored in the first memory 31 and the data stored in the second memory 32, the same input signal deletion processing as described above is performed.

(6) Sixth Case (case 6) When the input signal is a silent section, the duration of the silent section is longer than the set pause duration Tdel, and the ring memory 7 is in a state immediately before underflow. When it is determined, it is the sixth case.

In this case, the input signal is sent to the thinning processing section 24 via the multiplexer 20. The thinning-out section 24 performs thinning-out processing so that the compression rate is 1 / n, where n is the reproduction speed magnification of the VTR. For example, at the time of double speed reproduction, the input signal is thinned at a compression rate of 1/2, and at the time of triple speed reproduction, the input signal is thinned at a compression rate of 1/3. During 1 × speed reproduction, the input signal is output as it is.

The following method is used as the thinning processing by the 1 / n thinning processing section 24. Here, 2
A description will be given of the case of double speed reproduction as an example.

The pitch of the input signal is extracted using the time axis compression method using PICOLA or TDHS described above,
The pitch data portion is thinned out so that the compression ratio becomes 1/2.

Also, as shown in FIGS. 5A to 5C,
The waveform may be thinned every predetermined time Ts without performing the pitch extraction.

In the method of FIG. 5A, the waveform B and the waveform D among the waveforms A to D are thinned out, and a signal including the waveforms A and C is obtained.

In the method shown in FIG. 5B, the waveform B and the waveform D among the waveforms A to D are thinned out. The waveform A has a slope rising from 0 to 1 at the front end (function K4).
Is multiplied by a function having a slope (function K3) having a slope falling from 1 to 0 at the rear end to generate a waveform A ′. The waveform C ′ is formed by multiplying the waveform C by a function having a slope (function K4) rising from 0 to 1 at the front end and a slope (function K3) falling from 1 to 0 at the rear end. Is done. In this way, a signal consisting of four waveforms A to D is compressed into a signal consisting of two waveforms A 'and C'.

In the method shown in FIG. 5C, the waveform A is weighted linearly from 1 to 0 (weight function K1) to generate a waveform A '. A weight (weight function K2) from 0 to 1 is assigned to the waveform B,
A waveform B 'is created. Then, those waveforms A ′ and B ′ are added to create a waveform A ′ * B ′ having a length Ts.

Similarly, a weight (function K1) linearly going from 1 to 0 is assigned to the waveform C, and the waveform C ′
Is created. The waveform D 'is weighted from 0 to 1 (function K2) to generate a waveform D'. Then, those waveforms C ′ and D ′ are added to create a waveform C ′ * D ′ having a length Ts. In this way, a signal consisting of four waveforms A to D becomes two waveforms A '
It is compressed into a signal consisting of * B 'and C' * D '.

As described above, in the case corresponding to the sixth case, the thinning-out process is performed at a compression ratio of 1 / n, where n is the reproduction magnification of the VTR. However, the compression ratio is controlled as follows. You may make it.

When the decimation process is performed at a compression ratio of 1 / n, the ratio f between the sampling frequency fsDA of the D / A converter 8 and the sampling frequency fsAD of the A / D converter 2 is obtained.
When sDA / fsAD is equal to the compression ratio 1 / n,
The storage amount of the ring memory 7 does not change. However, the calculation accuracy of the compression ratio 1 / n and the sampling frequency fs
Depending on the clock accuracy of AD and fsDA, fsDA /
It is possible that fsAD does not equal the compression ratio 1 / n.

When fsDA / fsAD becomes larger than the compression ratio 1 / n (fsDA / fsAD> 1 / n),
Assuming that fsDA / fsAD = 1 / a (a> 0), ｛(1
/ A)-(1 / n)}, the compression ratio decreases, the degree of thinning increases, the storage amount of the ring memory 7 decreases, and the storage amount of the ring memory 7 may underflow. .

On the other hand, fsDA / fsAD is equal to the compression ratio 1 /
n (fsDA / fsAD <1 /
In n), assuming that fsDA / fsAD = 1 / a (a> 0), the compression ratio increases by {(1 / n)-(1 / a)}, the degree of thinning decreases, and the ring memory 7 Is increasing.

Therefore, when performing the thinning process,
After confirming the storage amount of the ring memory 7, the compression ratio is controlled as follows. As fsDA / fsAD = 1 / a (a> 0), α that satisfies the condition of (1 / n) −α <1 / a <(1 / n) + α is selected. Here, α is a value of 0 or more and 1 or less, for example, a value in a range of 0.001 to 0.1.

When fsDA / fsAD becomes larger than the compression ratio 1 / n, that is, when the amount of storage in the ring memory 7 decreases, the compression ratio is changed from 1 / n to ｛(1 / n).
n) + α｝. That is, the compression ratio is increased, and the amount of storage in the ring memory 7 is increased.

When fsDA / fsAD becomes smaller than the compression ratio 1 / n, that is, when the amount of storage in the ring memory 7 increases, the compression ratio is changed from 1 / n to ｛(1 / n).
n) -α｝. That is, the compression ratio is reduced, and the amount of storage in the ring memory 7 is reduced.

In the above description, the compression ratio is changed based on the storage amount of the ring memory 7, but when the thinning process is performed, the compression ratio is set to {(1 / n) -α} or { (1 / n) + α｝ may be alternately changed.

FIGS. 6 and 7 show a processing procedure by the speech speed conversion unit 6. FIG.

The processing performed by the speech speed conversion unit 6 in the case of double speed reproduction of a VTR will be described below.

(1) Processing at the Start of Reproduction When reproduction is started and the average power value P of the first frame is calculated by the power calculator 11 (step 1), the calculated average power value P is a threshold. Whether or not the value is equal to or greater than Th is determined based on the output of the comparison unit 12 (step 2).

When the input speech signal starts from a silent section, in the first frame, the average power value P becomes smaller than the threshold value Th, and the process proceeds to step S11. Then, the duration of the silent section (the number of frames in which the silent section continues) is calculated, and it is determined whether or not the calculated duration is equal to or longer than the pause duration Tdel set in the pause duration memory 17 (step). 12). This pause continuation length T
“del” is set to, for example, a length of four frames in terms of the number of frames.

In the processing for the first frame,
Since the continuation length of the silent section is less than the pause continuation length Tdel, it is determined whether or not the ring memory 7 is in the state immediately before the underflow based on the output of the ring memory storage amount state determination unit 16 (steps 13 and 14). .

In the processing for the first frame,
Since the ring memory 7 is in the state immediately before the underflow, the frame data is thinned by the thinning processing unit 24 at a compression rate of 1/2 (step 28), and the compressed data after the thinning processing is written to the ring memory 7. Thereafter, the process returns to step 1.

(2) Description of the First Case Process If it is determined in step 2 that the average power value P is equal to or greater than the threshold Th, it is determined that the current frame is a voice section, and step 3 is performed. Proceed to. In step 3,
Whether or not the previous frame was a deletion section is determined by the first flag F
1 is determined based on the state. If the previous frame is not a deletion section, it is determined whether or not the ring memory 7 is in the state immediately before overflow based on the output of the ring memory storage amount state determination unit 16 (steps 6 and 7). If the previous frame is a section to be deleted, after the processes of steps 4 and 5 are performed, it is determined whether or not the ring memory 7 is in a state immediately before overflow (steps 6 and 7). Steps 4 and 5 will be described later.

If it is determined in step 7 that the current state is not the state immediately before the overflow, the first case occurs, and the pitch compression / expansion means 23 converts the current frame data into the compression rate α determined by the compression / expansion rate adjustment means 42. Is compressed on the time axis (step 8). The compressed data is sent to and written to the ring memory 7. After this,
Return to step 1.

(2) Description of Processing in Second Case When it is determined in step 2 that the average power value P is equal to or greater than the threshold Th, the frame transmitted this time is determined to be a voice section. Then, go to step 3. In step 3, it is determined whether or not the previous frame is a deletion section based on the state of the first flag F1. If the previous frame is not the deletion section, it is determined whether or not the ring memory 7 is in the state immediately before the overflow based on the output of the ring memory storage amount state determination unit 16 (step 6,
7). If the previous frame is a section to be deleted, after the processes of steps 4 and 5 are performed, it is determined whether or not the ring memory 7 is in a state immediately before overflow (steps 6 and 7). Steps 4 and 5 will be described later.

If it is determined in step 7 that the state is immediately before the overflow, the second case occurs, and
The input signal is deleted by the input signal deletion unit 21 until the underflow detection signal is output from the ring memory storage amount state determination unit 16 (step 9). That is, writing to the ring memory 7 is stopped until the ring memory 7 is in a state immediately before underflow.

When the ring memory 7 is in the state immediately before underflow, the silence insertion section 22 writes a predetermined number of silence signals "0" of 200 or less into the ring memory 7 (step 10). Then, the process returns to step 1.

[0139] Instead of the processing in step 10 described above, FIG.
The processing shown in FIG. 9A or FIG. 9B may be performed. The method shown in FIG. 9A will be described. From the time when it is determined in step 7 that the state is immediately before the overflow, for example, for the waveform A corresponding to 200 input signals, the weight (linear) going from 1 to 0 linearly ( Weight function K
1) is applied to obtain a waveform A '. In addition, for waveform B corresponding to 200 input signals from immediately before the underflow to 200 before the underflow, weights from 0 to 1 (weight function K2)
To obtain a waveform B ′.

Then, the obtained two waveforms A ′ and B ′ are added to form a waveform A ′ * of 200 lengths.
Create B '. Then, 200 signals for this waveform A ′ * B ′ are written in the ring memory 7. The detection 200 times before the immediately before the underflow is performed based on the count value of the up / down counter 9. Thus, it is possible to effectively prevent a click sound from being generated at a joint between audio signals before and after the audio deletion section.

The method shown in FIG. 9B will be described. When it is determined in step 7 that the state is immediately before the overflow, for example, the waveform A for 100 input signals is linearly changed from 1 to 0. A weight A (weight function K1) is applied to obtain a waveform A '. Also, for waveform B for 100 input signals from immediately before the underflow to 100 before, the weights from 0 to 1 (weight function K
By adding 2), a waveform B 'is obtained. Then, 200 signals obtained by joining the two obtained waveforms A ′ and B ′ are written to the ring memory 7.

In step 9 described above, if it is determined that the state is immediately before the overflow, the input signal deletion unit 21 deletes the input signal until the underflow detection signal is output from the ring memory storage state determination unit 16. However, the data stored in the ring memory 7 may be deleted so that the ring memory 7 is in a state immediately before the underflow.

Specifically, as shown in FIG. 18B, the write start address of the ring memory 7 is changed from the address (point C) in the state immediately before the overflow shown in FIG. Is jumped to the address (point A) where the state immediately before the underflow occurs. Therefore, in the process of step 9, the data stored at the addresses from the point A to the point C is deleted. Thereafter, as shown in FIG. 18C, after the mute signal is written in step 10, the input data is written.

If the data stored in the ring memory 7 is deleted in step 9 so that the ring memory 7 is in the state immediately before the underflow, the mute signal is written in the ring memory 7 in step 10. 19 (a) or FIG. 19 (b).

Now, as shown in FIG. 18B, the write start address of the ring memory 7 is changed from the address (point C) in the state immediately before the overflow shown in FIG. It is assumed that the user jumps to the address (point A) in the immediately preceding state. For data S stored from a point A to a predetermined number, for example, an address 200 points ahead (point B in FIG. 19A), FIG.
As shown in (a), a waveform S ′ is obtained by applying a weight (weight function K1) linearly going from 1 to 0. Further, as shown in FIG. 19A, a weight (weight function K2) from 0 to 1 is applied to 200 pieces of input data (waveform T) written to the ring memory 7 thereafter. ,
A waveform T 'is obtained.

Then, the obtained two waveforms S ′ and T ′ are added, and a waveform S ′ * having a length of 200 pieces is obtained.
Create T '. Then, 200 signals for this waveform S ′ * T ′ are written into the ring memory 7 from the point A. As a result, it is possible to effectively prevent a click sound from being generated at a joint between audio signals before and after the accumulated data deletion section.

The method shown in FIG. 19B will be described. Data stored from a point A in FIG. 18B to a predetermined number of addresses, for example, 100 addresses ahead (point B in FIG. 19B). S is weighted linearly from 1 to 0 (weight function K1) to obtain a waveform S '. Also,
Thereafter, a weight (weight function K2) from 0 to 1 is applied to 100 pieces of input data (waveform T) written to the ring memory 7 to obtain a waveform T '. Then, 200 signals obtained by joining the two obtained waveforms S ′ and T ′ are written to the ring memory 7 from the point A.

(3) Description of the Process in the Third Case If it is determined in step 2 that the average power value P is smaller than the threshold Th, the continuation length of the silent section up to this time is calculated (step 11). The calculated duration is the pause duration Tde set in the pause duration memory 17.
It is determined whether it is equal to or greater than 1 (step 12). And
If it is determined that the duration of the silent section is less than the pause duration Tdel, the ring memory storage amount state determination unit 1
Based on the output of No. 6, it is determined whether or not the state is immediately before underflow (steps 13 and 14).

If the ring memory 7 is not in the state immediately before the underflow, it is determined whether or not it is in the state immediately before the overflow based on the output of the ring memory storage amount state determination section 16 (steps 6 and 7). If it is not the state immediately before the overflow, the third case occurs, and the pitch compression / expansion means 23 compresses the current frame data on the time axis at the compression rate α determined by the compression / expansion rate adjusting means 42 (step 8). The compressed data is sent to and written to the ring memory 7. Thereafter, the process returns to step 1.

(4) Description of Processing in Fourth Case When it is determined in step 2 that the average power value P is smaller than the threshold Th, the continuation length of the silent section up to this time is calculated (step 11). The calculated duration is the pause duration Tde set in the pause duration memory 17.
It is determined whether it is equal to or greater than 1 (step 12). And
If it is determined that the duration of the silent section is less than the pause duration Tdel, the ring memory storage amount state determination unit 1
Based on the output of No. 6, it is determined whether or not the state is immediately before underflow (steps 13 and 14).

If the ring memory 7 is not in the state immediately before the underflow, it is determined whether or not it is in the state immediately before the overflow based on the output of the ring memory storage amount state determination section 16 (steps 6 and 7). If the state is immediately before the overflow, the fourth case occurs, and the input signal is deleted by the input signal deletion unit 21 until the underflow detection signal is output from the ring memory storage amount state determination unit 16 (step 9). That is, the ring memory 7
Ring memory 7 until
Writing to is interrupted.

When the ring memory 7 is in the state immediately before underflow, the silence insertion section 22 writes a predetermined number of silence signals "0" of 200 or less into the ring memory 7 (step 10). Then, the process returns to step 1.

(5) Description of the Process in the Fifth Case If it is determined in step 2 that the average power value P is smaller than the threshold Th, the continuation length of the silent section up to this time is calculated (step 11). The calculated duration is the pause duration Tde set in the pause duration memory 17.
It is determined whether it is equal to or greater than 1 (step 12). And
When it is determined that the duration of the silent section is equal to or longer than the pause duration Tdel, the ring memory storage amount state determination unit 1
6, it is determined whether or not the state is immediately before the underflow (steps 15 and 16).

When the ring memory 7 is not in the state immediately before the underflow, the fifth case is set, and the first flag F1 indicating that the current frame is a deletion section by the input signal deletion section 25 is set (step 17). The first flag F1 has been reset (F1 = 0) in the initial setting when the power is turned on. Then, it is determined whether or not the second flag F2 indicating whether or not the current frame is the first frame of the deletion section by the input signal deletion unit 25 has been reset (step 18).

The second flag F2 has been reset (F2 = 0) in the initial setting when the power is turned on. Then, when the processing for the first frame of the deletion section by the input signal deletion unit 25 is completed, it is set to (F2 = 1). Then, when the processing for a series of deletion sections by the input signal deletion unit 25 is completed, reset (F2 =
0).

Therefore, when the current frame is the first frame of the section to be deleted by the input signal deleting section 25, the second flag F2 is reset (F2 = 0). When the second flag F2 is reset, the current frame data is stored in the first memory 31 by the waveform synthesis insertion unit 26 (step 19). Further, the writing of the current frame data to the ring memory 7 is stopped by the input signal deleting unit 25 (step 20).
That is, the current frame data is deleted. And
After the second flag F2 is set (F2 = 1) (Step 21), the process returns to Step 1.

Further, when a silent section continues,
The process proceeds to Step 16 through Steps 2, 11, 12, and 15, and it is determined whether or not the ring memory 7 is in the state immediately before the underflow based on the output of the ring memory storage amount state determination unit 16.

When the ring memory 7 is not in the state immediately before the underflow, the current frame is set to the input signal deleting section 25.
A first flag F1 indicating that the section is a deletion section is set (step 17). Then, it is determined whether or not the second flag F2 indicating whether or not the current frame is the first frame of the deletion section by the input signal deletion unit 25 is reset (step 18).

In this case, since the second flag F2 is set (F2 = 1), it is determined that the current frame is not the first frame of the deletion section by the input signal deletion unit 25. In this case, the current frame data is stored in the second memory 32 by the waveform synthesis insertion unit 26 (step 22). Further, the writing of the current frame data to the ring memory 7 is stopped by the input signal deleting unit 25 (step 23). Then, the process returns to step 1.

When the silent section continues and the ring memory 7 is not in the state immediately before the underflow, steps 2, 11, 12, 15, 16, 17, and
The processing of 18, 22, and 23 is repeated. That is,
The frame data in the second memory 32 is updated, and the writing of the frame data to the ring memory 7 is stopped.

Thereafter, when the frame data of the voice section is input, in step 2, the average power value P
Is greater than or equal to the threshold Th, it is determined based on the state of the first flag F1 whether or not the previous frame was a deletion section by the input signal deletion unit 25 (step 3). In this case, since the first flag F1 is set (F1 = 1), it is determined that the previous frame was a deletion section by the input signal deletion unit 25, and the process proceeds to Step 4. In step 4, the deletion process by the input signal deletion unit 25 is stopped, and the waveform synthesis insertion process by the waveform synthesis insertion unit 26 is performed.

That is, as described above with reference to FIG. 4A, the content of the first memory 31 is multiplied by a function that changes linearly from 1 to 0, and the content of the second memory 32 is multiplied by 0 to 1. Is multiplied by a function that changes linearly, and the results of both multiplications are added. The result of this addition (corresponding to A ′ * B ′ in FIG. 4A) is sent to the ring memory 7 via the demultiplexer 27 and written into the ring memory 7.

Thereafter, the first flag F1 and the second flag F2 are reset (F1 = F2 = 0) (step 5). Then, the process proceeds to Step 6.

By the way, when the deletion processing by the input signal deletion section 25 as described above is repeatedly performed on the continuous silent section, the ring memory 7
May be in the state immediately before underflow. In this case, the result of step 16 is YES, and the
Move on to In step 24, it is determined whether or not the previous frame is a deletion section by the input signal deletion unit 25 by the first flag F.
1 is determined based on the state.

In this case, since the first flag F1 has been set (F1 = 1), the routine proceeds to step 25, where the current frame data is stored in the second memory 32. Then, the deletion processing by the input signal deletion unit 25 is stopped, and the waveform synthesis insertion processing by the waveform synthesis insertion unit 26 is performed (step 26). Then, the first flag F1 and the second flag F2 are reset (F1 = F2
= 0) (step 27), and then proceed to step 1.

The waveform synthesizing and inserting process performed by the waveform synthesizing and inserting unit 26 in step 26 is substantially the same as the waveform synthesizing and inserting process described in step 4 except that the frame data stored in the second memory 32 is The difference from the processing described in step 4 above is that the frame data is after the ring memory 7 is in the state immediately before the underflow.

The processing in step 25 is omitted, and
If the answer is YES in step 24, the second memory 32
Alternatively, the process may proceed to step 26 without storing the current frame data. In this case, in the waveform synthesis insertion process performed in step 26, the second
The frame data before the underflow immediately before state (the previous frame data) stored in the memory 32 is used.

Further, the processing of step 22 may be omitted, and a step of storing frame data in the second memory 32 may be added between step 3 and step 4. In this case, step 4
In the step 19, the first memory 31
And the contents stored in step 3 and step 4
The waveform synthesis insertion process is performed based on the content stored in the second memory 32 in the step added between the steps.

(6) Description of the Process in the Sixth Case When it is determined in step 2 that the average power value P is smaller than the threshold Th, the continuation length of the silent section up to this time is calculated (step 11). The calculated duration is the pause duration Tde set in the pause duration memory 17.
It is determined whether it is equal to or greater than 1 (step 12). And
When it is determined that the duration of the silent section is equal to or longer than the pause duration Tdel, the ring memory storage amount state determination unit 1
6, it is determined whether or not the state is immediately before the underflow (steps 15 and 16).

When the ring memory 7 is in the state immediately before the underflow, it is determined whether or not the previous frame is a deletion section by the input signal deletion section 25 based on the state of the first flag F1 (step 24). First flag F1
Is reset (F1 = 0), that is, if the previous frame is not the section to be deleted by the input signal deleting section 25, the sixth case occurs and the process proceeds to step.
In step 28, the thinning processing section 24 thins out the current frame data at a compression ratio of 1/2. The thinned data is sent to the ring memory 7 and written. Thereafter, the process returns to step 1.

That is, even if the duration of the silent section is equal to or longer than the pause duration Tdel, if the ring memory 7 is in the state immediately before underflow and the previous frame is not the section to be deleted by the input signal deleting section 25, the frame The data is not deleted, and after being thinned out at a compression ratio of 1/2, is written to the ring memory 7.

In FIG. 7, in step 12,
It is determined whether or not the duration of the silent section is longer than the set pause duration Tdel.
A, as shown in FIG.
Whether it is less than the reference length T1 (T <T1), or less than the second reference length T2 (T1 <T2) set to be equal to or longer than the first reference length T1 in which the silence duration T is set (T1 ≦ T <T2) ),
Or a second reference length T2 in which a continuation length T of a silent section is set.
Whether (T ≧ T2) or not may be determined. As the first reference length, for example, the length for four frames is
As the second reference length, for example, a length for 40 frames is set.

Then, as shown in FIG. 8, the process may proceed to the following steps according to each determination result.
That is, when the continuation length T of the silent section is less than the set first reference length T1 (T <T1), the process proceeds to step S13. First reference length T1 in which duration T of silent section is set
If it is less than the second reference length T2 (T1 <T2) (T1 ≦ T <T2) set as described above, the process proceeds to step 28, where the thinning is performed by the 1 / n thinning process. The duration T of the silent section is equal to or longer than the set second reference length T2 (T ≧ T
If 2), go to step 15.

FIG. 10 shows a relationship between an input signal and an output signal at the time of double speed reproduction, and particularly shows a state where an input signal in a silent section is deleted. FIGS. 11 and 12 show a data write start point to the ring memory 7, a data read start point from the ring memory 7, and points A to H in FIG.
3 shows the state of the ring memory 7 in FIG.

At the start of double-speed playback, the input signal is in a silent section and the ring memory 7 is empty (see FIG. 11A). After the data is decimated by ２, the data is written to the ring memory 7.

When the accumulated amount Tm of the ring memory 7 reaches the underflow detection data Tmin, reading of data from the ring memory 7 is started (FIG. 11).
(B)).

Then, when the frame data for the voice section a of the input signal is sent (point A), the frame data is divided by 1 /
It is compressed by the pitch compression / expansion means 23 at a compression ratio α of 2 or more. Assuming that the compression ratio α is other than １ ／, the frame data is decompressed when the compression rate α is other than す る. In this sense,
FIG. 10 illustrates the decompression process. Then, the compressed data is written to the ring memory 7. At point A, as shown in FIG.
It remains at Tmin.

Output signal a for speech section a of the input signal
1 is read out with a delay of the accumulated amount TmA at the point A. Then, at the point in time when the voice section a of the input signal has been input (point B), as shown in FIG. 11D, the accumulation amount Tmin at point A, which is the start point of the current compression section,
The sum of the compressed data of the voice section a from the point A to the point B and the decompression amount StB with respect to the compression at the compression ratio of 1/2 is the storage amount TmB (= StB + Tmin) of the ring memory 7. Therefore, the output signal a1 for the speech section a of the input signal
Is output when TmB (= StB + Tmin) has elapsed from point B.

The frame data in a silent section shorter than the pause duration Tdel following the audio section a of the input signal is also compressed by the pitch compression / expansion means 23 at a compression rate α of 1/2 or more. When a voice section b is input following the silent section, the frame data of the voice section b is also compressed by the pitch compression / expansion means 23 at a compression rate α of 1/2 or more.

At the point in time when the voice section b of the input signal has been input (point C), as shown in FIG. 11 (e), the accumulated amount Tm at point A which is the start point of the current compression section.
The sum of in and the decompression amount StC of the compressed data corresponding to the input signals from point A to point C with respect to 圧 縮 compression is the accumulated amount TmC (= StC + Tmin) of the ring memory 7. Therefore, the output signal b1 for the voice section b of the input signal ends being output at the point when TmC (= StC + Tmin) has elapsed from the point C.

When a signal in a silent section having a length equal to or longer than the pause duration Tdel is transmitted following the speech section b of the input signal, the signal reaches the pause duration Tdel (D
(Point), the frame data is compressed by the pitch compression / expansion means 23 at a compression ratio α of 1/2 or more.

At point D, as shown in FIG. 11F, the accumulation amount Tmin at point A, which is the start point of the current compression section,
Of the compressed data corresponding to the input signal from the point A to the point D and the decompression amount StD with respect to the 圧 縮 compression is the accumulated amount TmD (= StD + Tmin) of the ring memory 7.
Therefore, the output signal for the silent section between the voice section b of the input signal and the point D is TmD (= StD +
Tmin), the output ends at the point when the minute has elapsed.

The frame data in the silent section after the pause duration Tdel is deleted by the input signal deletion unit 25 until the storage amount of the ring memory 7 becomes equal to or less than the underflow detection data Tmin. The length Std of the pause deletion part is D from the point A which is the start point of the current compression section.
It becomes equal to the extension StD of the compressed data corresponding to the input signal up to the point with respect to 圧 縮 compression. Input signal deletion unit 2
After the deletion process is performed by step 5, a synthesized waveform for preventing a click sound is inserted by the waveform synthesis insertion unit 26, but the inserted synthesized waveform portion is omitted in FIG.

The last point (E
12), the accumulated amount TmE of the ring memory 7 is equal to or less than the underflow detection data Tmin, as shown in FIG. Here, an example is shown in which the accumulated amount TmE has become equal to the underflow detection data Tmin.

The frame data for the silent section from the point E is thinned out by the thinning-out processing section 24 at a compression ratio of 1/2, and then written to the frame memory 7. And
When the signal of the voice section c is input (point F), the frame data of the voice section c is compressed by the pitch compression / expansion means 23 at a compression ratio α of 1/2 or more. That is, a new compression section is started. Then, the compressed data is written to the ring memory 7.

At the point F, as shown in FIG. 12 (h), the accumulation amount TmF of the ring memory 7 is the same as that at the point E.
in.

Output signal c for speech section c of the input signal
1 is output with a delay of the accumulation amount Tmin at the point F. Pause duration Tde following the voice section c of the input signal
Frame data in a silent section less than 1 (a silent section from the voice section c to the point G) is also compressed by the pitch compression / expansion means 23 at a compression ratio α of 1/2 or more.

At point G, as shown in FIG. 12 (i), the accumulation amount Tmin at point F which is the start point of the current compression section.
Of the compressed data corresponding to the input signal from the point F to the point G and the decompression amount StG with respect to the 圧 縮 compression is the accumulated amount TmG (= StG + Tmin) of the ring memory 7.
Therefore, the output signal of the input signal for the silent section from the voice section c to the point G is TmG (= StG +
Tmin), the output ends at the point when the minute has elapsed.

The frame data in the silent section after the pause duration Tdel is deleted by the input signal deletion unit 25 until the storage amount of the ring memory 7 becomes the underflow detection data Tmin. Length S of this pause deletion part
td is equal to the decompression amount StG of the compressed data corresponding to the input signal from the point F to the point G, which is the start point of the current compression section, with respect to 圧 縮 compression.

The last point of the section from which the input signal has been deleted (H
12), the accumulated amount TmH of the ring memory 7 is equal to or less than the underflow detection data Tmin, as shown in FIG. Here, an example is shown in which the accumulated amount TmH is equal to the underflow detection data Tmin.

The frame data for the silent section from the point H is thinned out by the thinning-out processing section 24 at a compression ratio of 、, and then written into the frame memory 7. And
When the signal of the voice section d is input, the frame data of the voice section d is converted into 1 /
It is compressed at a compression ratio α of 2 or more. Then, the decompressed data is written to the ring memory 7.

FIG. 13 shows the relationship between the input signal and the output signal at the time of 2 × speed reproduction, and particularly shows how the input signal is deleted when the state immediately before the overflow occurs.
FIG. 14 shows the ring memory 7 at each of points S to U in FIG.
The state of is shown.

From a certain point of time to a point T, a voice section a,
The frame data for a series of input signals including b, c, etc. and a silent section is
It is assumed that the image data is compressed at a compression ratio α of 2 or more (when the compression ratio α is other than 1/2, the compression is extended at a compression ratio of 1/2). In this case, the extension is accumulated in the ring memory 7.

At the input start point (point S) of the voice section b, as shown in FIG. 14A, the accumulation amount Tmin at the start point of the compression processing of the series of input signals and the compression processing of the compression processing are performed. The sum of the compressed data corresponding to the input signal from the start point to the point S and the decompression amount StS with respect to 圧 縮 compression is the accumulated amount TmS (= StS + Tmin) of the ring memory 7.
Therefore, the output signal b1 for the voice section b is started to be output at the point when TmS (= StS + Tmin) has elapsed from the point S.

It is assumed that the ring memory 7 is in the state immediately before the overflow at the time (point T) when the compressed data corresponding to the input signal of the voice section c is written into the ring memory 7. That is, it is assumed that, at the point T, the accumulated amount of the ring memory 7 is equal to or larger than the overflow detection data Tmax.

At the point T, as shown in FIG. 14 (b), the accumulated amount Tmin at the start point of the compression processing for the series of input signals and the input signal from the compression processing start point to the point T are determined. The sum of the corresponding compressed data and the decompression amount StT with respect to １ ／ compression is the accumulated amount TmT of the ring memory 7.
(= StT + Tmin). In other words, the total number of words in the ring memory 7 is set to TOTAL, the overflow detection data is set to Tmax, and TOTAL and Tmax are used.
Is Dmin, the accumulated amount Tmt at point T is
Since it is equal to Tmax, it becomes TOTAL-Dmin.

Therefore, the output signal corresponding to the series of input signals starts from the point T and the accumulated amount TmT (= StT + Tmi)
n) Output is completed at the point of time delayed.

At the point T, when the ring memory 7 enters the state immediately before the overflow, the input signals thereafter are unconditionally deleted by the input signal deletion section 21 until the ring memory 7 enters the state immediately before the underflow. You.
After the deletion process is performed by the input signal deletion unit 21, silence is inserted by the silence insertion unit 22, but the inserted silence part is omitted in FIG. After the ring memory 7 enters the state immediately before overflow (T
Point), and the frame data is deleted, as shown in FIG.
It is assumed that the ring memory 7 is in a state immediately before underflow (accumulated amount TmU = Tmin) at point U as shown in FIG. In this case, an input signal including four silent sections and three voice sections d, e, and f from point T to point U is deleted. Therefore, the input signal from point T to point U does not appear as an output signal.

When the signal of the voice section g is input after the point U, the frame data for this voice section is compressed by the pitch compression / expansion means 23 at a compression rate α of 1/2 or more (the compression rate α is 1 / When the compression ratio is other than 2, the data is decompressed at a compression rate of 1/2, and then written into the ring memory 7. The output signal g for the voice section g is started to be output with a delay of the accumulation amount Tmin of the ring memory 7 at the point U.

In the above embodiment, the voice section and the silent section of the input signal are determined based on the average power value of each frame, but may be determined based on the average amplitude of each frame. In this case, as shown in FIG. 15, an average amplitude calculator 11A that calculates an average amplitude value in frame units is provided instead of the power calculator 11 of FIG. 2, and the threshold memory 13A includes, for example, A threshold value of ²⁶ is set. Then, the average amplitude value calculated by the average amplitude calculation unit 11A and the threshold value memory 13A
Is compared with the threshold value by the comparison unit 12A to determine whether it is a voice section or a silent section.

That is, if the average amplitude value is equal to or larger than the threshold value, it is determined to be a voice section, and if the average amplitude value is less than the threshold value, it is determined to be a silent section. The average amplitude value W for each frame is calculated based on the following equation (3), where i ₀ , i ₁ ,... I _N−1 (where N = 200) is the amplitude of each audio signal in one sampled frame. Is calculated.

[0202]

(Equation 3)

The other processing is the same as the processing performed by the speech speed conversion unit 6 in FIG. 2, and a description thereof will be omitted.

In this case, the threshold value may be changed in the following manner. That is,
As shown by a dotted line in FIG. 15, an average amplitude steady state detection and threshold value updating unit 14A is provided. The average amplitude steady state detection and threshold update unit 14A includes an average amplitude calculation unit 11
It is determined whether or not the average amplitude value W from A is constant over a predetermined number of frames. If the average amplitude value W is constant (steady state), a value twice the average amplitude value W at that time is stored in the threshold memory 13A. To update the threshold. However, the maximum value of the threshold to be updated, a predetermined value is limited for example to 2 ^8.

Further, the speech section and the silence section of the input signal may be determined based on the cumulative amplitude value Wa of the speech signal of each frame and the given threshold value as shown in the following Expression 4. .

[0206]

(Equation 4)

[0207] The voice section and the silent section of the input signal are detected as the voice section if the periodicity of the signal of each frame is detected and the detected cycle is within a predetermined pitch cycle range of the voice signal. If the detected cycle is outside the predetermined pitch cycle range of the audio signal, it may be determined to be a silent section.

In this case, as shown in FIG.
Is provided with a pitch cycle detecting section 11B for detecting the periodicity of each frame based on the autocorrelation method, and the pitch memory range of the voice signal is set in the threshold value memory 13B. You. Then, the cycle detected by the pitch cycle detecting unit 11B and the threshold memory 13
The comparison unit 12B compares the pitch cycle range of the audio signal set to B.

The range of the pitch period of the audio signal to be set is as follows.
It depends on the reproduction speed, and is set, for example, in the range of 66 × n (Hz) to 320 × n (Hz) at the time of n × speed reproduction. Therefore, at the time of double speed reproduction, the pitch cycle range of the audio signal is set to a range of 132 Hz to 640 Hz. Other processes are the same as the processes performed by the speech speed conversion unit 6 in FIG. 2, and thus description thereof is omitted.

[0210] Further, the voice section and the silent section of the input signal may be determined by comparing the power spectrum of the signal of each frame with the power spectrum of the steady state.

In this case, as shown in FIG.
Is provided with a power spectrum calculator 11C for calculating a power spectrum for one or a plurality of predetermined frequency bands for each frame.
The power spectrum in the steady state for the predetermined one or a plurality of frequency bands is stored in the power spectrum storage unit 1.
3C.

The contents of the power spectrum storage section 13C are as follows.
When the power spectrum steady state detection unit 14B detects that the power spectrum is in the steady state based on the change state of the power spectrum calculated by the power spectrum calculation unit 11C, the power spectrum is updated to the detected power spectrum in the steady state.

When the input signal is the power spectrum calculator 11C
, A power spectrum for one or more predetermined frequency bands is calculated for each frame.
Then, the comparison unit 12C compares the calculated power spectrum with the power spectrum in the steady state stored in the power spectrum storage unit 13C.

If the calculated power spectrum fluctuates with respect to the steady-state power spectrum, the frame is determined to be a voice section. Conversely, if the calculated power spectrum does not fluctuate from the steady-state power spectrum, the frame is determined to be a silent section.

Specifically, the power spectrum storage unit 13
C includes the predetermined one based on the steady-state power spectrum for the predetermined one or more frequency bands.
Alternatively, threshold values for a plurality of frequency bands are stored. Then, it is stored in the power spectrum storage unit 13C. The power spectrum for the predetermined one or more frequency bands calculated by the power spectrum calculation unit 11C is compared with the corresponding threshold value stored in the power spectrum storage unit 13C, so that the input signal is It is determined whether the section is a section or a silent section.

For example, it is assumed that the power spectrum in the steady state is a power spectrum of only noise as shown in FIG. Also, it is assumed that the power spectrum of the voice without noise is shown in FIG. In the steady state, when noise shown by the power spectrum of FIG.
When an audio signal having the power spectrum shown by 1 (b) is input, the power spectrum is obtained by combining both power spectra as shown in FIG. 21 (c).

Therefore, for example, a frequency band f in which power is relatively small in a power spectrum in a steady state.
The powers for a and fb increase significantly in the power spectrum of the voice section. That is, by comparing the steady-state power in one or more frequency bands having relatively small power in the steady-state power spectrum with the power in the one or more frequency bands of the power spectrum of the input signal, Is a voice section or a silent section.

If it is known that the steady-state noise is noise in a high frequency band, a power spectrum for a low frequency band (for example, a frequency band of 4 KHz or less) where the influence of the noise is small is calculated. Depending on whether the calculated power spectrum is equal to or greater than a predetermined threshold,
It is also possible to determine whether the input signal is a voice section or a silent section.

Further, by comparing the average power value P of each frame with the threshold value Th, when discriminating between a voice section and a silent section, the threshold value is determined based on the storage amount of the ring memory 7. Th may be changed. That is, the threshold value T is set such that the smaller the storage amount of the ring memory 7, in other words, the larger the empty area of the ring memory 7, the smaller the missing portion of the voice section.
h is reduced. This makes the output sound closer to nature.

That is, as shown in FIG. 22, a threshold adjusting means 51 is provided. The threshold adjusting unit 51 obtains the storage amount of the ring memory 7 from the ring memory storage amount state determination unit 16. Then, the storage time Tm is calculated by dividing the obtained storage amount of the ring memory 7 by the sampling frequency of the D / A converter 8. Then, the threshold value Th is determined based on the calculated accumulation time Tm, and the content of the threshold value memory 13 is updated.

More specifically, the storage amount of the ring memory 7 obtained from the ring memory storage amount state determination unit 16 is 800, which is the sampling frequency of the D / A conversion unit 8.
By dividing by 0, the accumulation time Tm is obtained.
Then, the threshold value Th for the accumulation time Tm is obtained based on the data of the threshold value Th for the accumulation time Tm created in advance.

The following table shows an example of data of the threshold value Th with respect to the accumulation time Tm when the number of quantization bits of the A / D converter 2 is 12 bits.

[0223]

[Table 3]

When the speech section and the silent section are discriminated by comparing the power accumulated value Pa of each frame with the threshold value, the average amplitude value W of each frame is compared with the threshold value. When discriminating between the voice section and the silent section, the power section of each frame is compared with the threshold value by comparing the amplitude cumulative value Wa of each frame with the threshold value. Is also determined in the same manner as described above.
The threshold value may be changed based on the accumulated amount of.

Further, the pause continuation length Tdel for determining the start point of deletion of a silent section may be changed based on the storage amount of the ring memory 7. That is,
The pause continuation length Tdel is increased such that the smaller the storage amount of the ring memory 7, in other words, the larger the empty area of the ring memory 7, the smaller the number of silence sections to be deleted. This makes the output sound closer to nature.

That is, as shown in FIG. 22, a pause continuation length adjusting means 52 is provided. Pause continuation length adjusting means 52
Obtains the storage amount of the ring memory 7 from the ring memory storage amount state determination unit 16. And the obtained ring memory 7
Is divided by the sampling frequency of the D / A converter 8 to calculate the accumulation time Tm. Then, based on the calculated accumulation time Tm, the pause duration Tde
is determined, and the contents of the pause continuation length setting memory 17 are updated.

More specifically, the storage amount of the ring memory 7 obtained from the ring memory storage amount state determination unit 16 is 800, which is the sampling frequency of the D / A conversion unit 8.
By dividing by 0, the accumulation time Tm is obtained.
Then, the pause duration Tdel for the accumulation time Tm is obtained based on the data of the pause duration Tdel for the accumulation time Tm created in advance.

The following table shows an example of the data of the pause duration Tdel with respect to the accumulation time Tm at the time of double speed reproduction of the VTR.

[0229]

[Table 4]

In the above, the case where the input signal is an analog signal has been described. However, the present invention can be applied to a case where the input signal is digital data. For example,
When a compressed digital audio signal is sent from an IC memory, a magnetic disk, a digital communication line, or the like, the compressed digital audio signal is expanded and the PC
The converted PCM audio signal is temporarily stored in a buffer. Thereafter, the PCM audio data is read from the buffer at a speed corresponding to the set reproduction speed magnification, and is sent to the frame memory 5 in FIG.

Next, an embodiment in which the speech speed conversion device according to the present invention is applied to a TV telephone will be described. Note that the following speech speed converter 109 corresponds to the speech speed converter shown in FIG.

FIG. 23 is a schematic configuration diagram of a TV telephone with a speech speed conversion function to which the present invention is applied. An input signal in which video and audio are mixed is separated into a video signal and an audio signal by a signal separation / synthesis unit 101, and transmitted to a video signal processing block 102 and an audio signal processing block 103, respectively, for processing. The video and audio signals processed by the respective signal processing blocks 102 and 103 are combined by the signal separation / combination unit 101 to become an output signal in which the video signal and the audio signal are mixed.

Next, the processing in the video signal processing block 102 will be described. The video input signal separated by the signal separation / synthesis unit 101 is received by the video reception unit 104, and the video is displayed on a monitor 105. Further, the video image captured by the camera 107 is transmitted by the video transmission unit 106 to the signal separation / combination unit 101 as a video signal.

Similarly, in the audio signal block 103, the audio input signal separated by the signal separation / combination unit 101 is received by the receiving unit 108, and the audio is converted to a speed that is easy for the receiver to hear by the speech speed conversion device 109 according to the present invention. The voice is uttered by the speaker 110 as the speed-controlled voice. The voice collected by the microphone 112 is transmitted to the signal separation / combination unit 101 as a voice signal by the transmission unit 111. At this time, it is preferable that the video and the audio mixed in the input / output signal have no time lag, and the speech speed conversion device 109 according to the present invention can provide a TV telephone with the minimum lag between the video and the audio.

In addition, even when the TV telephone does not transmit and receive a video signal, and transmits and receives only an audio signal as in a telephone, the speech speed conversion device 109 according to the present invention makes it possible to determine the timing of the conversation between the transmitter and the receiver. Needless to say, it is possible to provide a telephone with a speech speed conversion function that is less likely to be shifted and that can change the speech speed at a speed that is easy for the listener to hear.

[0236]

According to the present invention, it is possible to obtain a speech speed conversion device capable of reducing the processing load, reducing the difference between the video and the audio, and not increasing the capacity of the memory for storing the audio signal.

Further, according to the present invention, it is possible to reduce the number of missing voices in the voice section of the input signal as much as possible.
The audio reproduction speed for the audio in the audio section can be reduced with respect to the set reproduction speed magnification.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an overall configuration of a speech speed conversion device.

FIG. 2 is a block diagram illustrating a configuration of a speech speed conversion unit.

FIG. 3 shows that an input signal is compressed at a compression ratio of 2 /
FIG. 3 is an explanatory diagram showing a method of performing compression in No. 3;

FIG. 4 is an explanatory diagram for describing processing by a waveform synthesis processing unit;

FIG. 5 is an explanatory diagram for explaining various thinning processing methods performed by a thinning processing unit.

FIG. 6 is a flowchart illustrating a processing procedure by a speech speed conversion unit;

FIG. 7 is a flowchart illustrating a processing procedure by a speech speed conversion unit;

FIG. 8 is a flowchart illustrating a modified example of the processing procedure by the speech speed conversion unit and corresponding to FIG. 7;

FIG. 9 is an explanatory diagram for explaining a process that can be replaced with the process of step 10 in FIG. 6;

FIG. 10 is a time chart showing a relationship between an input signal and an output signal at the time of double-speed reproduction, and particularly showing a state in which an input signal in a silent section is deleted.

FIG. 11 shows a starting point of data writing to the ring memory 7,
FIG. 11 is a schematic diagram showing a state of starting reading data from the ring memory 7 and a state of the ring memory 7 at points A to D in FIG. 10.

12 is a schematic diagram showing the state of the ring memory 7 at points E to H in FIG.

FIG. 13 shows a relationship between an input signal and an output signal at the time of 2 × speed reproduction.
6 is a time chart illustrating a state in which an input signal is deleted.

14 is a diagram showing a ring memory 7 at each of points S to U in FIG. 13;
It is a schematic diagram which shows the state of.

FIG. 15 is a block diagram illustrating a modified example of a circuit for determining a voice section and a silent section, corresponding to FIG. 2;

FIG. 16 is a block diagram showing another modified example of the circuit for distinguishing between a voice section and a silent section, and corresponds to FIG. 2;

FIG. 17 shows a compression ratio of 2 /
FIG. 3 is an explanatory diagram showing a method of performing compression in No. 3;

FIG. 18 is an explanatory diagram for explaining a process that can be replaced with the process of step 9 in FIG. 6;

19 is an explanatory diagram for explaining a process that can be replaced with the process of step 10 of FIG. 6 when the process of FIG. 18 is adopted as the process of step 9 of FIG. 6;

FIG. 20 is a block diagram showing still another modification of the circuit for discriminating between a voice section and a silent section, corresponding to FIG. 2;

FIG. 21 is a graph showing a power spectrum in a steady state, a power spectrum of a voice without noise, and a power spectrum of a voice section.

FIG. 22 is a block diagram showing a speech speed conversion unit to which a threshold adjustment unit and a pause duration adjustment unit are added.

FIG. 23 is a schematic configuration diagram of a TV phone with a speech speed conversion function to which the present invention is applied.

[Explanation of symbols]

 Reference Signs List 2 A / D converter 4 DSP 5 Frame memory 6 Speech speed converter 7 Ring memory 8 D / A converter 9 Up / down counter 11 Power calculator 11A Average amplitude calculator 11B Pitch cycle detector 11C Power spectrum calculator 12 , 12A, 12B, 12C Comparing unit 15 Conditional branching unit 16 Ring memory storage amount state discriminating unit 21, 25 Input signal deleting unit 23 Pitch compression / expansion unit 24 Decimation processing unit 42 Compression / expansion rate adjusting unit 51 Threshold adjusting unit 52 Pause Duration adjustment means

──────────────────────────────────────────────────続 き Continuation of the front page (72) Inventor Masanori Miyatake 2-5-5 Keihanhondori, Moriguchi-shi, Osaka Sanyo Electric Co., Ltd. (56) References JP 3-205656 (JP, A) JP Hei 8-137492 (JP, A) (58) Field surveyed (Int. Cl. ⁷ , DB name) G10L 21/04 G11B 20/02

Claims (21)

(57) [Claims] 1. A speech speed conversion processing means for performing speech speed conversion processing of an input voice signal, a ring memory in which an output of the speech speed conversion processing means is written,
And a means for reading data from the ring memory at a constant speed. When the input voice signal is in the voice section and the ring memory is not in the state immediately before the overflow, the speech speed conversion processing means sets the playback speed magnification to 1 / n Means for performing a compression / expansion process on the input audio signal at the compression rate determined according to the program type set by the operator. 2. A / D conversion means for sampling an input analog audio signal at a sampling frequency corresponding to a set reproduction speed magnification, a frame memory to which an audio signal output from the A / D conversion means is input, Each time a required number of voice signals are input to the frame memory, voice speed conversion processing means for performing voice speed conversion processing on those voice signals, a ring memory in which the output of the voice speed conversion processing means is written, 1 × speed reproduction Reading means for reading data from the ring memory based on a read signal having a frequency equal to the sampling frequency at the time, and storage amount calculating means for calculating the storage amount of the ring memory based on a write signal and a read signal of the ring memory. The speech speed conversion processing means includes an input voice corresponding to a required number of voice signals input to the frame memory. A signal for performing compression / expansion processing or deletion processing on the required number of audio signals in accordance with an output of the section determination means and an output of the accumulation amount calculation means, When the input sound is a voice section and the ring memory is not in a state immediately before overflow, the signal processing means has a compression ratio of 1 / n or more, where n is a set reproduction speed magnification, and is set by an operator. A speech speed conversion device including means for performing a compression / expansion process at a compression rate determined according to a selected program type. 3. A frame memory in which an input digital audio signal is written at a speed corresponding to a set reproduction speed magnification. Each time a required number of audio signals are input to the frame memory, the digital audio signal is Speed conversion processing means for performing voice speed conversion processing, a ring memory to which the output of the voice speed conversion processing means is written, reading means for reading data from the ring memory at a constant speed, and a write signal and a read signal of the ring memory. The speech rate conversion processing means determines whether the input voice corresponding to the required number of voice signals input to the frame memory is a voice section or a silent section. Section discriminating means for discriminating, and according to the output of the section discriminating means and the output of the accumulation amount calculating means, Signal processing means for performing a decompression / expansion processing or a deletion processing, wherein the signal processing means performs compression of 1 / n or more with the set reproduction speed magnification being n when the input voice is a voice section and the ring memory is not in a state immediately before overflowing. A speech speed conversion device including means for performing compression / expansion processing at a compression rate determined according to a program type set by an operator. 4. A speech speed conversion processing means for performing speech speed conversion processing of an input voice signal, a ring memory in which an output of the speech speed conversion processing means is written,
And a means for reading data from the ring memory at a constant speed. When the input voice signal is in the voice section and the ring memory is not in the state immediately before the overflow, the speech speed conversion processing means sets the playback speed magnification to 1 / n Means for performing compression / expansion processing on the input audio signal at the compression rate determined in accordance with the program type and the amount of storage in the ring memory, the compression rate being set by the operator. Conversion device. 5. An A / D converter for sampling an input analog audio signal at a sampling frequency corresponding to a set reproduction speed magnification, a frame memory to which an audio signal output from the A / D converter is input, Each time a required number of voice signals are input to the frame memory, voice speed conversion processing means for performing voice speed conversion processing on those voice signals, a ring memory in which the output of the voice speed conversion processing means is written, 1 × speed reproduction Reading means for reading data from the ring memory based on a read signal having a frequency equal to the sampling frequency at the time, and storage amount calculating means for calculating the storage amount of the ring memory based on a write signal and a read signal of the ring memory. The speech speed conversion processing means includes an input voice corresponding to a required number of voice signals input to the frame memory. A signal for performing compression / expansion processing or deletion processing on the required number of audio signals in accordance with an output of the section determination means and an output of the accumulation amount calculation means, When the input sound is a voice section and the ring memory is not in a state immediately before overflow, the signal processing means has a compression ratio of 1 / n or more, where n is a set reproduction speed magnification, and is set by an operator. A speech speed conversion device including means for performing compression / expansion processing at a compression ratio determined in accordance with a selected program type and an amount of storage in a ring memory. 6. A frame memory in which an input digital audio signal is written at a speed corresponding to a set reproduction speed magnification. Speed conversion processing means for performing voice speed conversion processing, a ring memory to which the output of the voice speed conversion processing means is written, reading means for reading data from the ring memory at a constant speed, and a write signal and a read signal of the ring memory. The speech rate conversion processing means determines whether the input voice corresponding to the required number of voice signals input to the frame memory is a voice section or a silent section. Section discriminating means for discriminating, and according to the output of the section discriminating means and the output of the accumulation amount calculating means, Signal processing means for performing a decompression / expansion processing or a deletion processing, wherein the signal processing means performs compression of 1 / n or more with the set reproduction speed magnification being n when the input voice is a voice section and the ring memory is not in a state immediately before overflowing. A speech speed conversion device including means for performing a compression / expansion process at a compression ratio determined according to a program type and a storage amount of a ring memory set by an operator. 7. A / D conversion means for sampling an input analog audio signal at a sampling frequency corresponding to a set reproduction speed magnification, a frame memory to which an audio signal output from the A / D conversion means is input, Each time a required number of voice signals are input to the frame memory, voice speed conversion processing means for performing voice speed conversion processing on those voice signals, a ring memory in which the output of the voice speed conversion processing means is written, 1 × speed reproduction Reading means for reading data from the ring memory based on a read signal having a frequency equal to the sampling frequency at the time, and storage amount calculating means for calculating the storage amount of the ring memory based on a write signal and a read signal of the ring memory. The speech speed conversion processing means includes an input voice corresponding to a required number of voice signals input to the frame memory. A signal for performing compression / expansion processing or deletion processing on the required number of audio signals in accordance with an output of the section determination means and an output of the accumulation amount calculation means, Signal processing means, when the input voice is a voice section and the ring memory is not in a state immediately before overflow,
When the fixed compression rate mode is selected, the compression / expansion process is performed at a compression rate of 1 / n or more, where n is the set reproduction speed magnification, and is determined according to the program type set by the operator. When the compression rate fluctuation mode is selected, the set reproduction speed magnification is set to n and 1 / n
A speech speed conversion device including means for performing a compression / expansion process at the above-mentioned compression rate and at a compression rate determined according to the program type and the storage amount of the ring memory set by the operator. 8. A frame memory in which an input digital audio signal is written at a speed corresponding to a set reproduction speed magnification, and each time a required number of audio signals are input to the frame memory, the digital audio signal is Speed conversion processing means for performing voice speed conversion processing, a ring memory to which the output of the voice speed conversion processing means is written, reading means for reading data from the ring memory at a constant speed, and a write signal and a read signal of the ring memory. The speech rate conversion processing means determines whether the input voice corresponding to the required number of voice signals input to the frame memory is a voice section or a silent section. Section discriminating means for discriminating, and according to the output of the section discriminating means and the output of the accumulation amount calculating means, A signal processing means for performing a reduced expansion processing or deletion processing, the signal processing means, at the time the input speech is the speech section and the ring memory is not overflow state immediately before,
When the fixed compression rate mode is selected, the compression / expansion process is performed at a compression rate of 1 / n or more, where n is the set reproduction speed magnification, and is determined according to the program type set by the operator. When the compression rate fluctuation mode is selected, the set reproduction speed magnification is set to n and 1 / n
A speech speed conversion device including means for performing a compression / expansion process at the above-mentioned compression rate and at a compression rate determined according to the program type and the storage amount of the ring memory set by the operator. 9. The signal processing means includes: (1) a first mode in which the input voice is a voice section and the ring memory is not in a state immediately before overflowing, based on an output of the section determination means and an output of the storage amount calculation means; 2) the second mode in which the input voice is a voice section and the ring memory is in a state immediately before overflow; (3) the input voice is a silent section and the duration of the silent section is less than a predetermined silent deletion start point discriminating value. And a third mode in which the ring memory is not in the state immediately before the overflow, and (4) the input voice is in the silent section, the duration of the silent section is less than the predetermined silent deletion start point determination value, and the ring memory is in the state immediately before the overflow. A fourth mode, (5) the input voice is a silent section, the duration of the silent section is greater than or equal to a predetermined silent deletion start point determination value, and (5) The input mode is a silent section, the duration of the silent section is greater than or equal to a predetermined silent deletion start point determination value, and the ring memory is in a state immediately before the underflow. A mode discriminating means for discriminating which of the sixth mode and the third mode. When the mode is determined to be the first mode or the third mode, the set reproduction speed magnification is set to n and the compression ratio is 1 / n or more. So,
First processing means for performing a compression / expansion process at a compression ratio determined according to a program type set by an operator; when the mode is determined to be the second mode or the fourth mode, the storage amount of the ring memory is immediately before the underflow; A second processing unit for deleting the audio signal until the state is reached; a third processing unit for deleting the audio signal in a silent section when the fifth mode is determined; and a setting when the sixth mode is determined. Playback speed magnification is n
And a fourth processing means for performing a compression / expansion process at a compression ratio of 1 / n ± α (where α is a value of 0 or more and 1 or less). The speech speed conversion device according to 1. 10. The signal processing means includes: (1) a first mode in which an input voice is a voice section and a ring memory is not in a state immediately before an overflow, based on an output of the section determining means and an output of the storage amount calculating means; 2) the second mode in which the input voice is a voice section and the ring memory is in a state immediately before overflow; (3) the input voice is a silent section and the duration of the silent section is less than a predetermined silent deletion start point discriminating value. And a third mode in which the ring memory is not in the state immediately before the overflow, and (4) the input voice is in the silent section, the duration of the silent section is less than the predetermined silent deletion start point determination value, and the ring memory is in the state immediately before the overflow. A fourth mode, (5) the input voice is a silent section, the duration of the silent section is equal to or greater than a predetermined silent deletion start point determination value, and A fifth mode in which the memory is not in the state immediately before the underflow; and (6) the input voice is in the silent section, the duration of the silent section is equal to or greater than the predetermined silent deletion start point determination value, and the ring memory is in the state immediately before the underflow. A mode discriminating means for discriminating which of the sixth mode and the third mode. When the mode is determined to be the first mode or the third mode, the set reproduction speed magnification is set to n and the compression ratio is 1 / n or more. So,
First processing means for performing compression / expansion processing at a compression rate determined according to the program type set by the operator and the amount of storage in the ring memory; when the mode is determined to be the second mode or the fourth mode, A second processing unit for deleting the audio signal until the accumulated amount becomes a state immediately before the underflow; a third processing unit for deleting an audio signal in a silent section when the fifth mode is determined; and a sixth mode. Is set, the set playback speed magnification is set to n
And a fourth processing means for performing a compression / expansion process at a compression ratio of 1 / n ± α (where α is a value of 0 or more and 1 or less). The speech speed conversion device according to 1. 11. The signal processing means includes: (1) a first mode in which an input voice is a voice section and a ring memory is not in a state immediately before an overflow, based on an output of the section determining means and an output of the storage amount calculating means; 2) the second mode in which the input voice is a voice section and the ring memory is in a state immediately before overflow; (3) the input voice is a silent section and the duration of the silent section is less than a predetermined silent deletion start point discriminating value. And a third mode in which the ring memory is not in the state immediately before the overflow, and (4) the input voice is in the silent section, the duration of the silent section is less than the predetermined silent deletion start point determination value, and the ring memory is in the state immediately before the overflow. A fourth mode, (5) the input voice is a silent section, the duration of the silent section is equal to or greater than a predetermined silent deletion start point determination value, and A fifth mode in which the memory is not in the state immediately before the underflow; and (6) the input voice is in the silent section, the duration of the silent section is equal to or greater than the predetermined silent deletion start point determination value, and the ring memory is in the state immediately before the underflow. A mode discriminating means for discriminating which of the following six modes: When the mode is determined to be the first mode or the third mode, and when the fixed compression ratio mode is selected, the set reproduction speed magnification is set. Where n is a compression rate of 1 / n or more, the compression / decompression processing is performed at a compression rate determined according to the program type set by the operator, and when the compression rate fluctuation mode is selected, the set reproduction is performed. The compression rate is 1 / n or more, where n is the speed magnification, and the compression rate is determined according to the program type set by the operator and the compression rate determined according to the amount of storage in the ring memory. A first processing means for performing a long process; a second processing means for deleting an audio signal until the storage amount of the ring memory becomes a state immediately before an underflow when the second mode or the fourth mode is determined; A third processing means for deleting the audio signal in the silent section when determined, and a set reproduction speed magnification of n when determined to be the sixth mode.
And a fourth processing means for performing a compression / expansion process at a compression ratio of 1 / n ± α (where α is a value of 0 or more and 1 or less). The speech speed conversion device according to 1. 12. The section discriminating means includes means for calculating a power average value of a required number of audio signals input to the frame memory, and based on the calculated power average value and a given threshold value. Determining means for determining whether the input voice is a voice section or a silent section.
12. The speech speed conversion device according to any one of claims 11 and 11. 13. The section discriminating means includes means for calculating a cumulative power value of a required number of audio signals input to a frame memory, and based on the calculated cumulative power value and a given threshold value. Determining means for determining whether the input voice is a voice section or a silent section.
12. The speech speed conversion device according to any one of claims 11 and 11. 14. The section discriminating means includes means for calculating an amplitude average value of a required number of audio signals input to the frame memory, and based on the calculated amplitude average value and a given threshold value. Determining means for determining whether the input voice is a voice section or a silent section.
12. The speech speed conversion device according to any one of claims 11 and 11. 15. The section discriminating means includes means for calculating a cumulative amplitude value of a required number of audio signals input to the frame memory, and based on the calculated cumulative amplitude value and a given threshold value. Determining means for determining whether the input voice is a voice section or a silent section.
12. The speech speed conversion device according to any one of claims 11 and 11. 16. The section discriminating means includes: detecting means for detecting the periodicity of a required number of audio signals input to the frame memory; and determining whether the input voice is a voice section or a silent section based on the detected cycle. And determining means for determining.
12. The speech speed conversion device according to any one of claims 11 and 11. 17. The section discriminating means includes: calculating means for calculating a power spectrum of a required number of audio signals input to the frame memory for one or more predetermined frequency bands; and calculating the power spectrum and a given power spectrum. Determining means for determining whether the input voice is a voice section or a silent section based on the threshold value.
12. The speech speed conversion device according to any one of claims 11 and 11. 18. The apparatus according to claim 12, wherein said threshold value is adjusted according to the amount of storage in said ring memory.
18. The speech speed conversion device according to any one of claims 17 and 17. 19. The speech speed conversion device according to claim 9, wherein said first processing means performs compression / expansion processing in units of a pitch cycle or in integral multiples of a pitch cycle. 20. A speech speed conversion apparatus according to claim 9, wherein said first processing means performs compression / expansion processing in fixed frame length units. 21. The speech speed conversion device according to claim 9, wherein the silence deletion start point discrimination value is adjusted according to the amount of storage in the ring memory.