JPH07153188A - Voice reproducing device - Google Patents

Abstract

PURPOSE:To efficiently reproduce a voice being easy to listen by analyzing voice data previously read by plural analyzing signal processing means and performing the prescribed voice processing. CONSTITUTION:A previous reading reproducing control means 12 previously reads voice data read previous to voice data reproduced by a reproducing control means 18 as necessary, and an uttering speed arithmetic means 15 estimates an uttering speed for the voice data. A storage means 16 stores addresses of a voice section and a non-voice section discriminated by a voice/non-voice discriminating means 14 and voice speed information estimated by the means 15. A voice output control means 17 decides a data length deleting and adding a non-voice section based on a analyzed result of given reproducing conditions and previously read voice data stored in the means 16, further, decides parameters of speed conversion processing changing an uttering speed of the voice section. Thereby, voice can be made easy to listen.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention
The present invention also relates to an audio reproducing device for reproducing in an easily audible manner.

[0002]

2. Description of the Related Art Conventionally, for the purpose of efficiently or repeatedly listening to a voice signal recorded on a tape recorder or the like, special playback such as fast playback, slow playback playback, skip, repeat playback, etc. is performed. A voice reproducing device capable of performing is used.

A conventional audio reproducing apparatus as described above will be described below with reference to the drawings. FIG. 4 shows the structure of a conventional audio reproducing apparatus. In FIG. 4, 41 is a magnetic tape reproducing device, 42 is a buffer memory, 43
Is a memory control circuit, and 44 is a reproduction control circuit. The operation of the audio reproducing apparatus configured as above will be described below. First, the magnetic tape reproducing device 41 has a reproduction control circuit 44.
The reproduction speed is controlled by and the read audio signal is periodically written in the buffer memory 42 under the control of the memory control circuit 43. Then, the reproduction control circuit 44 controls the memory control circuit 43 and the magnetic tape reproducing device 41 so that predetermined data is read and read by the buffer memory 42 based on the information of the set reproduction speed.

Under the control of the reproduction control circuit 44, the memory control circuit 43 controls the buffer memory 42 so as to periodically output a part of the input signal of the buffer memory 42. According to the data reproducing method using the buffer memory 42,
It is possible to reproduce an audio signal even with signals of various reproduction speeds without changing the pitch. FIG. 5 schematically shows a method of reading audio data when the reproduction speed is changed. In FIG. 5, each number box indicates a data block of a specific time length. This time length depends on the capacity of the memory used, but here it is 3 seconds. Normally, all the same data as the recorded audio data can be output during playback, but 1,2,3 during double speed playback.
Audio signals of blocks 5, 7, and 9 can be output, but
The audio signals of blocks 4, 6 and 8 are discarded and not reproduced. Similarly, during 5 × speed reproduction, blocks 1, 5 and 10 are output, and all blocks in between are discarded. In order to increase the reproduction speed in this way, the ratio of data to be discarded may be increased.

[0005]

However, in the above configuration, when the time axis is compressed to increase the speed, consonants and the like are lost to thin out the data, resulting in a decrease in clarity and further block. The connection point of is discontinuous, and although the connection point is muted in order to reduce it, there is a problem that only amplitude and phase discontinuous voices with poor naturalness can be obtained. Also, for signals that do not need to be reproduced, such as in silent sections, time-based compression and
Since the decompression process is performed, there is much waste in efficiently understanding the voice information. In view of the above problems, the present invention performs a speed conversion process in consideration of a non-voice section and a voice section to expand / contract the time axis based on the analysis result obtained by prefetching to obtain a desired reproduction speed, and at the same time formant emphasis. It is an object of the present invention to provide a voice reproducing device that can obtain variable-speed reproduced voice that is easy to hear by applying signal processing such as or consonant enhancement to an effective portion.

[0006]

In order to achieve the above object, the audio reproducing apparatus of the present invention analyzes preread audio data by a plurality of analysis signal processing means, and the result is comprehensively analyzed by the audio output control means. The audio processing is performed by the reproduction control means of the audio data and the signal processing means. Further, the voice reproduction device of the present invention analyzes the preread voice data by using the voice / non-voice discriminating means and the utterance speed calculating means, and uses the result to delete or add the non-voice section by the voice output controlling means. And the rate of changing the speaking rate in the voice section are determined. Further, the voice reproduction apparatus of the present invention analyzes the preread voice data by using the voice / non-voice discriminating means, and based on the result, the ratio and the formant emphasis at which the non-voice section is deleted and added by the voice output control means. It is configured to determine the section for performing. Further, the voice reproduction device of the present invention analyzes the preread voice data by using the phonological analysis means, and using the result, the rate at which the non-voice section is deleted and added by the voice output control means and the section for consonant emphasis. It is configured to determine.

[0007]

With these configurations, the non-speech section is deleted and added, and the signal processing such as speed conversion processing, consonant emphasis, and formant emphasis is applied to each effective portion to process the sound. , It is possible to reproduce voice information more efficiently and easily heard.

[0008]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment of the present invention will be described below with reference to the drawings. FIG. 1 shows a block diagram of an audio reproducing apparatus in a first embodiment of the present invention. Figure 1
In FIG. 11, 11 is a voice storage medium, 12 is a first read-ahead reproduction control means, 13 is a second read-ahead reproduction control means, 14 is a voice / non-voice discrimination means, 15 is a speech rate calculation means, 16 is a storage means, 17 Is audio output control means, 18 is reproduction control means, and 19 is speed conversion processing means.

The operation will be described below. It is assumed that the audio signal is already recorded in the audio storage medium 11. Before outputting the recorded audio signal, the audio signal is analyzed in advance, and based on the result, the operations of the reproduction control means 18 and the speed conversion processing means 19 are changed.
The first pre-reading reproduction control means 12 pre-reads the audio data prior to the audio data reproduced by the reproduction control means 18 as necessary, and the audio / non-voice discrimination means for the audio data.
14 determines the voice section / non-voice section. The second read-ahead reproduction control means 13 also pre-reads the voice data prior to the voice data reproduced by the reproduction control means 18, if necessary, and the utterance speed calculation means 15 estimates the utterance speed for the voice data. Further, the storage means 16 stores the address of the voice section and the address of the non-voice section judged by the voice / non-voice discrimination means 14 and the information of the utterance speed estimated by the utterance speed calculation means 15. The voice output control means 17 determines the data length for deleting and adding the non-voice section based on the given reproduction condition and the analysis result of the preread voice data stored in the storage means 16, and further determines the voice section. The parameters of the speed conversion process for changing the utterance speed are determined. This determination method can be effectively obtained for a given reproduction condition by analyzing in advance the vocalization rate and the non-voice section information of the audio data to be reproduced. That is, when performing the speed conversion processing for independently expanding and contracting the non-voice section and the voice section on the time axis, each parameter can be determined in consideration of the ratio of the time when the voice data exists. This allows
It is possible to realize a natural speech rate conversion process similar to the case where a human changes the speech rate. Further, even if a reproducing condition such as "repeating two sentences and repeatedly reproducing" is given, the audio data address to be reproduced in the audio storage medium can be easily obtained by referring to the storage means. On the other hand, even when a reproduction condition such as “play the next sentence currently being reproduced” is given, if the number of prefetched data is sufficiently large, refer to the storage means as in the previous example. The voice data address to be reproduced in is easily obtained.

As described above, according to the present embodiment, the result of analysis of the preread voice data by the voice / non-voice discrimination means and the utterance speed calculation means is stored, so that the non-voice section is deleted. ， The parameter setting of the speed conversion processing including addition can be given more accurately, and at the same time,
This makes it possible to easily perform skip reproduction and repeated reproduction in units of voice sections.

Next, a second embodiment of the present invention will be described with reference to the drawings. FIG. 2 shows a block diagram of an audio reproducing apparatus according to the second embodiment of the present invention. Figure 2
In FIG. 11, 11 is a voice storage medium, 12 is a first read-ahead reproduction control means, 14 is a voice / non-voice discrimination means, 16 is a storage means, 17 is a voice output control means, 18 is a reproduction control means, and 20 is a formant emphasis means. Is.

The operation will be described below. The basic operation is the same as that of the first embodiment, the voice signal is analyzed in advance before the recorded voice signal is output, and based on the result, the reproduction control means 18 and the formant emphasizing means 20 to be performed later. Change the behavior of. First prefetch reproduction control means 12
If necessary, the audio data prior to the audio data reproduced by the reproduction control means 18 is pre-read, and the audio / non-audio discrimination means 14 determines the audio section / non-voice section for the audio data. Further, the storage means 16 is a voice / non-voice discrimination means 14
The information of the address of the voice section and the address of the non-voice section determined by is stored. The voice output control means 17 determines the data length for deleting and adding the non-voice section based on the given reproduction condition and the analysis result of the preread voice data stored in the storage means 16, and further determines the voice section. Determine the formant enhancement parameters for. The formant enhancement processes only the voice section in order to make the voice signal easier to hear, and does not process the non-voice section in order to prevent the characteristic change of the signal. These determination methods can be effectively obtained for given reproduction conditions by analyzing in advance the information of the non-voice section of the audio data to be reproduced. As described above, according to the present embodiment, by storing the result of analysis of the preread voice data by the voice / non-voice discriminating means, it is possible to effectively delete and add the non-voice section and formant enhancement. At the same time, skip playback and repeat playback can be easily performed in voice section units.

Next, a third embodiment of the present invention will be described with reference to the drawings. FIG. 3 shows a block diagram of an audio reproducing apparatus in a third embodiment of the present invention. Figure 3
In the figure, 11 is a voice storage medium, 12 is a first prefetch reproduction control means, 16 is a storage means, 17 is a voice output control means,
Is a reproduction control means, 30 is a phoneme analysis means, and 31 is a consonant emphasis means.

The operation will be described below. The basic operation is the same as that of the first embodiment, the voice signal is pre-analyzed before the recorded voice signal is output, and based on the result, the reproduction control means 18 and the consonant emphasizing means 31 to be performed later.
Change the behavior of. The first look-ahead reproduction control means 12 pre-reads voice data prior to the voice data reproduced by the reproduction control means 18 as needed, analyzes the voice data by using the phonological analysis means 30, and outputs the non-voice data. At the same time as deleting and adding a section, information such as vocal speed conversion, consonant emphasis, formant emphasis, etc. of a voice section is obtained. The storage means 16 stores the information analyzed by the phonological analysis means 30. The voice output control means 17 determines the data length for deleting and adding the non-voice section based on the given reproduction condition and the analysis result of the preread voice data stored in the storage means 16, and further determines the voice section. The consonant emphasis parameter is determined according to the type of consonant included in. This determination method can be effectively obtained for a given reproduction condition by previously analyzing the information of the non-voice section of the audio data to be reproduced. As described above, according to the present embodiment, by storing the result of the analysis performed by the phoneme analysis unit for the preread voice data, it is possible to effectively delete and add the non-speech section and enhance the consonant. This makes it possible to easily perform skip reproduction and repeated reproduction in units of voice sections.

[0015]

According to the present invention, the audio signal pre-read by the pre-reading reproduction control means is analyzed in advance by the analysis signal processing means, and the audio data to be reproduced is determined based on the analysis result and the given reproduction condition. An excellent processing effect can be obtained by determining the application method of the signal processing means. By using the speech / non-speech discrimination means as the analysis signal processing means, it is possible to delete and add the non-speech section, and at the same time, obtain useful information for the conversion of the speech rate of the speech section and the formant enhancement. . Further, by using the phonological analysis means as the analysis signal processing means, it is possible to delete or add the non-speech section and at the same time obtain useful information for the conversion of the vocalization rate of the speech section, the consonant emphasis, and the formant emphasis. . With these configurations, it is possible to realize an excellent audio reproduction device that not only naturally reproduces the audio signal of the audio storage medium at a variable speed but also makes it easier to hear the audio section of the consonant part or the vowel part. .

[Brief description of drawings]

FIG. 1 is a configuration diagram of an audio reproducing device according to a first embodiment of the present invention.

FIG. 2 is a configuration diagram of an audio reproducing device according to a second embodiment of the present invention.

FIG. 3 is a configuration diagram of an audio reproducing device according to a third embodiment of the present invention.

FIG. 4 is a configuration diagram of a conventional audio reproducing device.

FIG. 5 is a schematic diagram of an output signal of a conventional audio reproducing device.

[Explanation of symbols]

11 ... Voice storage medium, 12 ... First look-ahead playback control means, 13 ... Second look-ahead playback control means, 14 ... Voice / non-voice discrimination means, 15 ... Speaking speed calculation means, 16 ... Storage means, 17 ... Voice Output control means, 18 ... Regeneration control means,
19 ... Velocity conversion processing means, 20 ... Formant emphasis means,
30 ... Phonological analysis means, 31 ... Consonant emphasizing means, 41 ... Magnetic tape reproducing device, 42 ... Buffer memory, 43 ... Memory control circuit, 44 ... Reproduction control circuit.

Claims (9)

[Claims] 1. A voice storage medium for recording a voice signal, a prefetch playback control means for prefetching and playing back the voice data stored in the voice storage medium, and an output of the prefetch playback control means for analysis or signal processing. Analysis signal processing means, storage means for storing analysis and calculation results of the analysis signal processing means, and audio output control means for determining an audio data reproduction method with reference to a given reproduction condition and a value of the storage means. A reproduction control means for receiving information about data to be read from the audio output control means and reading the audio data; and a predetermined process for the output signal of the reproduction control means under the control of the audio output control means. And a signal processing means for reproducing the time-series audio signal to be reproduced from the result of the analysis of the pre-reading data and subjecting it to predetermined processing. A sound reproducing device characterized by reproducing sound. 2. A voice / non-voice discrimination means and a utterance speed calculation means are provided as analysis signal processing means, and a ratio of deleting and adding a non-voice section by the voice output control means based on given reproduction conditions. The rate of changing the utterance speed in the voice section is determined, the non-voice section is deleted and added by the reproduction control means, the signal processing for changing the utterance speed is performed by the speed conversion processing means, and the voice reproduction speed is changed. The audio reproducing device according to claim 1, wherein 3. A voice / non-voice discriminating means is provided as the analysis signal processing means, and a ratio of deleting and adding a non-voice section by the voice output control means based on a given reproduction condition and a section for performing formant enhancement. Then, the reproduction control means deletes and adds the non-speech section, and the formant emphasizing means performs signal processing for emphasizing a predetermined formant, thereby emphasizing the formant of the sound part while deleting and adding the non-speech section. The audio reproducing device according to claim 1, wherein 4. A phoneme analysis means is provided as an analysis signal processing means, and a ratio of deleting and adding a non-voice section by a voice output control means and a utterance speed of a voice section are changed based on a given reproduction condition. 2. The voice reproduction speed is variable according to claim 1, wherein the ratio is determined, the non-voice section is deleted and added by the reproduction control means, the signal processing for changing the utterance speed is performed by the speed conversion processing means, and the voice reproduction speed is changed. Audio playback device. 5. A phoneme analysis means is provided as an analysis signal processing means, and a ratio of deleting and adding a non-voice section and a section for consonant emphasis are determined by the voice output control means based on a given reproduction condition. The reproduction control means deletes and adds a non-speech section, and the consonant emphasizing section performs signal processing for emphasizing a predetermined consonant so that the consonant portion is emphasized while expanding and contracting the non-speech section. Item 1. The audio reproduction device according to item 1. 6. A phoneme analysis means is provided as an analysis signal processing means, and a ratio of deleting and adding a non-voice section and a section for performing formant enhancement are determined by the voice output control means based on a given reproduction condition. , The reproduction control means deletes and adds non-speech intervals, and the formant emphasizing means performs signal processing for formant emphasis on a predetermined vowel, thereby expanding and contracting the non-speech intervals to emphasize the vowel formants. The audio reproducing device according to claim 1. 7. A phoneme analysis means is provided as an analysis signal processing means, and a ratio of deleting and adding a non-voice section by a voice output control means and a utterance speed of a voice section are changed based on a given reproduction condition. Decide the ratio and the section to emphasize the consonant,
The reproduction control means deletes and adds a non-voice section, the speed conversion processing means performs signal processing for changing the utterance speed, and the consonant emphasis means performs signal processing for emphasizing a predetermined consonant to determine the voice reproduction speed. The audio reproducing apparatus according to claim 1, wherein the consonant part is varied and emphasized. 8. A phoneme analysis means is provided as an analysis signal processing means, and a ratio of deleting and adding a non-voice section by a voice output control means and a utterance speed of a voice section are changed based on a given reproduction condition. The ratio and the section for which formant emphasis is performed are determined, and the non-voice section is deleted or added by the reproduction control means.
The speed conversion processing means performs signal processing for changing the utterance speed, and the formant emphasizing means performs signal processing for formant emphasizing on a predetermined vowel, varying the voice reproduction speed and emphasizing the vowel formant. The audio reproducing device according to claim 1. 9. A phoneme analysis means is provided as an analysis signal processing means, and a ratio of deleting and adding a non-voice section by a voice output control means and a utterance speed of a voice section are changed based on a given reproduction condition. The ratio, the section in which the consonant emphasis is performed and the section in which the formant emphasis is performed are determined, the non-speech section is deleted and added by the reproduction control unit, and the signal processing for changing the utterance speed is performed by the speed conversion processing unit, and the consonant emphasis unit. The signal processing for emphasizing a given consonant is performed with the formant emphasizing means for the formant emphasizing with respect to the given vowel by the formant emphasizing means, and the voice reproduction speed is varied, while the consonant emphasis and the vowel formant are enhanced. The audio reproducing apparatus according to claim 1, wherein the audio reproducing apparatus is emphasized.