JP2001184100A - Speaking speed converting device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reproduce an inputted voice signal which has arbitrary signal time length (source sound recording time) in a desirable reproduction time. SOLUTION: This device is equipped with a voice analysis part 10 which digitally analyzes a continuous voice signal, a sound-recording buffer 11 which stores the digital signal analyzed by the voice analysis part, a source sound recording time detecting means 16 which detects the source sound recording time of the continuous voice signal, a reference speaking speed multiple calculating means 18 which calculates a reference speaking speed multification M represented as the ratio of the detected source sound recording time R0 and a desirable reproduction time R1, and a signal composition part 12 which reproduces a new digital signal with a speaking speed multification Y corresponding to the previously calculated reference speaking speed multification M by compositing the digital signals outputted from the sound-recording buffer 11 responding to reproduction instructions.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a speech speed conversion device for changing the speech speed of voice.

[0002]

2. Description of the Related Art For example, as a method of learning a foreign language, a learning method of actually learning by listening to the foreign language by ear is effective. In this case, it is important to listen to the same conversation repeatedly. Therefore, in order to improve the ability to listen to a foreign language, as a simple method, for example, learning is performed by playing a tape on which a conversation or a narration of several minutes to several tens of minutes has been recorded.

[0003] In this case, a tape recorder dedicated to language learning is configured so that the sound reproduction speed can be varied within a certain range.

[0004] In addition to language learning, speeches, narrations, and news manuscripts are temporarily recorded on a tape recorder,
Even in the case of reproducing later, it is desirable that the reproducing speed can be varied within a certain range.

[0005] However, simply changing the reproduction speed of the sound changes the frequency of the reproduced sound, and the sound sounds unnatural.

[0006] Even if the speaking speed, that is, the speaking speed is changed in order to solve such inconvenience, the frequency of the voice does not change, and the sound can be heard naturally simply by slowing down the speech or making the speaker speak faster. A speech speed conversion method has been proposed.

Next, FIG. 11 and FIG.
2 will be described. FIG. 11 shows, for example, “It's diffi
cult for me to finish ... ". FIG. FIG. 12 is an enlarged view of the audio signal 1. As is well known, voice has consonants and vowels, and voice signal 1 also has corresponding consonants and vowels. As shown, the consonant is composed of one unvoiced sound 2 and the vowel is composed of a plurality of voiced sounds 3. The audio signal 1 includes silence 4 when the audio is interrupted.

Here, the duration of the unvoiced sound 2 or the voiced sound 3 is referred to as a voiced period 5, and the duration of the silence 4 is referred to as a silent period 6.

The unvoiced sound 2 forming a consonant has a relatively high frequency component, and the plurality of voiced sounds 3 forming a vowel have substantially the same waveform. Therefore, in order to increase the speech speed, one or a plurality of voiced sounds 3 among the plurality of voiced sounds 3 constituting the vowel are thinned out, and the voiced sounds 3 before and after the thinned voiced sound 3 or unvoiced sounds 3 are mixed. 2 and voiced sound 3 or voiced sound 3
And silence 4 are connected. Therefore, the duration of the vowel can be shortened, and as a result, the entire time of the audio signal 1 can be shortened,
Speaking speed can be increased without changing the frequency or sound quality of the voice. In addition, by shortening the time of the silent period 6, the speech speed can be increased.

Conversely, when the speech speed is reduced, the same voiced sound 3 may be inserted into a plurality of voiced sounds 3 constituting a vowel to extend the duration of the vowel. Further, the speech speed can be reduced by extending the time of the silent period 6.

In order to automatically perform the speech speed conversion by shortening / extending the voiced period 5, it is necessary to distinguish the unvoiced sound 2, the voiced sound 3 and the voiceless sound 4 included in the voice signal 1. As a classification method, a vowel uses the property that a plurality of voiced sounds 3 continue, and calculates an autocorrelation function for the audio signal 1, thereby classifying the unvoiced sound 2 and the voiced sound 3 and each voiced sound. The duration (pitch) of 3 is detected.

The speech speed of the voice signal 1 is determined by the number of voiced sounds 3 to be thinned out or the number of voiced sounds 3 to be inserted among a plurality of voiced sounds 3 constituting a vowel.

[0013] Therefore, by using a voice reproducing apparatus incorporating such a speech speed conversion function, a language learner can listen to conversation or narration at a high speed or a low speed without a sense of incongruity.

Further, if necessary, the speech speed of a speech, a narration, or a news manuscript can be changed without a sense of discomfort.

[0015]

However, even in a sound reproducing apparatus incorporating the above-mentioned speech speed conversion function,
There were the following issues that still need to be improved.

That is, in lectures and broadcast stations, speeches, narrations, and news manuscripts are often recorded in time with a predetermined time. However, given times often change suddenly.

However, in a voice reproducing apparatus incorporating the above-mentioned voice speed conversion function, what percentage of the voice speed of the voice to be reproduced is raised relative to the normal voice speed, Although it was possible to specify whether to reduce the playback time, the playback time itself could not be specified. For this purpose, the playback speed has been appropriately specified and reproduced. As a result, there arises a problem that the reproduction time does not exactly fit within the given time.

In particular, the broadcasting time in a broadcasting station is strictly set to the second. Speech and narration, even in the uncompressed and uncompressed state, do not fit in seconds at a given time. Therefore, it is extremely difficult to set an appropriate speech rate magnification and strictly store reproduced speeches and narrations to the order of seconds within the broadcast time even though the recording time is uncertain.

The present invention has been made in view of such circumstances, and it is possible to accurately reproduce an audio signal having an arbitrary time length at a given desired reproduction time without giving a sense of incongruity only by giving a desired reproduction time. It is an object of the present invention to provide a speech speed conversion device capable of greatly improving usability.

[0020]

SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, in a speech speed conversion device according to the present invention, a voice analyzing section for digitally analyzing a continuous voice signal, and a voice analyzing section for analyzing the continuous voice signal. A recording buffer for storing a digital signal, an original sound recording time detecting means for detecting an original sound recording time of a continuous audio signal, and a reference speech speed magnification indicated by a ratio of the detected original sound recording time to a desired reproduction time. And a signal for reproducing a new digital signal at a speech rate corresponding to the reference speech rate calculated by synthesizing the digital signal output from the recording buffer in response to the playback instruction. A synthesizing unit.

Further, in the speech speed conversion device of the present invention, a voice analysis unit for digitally analyzing a continuous voice signal,
A recording buffer for storing the digital signal analyzed by the audio analyzing unit, an original sound recording time detecting means for detecting an original sound recording time of a continuous audio signal, and a ratio of the detected original sound recording time to a desired reproduction time. In response to a reproduction instruction, a reference speech speed magnification calculating means for calculating the indicated reference speech speed magnification,
A signal synthesizing unit for synthesizing a digital signal output from the recording buffer to reproduce a new digital signal at a specified speech speed magnification, and calculating a speech speed ratio corresponding to the calculated reference speech speed ratio; A speech speed calculation unit for sending to the synthesis unit.

In the thus constructed speech speed conversion device, the input continuous speech signal is classified (analyzed) by a speech analysis unit into, for example, a sound period, a silence period, an unvoiced sound, a voiced sound, a silence, and the like. You. A digital signal (digital audio signal) digitally analyzed by the audio analysis unit is stored and held in a recording buffer.

Also, the original sound recording time indicated by the signal time length of the input continuous audio signal is detected, and a reference speech speed magnification indicated by, for example, the ratio with the desired input reproduction time is calculated. Then, the speech speed magnification is calculated from the reference speech speed magnification. Here, the speech speed magnification is a magnification of the speech speed when the speech speed of the voice in a state where the speed conversion is not performed is set to 1 (reference).

When a reproduction instruction is issued from the outside, the analyzed digital signal (digital audio signal) stored and held in the recording buffer is converted into a new signal having a speech speed magnification designated by the speech speed calculator in the signal synthesizer. Reproduced as a digital signal (digital audio signal).

Therefore, even if the signal time length (original sound recording time) of the input audio signal is unknown, the operator can specify only the desired reproduction time, and the reproduced audio signal can be changed. Fits exactly in playback time.

According to another aspect of the present invention, there is provided a speech speed conversion device, wherein the speech speed conversion device according to the above-described invention further includes a cumulative sound period calculating means for calculating a cumulative sound period of a continuous audio signal. And a cumulative silent period calculating means for calculating a cumulative silent period of the continuous audio signal. Then, the speech speed calculation unit calculates a sound target speech speed ratio from which the reference speech speed ratio can be obtained by changing the speech speed ratio in the sound period from the calculated accumulated speech period and the accumulated silence period. Then, during the sound period of the digital signal, the speech speed magnification of the sound target speech speed ratio is sent to the signal synthesizing unit, and during the silence period of the digital audio signal, no change in the period is sent to the signal synthesizing unit.

[0027] In the speech speed conversion device thus constructed, as means for changing the speech speed magnification of the continuous voice signal, the silent period of the continuous voice signal is not changed but the time of the voice period is changed. For example, the voiced sound is changed by thinning out or adding the same voiced sound.

Further, in the speech speed conversion device of another invention, the speech speed conversion device of the invention described above further comprises a cumulative sound period calculating means for calculating the cumulative sound period of the continuous audio signal, And a cumulative silence period calculating means for calculating a cumulative silence period of the continuous audio signal. Then, the speech speed calculation unit calculates a silence target period in which the reference speech speed magnification is obtained by changing the silence period from the calculated accumulated speech period and the accumulated silence period, and calculates the speech period of the digital signal. Then, a silent target period is sent to the signal synthesizing unit, and a speech speed magnification of 1 is sent to the signal synthesizing unit during the sound period of the digital signal.

In the speech speed conversion device having the above-described structure, as a means for changing the speech speed magnification of the input voice signal, the voice period of the input voice signal is not changed without changing the voice period. Compressing or expanding duration.

Further, in the speech speed conversion device according to another invention, the speech speed calculation unit of the speech speed conversion device according to the invention described above is configured such that the speech speed calculation unit sets the time within a period from the start time of the continuous audio signal until a predetermined time elapses. As the time elapses, the signal speed is changed from a preset initial speed factor to a target speed factor in the vicinity of the reference speed factor, and after a lapse of a predetermined time, the speed factor for maintaining the target speed factor is sequentially calculated to synthesize a signal. To the department.

In the thus constructed speech speed conversion device, only the beginning portion of the reproduced speech or narration has a speech speed magnification close to the normal speech speed, and after a lapse of a predetermined time, a target speech speed magnification near the reference speech speed magnification. Will be played back. Thus, the beginning of the replayed speech or narration will not be missed.

[0032]

Embodiments of the present invention will be described below with reference to the drawings. (First Embodiment) FIG. 1 is a block diagram showing a schematic configuration of a speech speed conversion device according to a first embodiment of the present invention. A series of continuous audio signals a having the same configuration as the audio signal 1 shown in FIG. Therefore, as shown in FIG. 12, the audio signal a is composed of an unvoiced sound 2 corresponding to a consonant, a voiced sound 3 corresponding to a vowel, and a silent sound 4. Then, as shown in FIG.
Or the duration of the sound period 5 consisting voiced 3 and T _1, the duration of the silent period 6 made silent 4 and T _0.

The analog audio signal a input from the input terminal 7 is converted into a digital audio signal by the A / D converter 8 and then stored in the audio signal memory 9. The voice analysis unit 10 converts the series of digital voice signals a ₁ written in the voice signal memory 9 into the unvoiced sound 2, the voiced sound 3, and the
And is divided into Specifically, the signal level of the audio signal a ₁ is checked, and the sound period 5 and the silent period 6 are classified.
After that, the autocorrelation analysis is performed on the signal of each voiced period 5 to divide the voiced period 5 into the unvoiced sound 2 and the voiced sound 3. The audio signal a ₂ divided into the unvoiced sound 2, the voiced sound 3 and the silent sound 4 by the voice analysis unit 10 is temporarily stored in the recording buffer 1.
1 is written and stored.

When a reproduction start command is input from the outside, the signal synthesizing section 12 takes in the digitally analyzed digital audio signal a ₂ written in the recording buffer 11, and
Of the plurality of voiced sounds 3 constituting each vowel in the voiced period 5 of the captured audio signal a _2, the speech speed calculation unit 1
9 is thinned out or added by the number corresponding to the speech speed magnification Y specified in 9. In addition, the duration of the silence period 6 in the captured audio signal a ₂ is shortened or extended according to the speech speed magnification Y specified by the speech speed calculation unit 19. Then, the signal synthesis unit 12 connects the unvoiced sound ₂ in the input audio signal a _2, the voiced sound 3 after thinning or addition, and the shortened or expanded silence 4 to synthesize a new audio signal a ₃ . And output.

The new audio signal a ₃ output from the signal synthesizing unit 12 is temporarily stored in the output buffer 13 and then stored in the D / A
The signal is converted into an analog audio signal a ₄ by the converter 14 and output from the output terminal 15.

Therefore, the new analog audio signal a ₄ output from the output terminal 15 is shortened or reduced by the amount corresponding to the specified speech speed magnification Y with respect to the analog audio signal a input to the input terminal 7. The length of the expanded speech or narration is shortened or lengthened accordingly.

The digital audio signal a ₁ output from the A / D converter 8 is written into the audio signal memory 9 and
It is input to the recording time detection unit 16. Recording time detector 16
Is the input audio signal a (a ₁ ), as shown in FIG.
The original sound recording time R ₀ indicated by the signal time length of
It is sent to the next reference speech speed magnification calculator 18.

The desired reproduction time input unit 17 the operator has operated the input, shown in FIG. 2, the following criteria story playback time R _I is the signal duration of the audio signal a ₄ analog output from the output terminal 15 It is sent to the speed magnification calculator 18.

The reference speech speed magnification calculator 18 calculates a reference speech speed ratio M by dividing the original sound recording time R ₀ of the input audio signal a (a ₁ ) by the desired reproduction time R _I of the output audio signal a _4. Is calculated and sent to the next speech speed calculation unit 19.

M = R ₀ / R _I The speech speed calculator 19 of the speech speed converter of the first embodiment is
The input reference speech speed magnification M is directly transmitted to the signal synthesizing unit 12 as the speech speed magnification Y.

Y = M As described above, the signal synthesizing unit 12 compresses and expands the captured digital audio signal a ₂ and outputs it as a digital audio signal a ₃ having a speech speed magnification Y.

In the speech speed conversion device of the first embodiment thus configured, the original sound recording time R ₀ indicating the signal time length of the audio signal a inputted from the input terminal 7 is automatically measured. Then, when the operator inputs the desired reproduction time R ₁ , the reference speech speed magnification M is automatically calculated and applied to the signal synthesizing unit 12 as the speech speed magnification Y.

[0043] Thus, as the operator, regardless of the original sound recording time R ₀ of the input speech signal a, only the operation input the desired playback time R _1, audio signal a having a desired playback time R ₁ of high precision ₄ can be played.

[0044] Therefore, when using this speech speed converting device in an environment controlled in seconds airtime as broadcasters, since high accuracy desired playback time R ₁ is ensured, the speech speed conversion The usability of the device can be greatly improved.

(Second Embodiment) FIG. 3 is a block diagram showing a schematic configuration of a speech speed conversion device according to a second embodiment of the present invention. The same parts as those of the speech speed conversion device of the first embodiment shown in FIG. 1 are denoted by the same reference numerals, and detailed description of the overlapping parts will be omitted.

In the speech speed conversion device of the second embodiment, the speech output from the speech analyzer 10 is divided (analyzed) into a voiced period 5, a silent period 6, an unvoiced sound 2, a voiced sound 3, and a silence 4. The digital audio signal a ₂ is sequentially written into the recording buffer 11 and is input to the cumulative sound period calculating unit 20 and the cumulative silent period calculating unit 21.

The cumulative sound period calculation unit 20 is configured as shown in FIG.
Thus, the input one original sound recording time R ₀Minute audio signal a
_TwoAll sound periods T included in (a)₁Has accumulated
Sound period T_S1Is calculated and sent to the speech speed calculation unit 19a. same
As shown in FIG. 2, the cumulative silence period calculation unit 21
Input original sound recording time R₀Minute audio signal a_Two(A)
All included silent periods T₀Cumulative silence period T
_S0Is calculated and sent to the speech speed calculation unit 19a.

Further, when a reproduction instruction is input and the signal synthesizing unit 12 starts reading out the digital audio signal a ₂ stored in the recording buffer 11, the audio signal a ₂ is input from the recording buffer 11 to the signal synthesizing unit 12. At the same time, it is input to the speech speed calculation unit 19a. Therefore, the speech speed calculation unit 1
The 9a, the audio signal a ₂ is input before the start of the digital, the reference speech rate magnification M, the cumulative voiced period T _S1, accumulated silence period T _S0
Is entered.

The speech speed calculator 19a calculates the speech speed magnification Y in accordance with the flowchart shown in FIG.
2 is executed.

As described above, the reference speech speed magnification M, the accumulated sound period T _S1 , and the accumulated silence period T _S0 are taken (P1). Then, to calculate the sound target speech speed ratio N ₁ (P2). This voiced target speech speed magnification N ₁ is the same as each silent period T _S0 ,
The speech speed magnification Y is applied to each sound period T _S1 in order to compress / expand each sound period T _S1 to obtain the desired reproduction time R ₁ . The voiced target speech speed ratio N ₁ is derived from the following formula.

(T _S1 / N ₁ ) + T _S0 = (T _S1 + T _S0 ) / M N ₁ = (M · T _S1 ) / (T _S1 + T _S0 -M · T _S0 ) and the digital audio signal a ₂ is initiated input (P3), for example, 0.
After waiting for the minute time Δt of 01 seconds, etc. (P4), to confirm that the audio signal a ₂ of the digital has not been completed (P5), the digital signal state of the audio signal a ₂ of the present time Yes In the case of the sound period 5 (T ₁ ) (P 6), the sound target speech speed magnification N ₁ obtained above is transmitted to the signal synthesizing unit 12 as the speech speed magnification Y (P 7). Y = N ₁ The signal state of the digital audio signal a _{2 at} the current time is a silent period 6
In the case of (T ₀ ), (P 6), the speech speed magnification 1 is transmitted to the signal synthesizing unit 12 as the speech speed magnification Y, which is an instruction for no change in the silence period 6 (T ₀ ) (P 8). Y = 1 Then, at P5, the audio signal a ₂ digital is completed, ends the speech speed ratio calculation process.

In the speech speed converter of the second embodiment configured as described above, similarly to the speech speed converter of the first embodiment described above, when the desired reproduction time R ₁ is designated,
An audio signal a ₄ having the desired playback time R ₁ can child play.

Further, in the speech speed conversion device of the second embodiment, the speech speed magnification Y input from the speech speed calculation unit 19a to the signal synthesis unit 12 is set to 1 during the silent period 6 (T ₀ ). That is, the soundless period 6 (T ₀ ) is set without change, and the sound target speech speed magnification N ₁ is set in the sound period 5 (T ₁ ). Therefore, even if the input audio signal a is greatly shortened or expanded, the silence period 6 between words is secured without being changed, so that it sounds more natural.

(Third Embodiment) FIG. 5 is a flowchart showing a process of calculating a speech speed magnification Y of a speech speed calculating unit 19a incorporated in a speech speed conversion device according to a third embodiment of the present invention. The block diagram of the speech speed converter of the third embodiment is the same as the block diagram of the speech speed converter of the second embodiment shown in FIG. The only difference is the content of the processing for calculating the speech speed magnification Y by the speech speed calculation unit 19a.

Then, the speech speed calculator 19a of the third embodiment
Performs a process of calculating the speech speed magnification Y and sending it to the signal synthesizing unit 12 according to the flowchart shown in FIG.

As in FIG. 4, the reference speech speed magnification M, the accumulated speech period T _S1 , and the accumulated silence period T _S0 are taken in (Q1), and the target silence target speech speed N ₀ for obtaining the target silence period is calculated. It is calculated (Q2). Silence target speech speed magnification N ₀ for obtaining the silence target period, each voice period T _{S 1} is intact, in order to obtain a desired playback R ₁ and compression and expansion of each silent period T _S0,
This is the speech speed magnification Y to act on each silent period T _S0 . This silent target speech speed magnification N ₀ is derived from the following equation.

T _S1 + (T _S0 / N ₀ ) = (T _S1 + T _S0 ) / M N ₀ = (M · T _S0 ) / (T _S1 + T _S0 −M · T _S1 ) and output from the recording buffer 11. It has been the digital audio signal a ₂ is initiated input (Q3), for example, 0.
After waiting for the minute time Δt of 01 seconds, etc. (Q4), and confirm that the audio signal a ₂ of the digital has not been completed (Q5), the digital signal state of the audio signal a ₂ of the present time silence In the case of the period 6 (T ₀ ) (Q 6), the silence target speech speed magnification N ₀ obtained above is transmitted to the signal synthesizing unit 12 as the speech speed magnification Y (Q 7). Y = N ₀ digital signal state of the audio signal a ₂ of the present time is in the case of voiced period _{5 (T 1) (Q6)} , and sends to the signal combining unit 12 speech speed ratio 1 as speech speed ratio Y ( Q8). Y = 1 Then, at Q5, when the audio signal a ₂ digital is completed, ends the speech speed ratio calculation process.

In the third speech speed converter constructed as described above, when the desired playback time R ₁ is designated, similarly to the speech speed converter of the first embodiment described above, the desired playback time R _{1 is set.} can child playing audio signal a ₄ with.

Further, in the speech speed conversion device according to the third embodiment, the speech speed magnification Y input from the speech speed calculation unit 19a to the signal synthesis unit 12 is set to 1 in the sound period 5 (T ₁ ). Then, in the silence period 6 (T ₀ ), the silence target speech speed magnification N ₀ is set.

Therefore, for example, the silent period 6 (T ₀ )
There are many, in the case of shortening the speech and narration that was slow to the desired playback time R _1, by adopting this approach, it is possible to be that tight to hear easy playback voice.

(Fourth Embodiment) FIG. 6 shows the time characteristic of the speech speed magnification Y applied from the speech speed calculator 19a incorporated in the speech speed converter of the fourth embodiment of the present invention to the signal synthesizer 12. FIG. The block diagram of the speech speed converter of the fourth embodiment is the same as the block diagram of the speech speed converter of the second embodiment shown in FIG. The only difference is the content of the processing for calculating the speech speed magnification Y by the speech speed calculation unit 19a.

[0062] The speech speed calculation unit 19a, as shown in FIG. 6, until the specified time T _B has elapsed from the start time of the digital audio signal a ₃ which is input to the speech rate calculating section 19a (t = 0) Is changed from an initial speech speed magnification B (= 1) set in advance with time to a target speech speed N near the reference speech speed M, and after a lapse of a specified time T, the target speech speed is increased. The speech speed magnification Y that maintains the magnification N is sequentially calculated and sent to the signal synthesis unit 12.

More specifically, the speech speed calculation unit 19a performs a process of calculating the speech speed magnification Y and sending it to the signal synthesis unit 12 in accordance with the flowchart shown in FIG.

First, the reference speech speed magnification M, the specified time T _B , and the original sound recording time R ₀ are taken in (S1), and the target speech speed magnification N is calculated (S2). Specifically, as shown in FIG. 6, the area S ₁ where the voice speed magnification Y is lower than the reference voice speed M is equal to the area S ₂ where the voice speed Y exceeds the reference voice speed M. The target speech speed magnification N is calculated.

S ₁ = (M−1) T _B / 2 S ₂ = (N−M) (R ₀ −T _B ) N = [(M−1) T _{B /} 2 (R ₀ −T _B )] + M When the input of the digital audio signal a ₂ output from the recording buffer 11 is started, the elapsed time t
Is initialized (t = 0). For example, after elapse of a minute time Δt such as 0.01 second (S4), the elapsed time t is determined in S5.
Is updated (t = t + Δt). When the elapsed time t after the update is less than the predetermined time T _B (S6), calculates the speech speed ratio Y shown in the following equation (S7). Y = [(N−1) / T _B ] t + 1 The calculated speech speed magnification Y is sent to the signal synthesis unit 12 (S
8). Then, the process returns to S4 and waits for the elapse of the next minute time Δt.

At S 6, when the elapsed time t after the update reaches the specified time T _B , the target speech speed magnification Y = N is sent to the signal synthesizing section 9.

In the speech speed converter according to the fourth embodiment thus constructed, as shown in FIG. 6, a digital audio signal a ₂ analyzed from the recording buffer 11 at time (elapsed time) t = 0. Is the signal synthesis unit 12 and the speech speed calculation unit 19a
If the input starts to speech speed of the speech signal a ₄ output from the output terminal 15 is typically speech speed (Y = 1). And
As the elapsed time t increases, the speech speed magnification Y also increases.

[0068] When the elapsed time t reaches the predetermined time T _B, it reaches the target speech speed ratio N over speech speed ratio Y slightly reference speech speed magnification M is (Y = N). After a lapse of specified time T _B is speaking rate ratio Y maintains the target speech speed magnification N.

Therefore, by adopting the speech speed conversion device of the fourth embodiment, when the desired playback time R ₁ is designated, as in the speech speed conversion device of the first embodiment described above, the desired playback time is designated. can child playing audio signal a ₄ having a time R _1. Further, only the beginning of the speech or the narration has a speech speed close to normal, and thereafter the speech speed becomes the target speech speed N. Therefore, you will not miss the beginning of your speech or narration.

(Fifth Embodiment) FIG. 8 shows the time characteristic of the speech speed magnification Y applied from the speech speed calculator 19a incorporated in the speech speed converter of the fifth embodiment of the present invention to the signal synthesizer 12. FIG. The block diagram of the speech speed conversion device of the fifth embodiment is the same as the block diagram of the speech speed conversion device of the second embodiment shown in FIG. The only difference is the content of the processing for calculating the speech speed magnification Y by the speech speed calculation unit 19a.

As shown in FIG. 8, the speech speed calculating unit 19a starts the sounding period 5 (T ₁ ) of the digital voice signal a ₃ input to the speech speed calculating unit 19a (t _S = T _S ). in age period t to _S defines time T _B has elapsed from 0), is set in advance with time the initial speech speed ratio B
(= 1) is changed to the target speech speed magnification N of the reference speech rate ratio M vicinity of and after the predetermined time T _B has elapsed, the target speech speed magnification N
Further, when the digital audio signal a ₃ changes to the silent period 6 (T ₀ ) during the elapsed time t _S = t _E, it becomes 1, and maintains ₁ until the next voiced period 5 (T ₁ ). The speech speed magnification Y is sequentially calculated and sent to the signal synthesizing unit 12. However, when the duration t _Q of the silent period 6 is shorter than the threshold time T _SH , the speech speed magnification Y does not return to 1.

More specifically, the speech speed calculation unit 19a performs a process of calculating the speech speed magnification Y and sending it to the signal synthesis unit 12 in accordance with the flowchart shown in FIG.

First, the reference speech speed M, the specified time T _B , and the speech speed B at the start of speech are taken in (A1), and the target speech speed N is obtained.
Is calculated (A2). Specifically, as shown in FIG.
The target speech speed magnification N is calculated so that an area S ₁ where the speech speed magnification Y is lower than the reference speech speed magnification M is equal to an area S ₂ where the speech speed magnification Y is higher than the reference speech speed magnification M. Note that the sound period 5
Since the time length T ₁ (t _S = 0 to t _S = t _E ) varies depending on the word or phrase, the target speech speed magnification N is calculated using an average value.

When the input of the digital audio signal a ₂ output from the recording buffer 11 is started, for example,
After a lapse of a minute time Δt such as 0.01 second (A3),
Taking the analysis results at the time of the digital audio signal a ₂ by A4 (A4). Then, if the digital audio signal a ₂ has the sound period 5 (A5), the state of the sound period flag is checked (A6). If it remains released to 0, it is determined that the sound period 5 has entered for the first time this time, and the sound period flag is set to 1 at A7, and the sound elapsed time t _S is initialized (t _S = 0). Further, the silent crotch flag is set to 0.
(A8).

At A6, the sound period flag is already 1
If it is still set to, it is determined that the sound period 5 has entered before the previous time, and the sound elapsed time t _S is updated at A9 (t _S = t _S + Δt).

Then, the speech speed magnification Y represented by the following equation is calculated (A10).

Y = N− (NB) exp [−t _S / (T _B / 5)] The calculated speech speed magnification Y is sent to the signal synthesis unit 12 (A 11). Then, the process returns to A3 and waits for the next minute time Δt to elapse.

[0078] Also, at A5, in the case the digital audio signal a ₂ at the current point of silent period 6 not voiced period 5, determine the status of the silent period flag (A12). If it is kept at 0, it is determined that the silence period 6 has entered for the first time this time. At A13, the silence period flag is set to 1 and the silence elapsed time t _Q is initialized (t _Q =
0). Then, the process proceeds to A9, and the sound elapsed time t _S is updated (t _S = t _S + Δt). Further, the above-mentioned speech speed magnification Y is calculated (A10) and transmitted to the signal synthesizing unit 12 (A1).
1). Then, the process returns to A3 and waits for the next minute time Δt to elapse.

At A12, if the silent period flag is already set to 1, it is determined that the silent period 6 has been entered before the previous time, and the silent elapsed time t _Q is updated at A14 (t _Q = t _Q + Δt). Then, the silence elapsed time t after the update
It is checked whether or not _Q has exceeded a preset threshold value T _SH (A15).

If the silence elapsed time t _Q does not exceed the threshold time T _SH , the process proceeds to A 9, where the speech elapsed time t _S is updated (t _S = t _S + Δt). Further, the speech speed magnification Y is calculated (A10) and transmitted to the signal synthesizing unit 12 (A11).
Then, the process returns to A3 and waits for the next minute time Δt to elapse.

At A15, when the silence elapsed time t _Q exceeds the threshold time T _SH , the sound period flag is reset to 0 for the first time (A16). Then, returning to A3, the next minute time Δ
Wait for t to elapse.

In the speech speed converter according to the fifth embodiment thus configured, as shown in FIG. 9, the analyzed digital audio signal a ₂ is recorded from the recording buffer 11 by the signal synthesizer 12 and the speech speed calculator. If the input starts to section 19a, speech speed ratio Y of the audio signal a ₄ output from the output terminal 15,
Each time the sound period 5 (T ₁₎ is started, an increase from the start speech speed ratio B talk close to normal speech speed, a specified time T in _B, until the target speech speed magnification N of the reference speech rate ratio M vicinity As long as the corresponding sound period 5 (T ₁ ) continues, the target speech speed magnification N is maintained.

Therefore, when the desired reproduction time R ₁ is designated by adopting the speech speed conversion device of the fifth embodiment, similarly to the speech speed conversion device of the first embodiment described above, the desired reproduction time is designated. can child playing audio signal a ₄ having a time R _1.

Further, when the speech in the speech or narration exceeds a certain threshold time T _SH , the speech speed magnification Y returns to almost the normal value B, so that the next speech starts at the normal speech speed. , Very easy to hear.

[0085]

As described above, in the speech speed converter according to the present invention, an audio signal having an arbitrary time length can be accurately reproduced at a given desired reproduction time without giving a sense of incongruity only by giving a desired reproduction time. it can. Therefore, usability can be greatly improved.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a schematic configuration of a speech speed conversion device according to a first embodiment of the present invention;

FIG. 2 is a schematic diagram showing a signal configuration of an audio signal to be recorded and reproduced;

FIG. 3 is a block diagram showing a schematic configuration of a speech speed conversion device according to a second embodiment of the present invention.

FIG. 4 is a flowchart showing a speech speed multiplication calculation process of a speech speed calculation unit incorporated in the speech speed conversion device according to the second embodiment;

FIG. 5 is a flowchart showing a process of calculating a speech speed magnification of a speech speed calculation unit incorporated in a speech speed conversion device according to a third embodiment of the present invention.

FIG. 6 is a diagram showing a change in a speech speed magnification output from a speech speed calculation unit incorporated in a speech speed conversion device according to a fourth embodiment of the present invention.

FIG. 7 is a flowchart showing a speech speed multiplication calculation process of a speech speed calculation unit incorporated in the speech speed conversion device according to the fourth embodiment;

FIG. 8 is a diagram showing a change in a speech speed magnification output from a speech speed calculation unit incorporated in a speech speed conversion device according to a fifth embodiment of the present invention.

FIG. 9 is a diagram showing a change in a speech speed magnification output from a speech speed calculator incorporated in the speech speed converter according to the fifth embodiment.

FIG. 10 is a flowchart showing a speech speed magnification calculation process of a speech speed calculation unit incorporated in the speech speed conversion device according to the fifth embodiment.

FIG. 11 is a diagram showing a general audio signal waveform.

FIG. 12 is a diagram showing details of a general audio signal.

[Explanation of symbols]

 2 ... unvoiced sound 3 ... voiced sound 4 ... silence 5 ... voiced period 5 ... silence period 8 ... A / D converter 9 ... audio signal memory 10 ... audio analysis unit 11 ... recording buffer 12 ... signal synthesis unit 13 ... output buffer 14 ... D / A converter 16 ... Recording time detection unit 17 ... Requested reproduction time input unit 18 ... Reference speech speed magnification calculation unit 19,18a ... Speech speed calculation unit 20 ... Cumulative sound period calculation unit 21 ... Cumulative silence period calculation unit

[Procedure amendment]

[Submission date] January 7, 2000 (2000.1.7)

[Procedure amendment 1]

[Document name to be amended] Drawing

[Correction target item name] All figures

[Correction method] Change

[Correction contents]

FIG.

FIG. 2

FIG. 4

FIG. 6

FIG. 9

FIG. 3

FIG. 5

FIG. 7

FIG. 11

FIG. 8

FIG. 10

FIG.

 ────────────────────────────────────────────────── ─── Continuing on the front page (72) Inventor Yoichi Katsuki 5-10-27 Minamiazabu, Minato-ku, Tokyo Anritsu Corporation F-term (reference) 5D045 AA07 BA02 5D080 AA05 BA01 DA02 FA31 FA39 GA02 GA16

Claims (5)

[Claims] An audio analysis unit for digitally analyzing a continuous audio signal; a recording buffer for storing a digital signal analyzed by the audio analysis unit; an original sound of the continuous audio signal; the original sound recording time detecting means for detecting the recording time (16), and calculates the detected original sound recording time (R ₀₎ and the desired playback time reference speech speed ratio represented by the ratio of (R ₁₎ (M) A reference voice speed magnification calculating means (18); and a voice speed ratio (Y) corresponding to the calculated reference voice speed ratio (M) by synthesizing a digital signal output from the recording buffer in response to a reproduction instruction. And a signal synthesizing section (12) for reproducing a new digital signal. 2. An audio analysis unit (10) for digitally analyzing a continuous audio signal, a recording buffer (11) for storing a digital signal analyzed by the audio analysis unit, and an original sound of the continuous audio signal. the original sound recording time detecting means for detecting the recording time (16), and calculates the detected original sound recording time (R ₀₎ and the desired playback time reference speech speed ratio represented by the ratio of (R ₁₎ (M) Reference speech speed magnification calculating means (18), and signal synthesis for reproducing a new digital signal at a specified speech speed magnification (Y) by synthesizing digital signals output from the recording buffer in response to a reproduction instruction. Unit (12); and a speech speed calculation unit (19) that calculates a speech speed ratio (Y) corresponding to the calculated reference speech speed ratio (M) and sends the result to the signal synthesis unit. Characteristic speech speed converter. 3. A cumulative sound period calculating means (20) for calculating a cumulative sound period of the continuous audio signal, and a cumulative silent period calculating means (21) for calculating a cumulative silent period of the continuous audio signal. The speech speed calculation unit (19a) is configured to change the speech speed ratio of the speech period from the calculated accumulated speech period and the accumulated silence period so that the reference speech speed ratio (M) is Calculate the obtained sound target speech speed magnification (N ₁ ) and send the speech speed magnification of the sound target speech speed magnification to the signal synthesizing section during the sound period of the digital signal. 3. The speech speed conversion device according to claim 2, wherein no change in period is sent to the signal synthesis unit during a period. 4. A cumulative sound period calculating means (20) for calculating a cumulative sound period of the continuous audio signal, and a cumulative silent period calculating means (21) for calculating a cumulative silent period of the continuous audio signal. The speech speed calculation unit (19a) includes: a silence target period in which the reference speech speed magnification (M) is obtained by changing the silence period from the calculated accumulated speech period and accumulated silence period. (M ₀ ) is calculated, and the speech rate magnification of the target silence period (N ₀ ) is sent to the signal synthesizing section during the sound period of the digital signal, and the speech speed of 1 is output during the sound period of the digital signal. 3. The speech speed conversion device according to claim 2, wherein a magnification is transmitted to the signal synthesis unit. 5. The speech speed calculation unit (19a), based on an initial speech speed magnification set in advance as time elapses, within a period from the start time of the continuous audio signal until a specified time elapses, sets the reference value. The method according to claim 1, further comprising: changing a speech speed ratio to a target speech speed ratio in the vicinity of the speech speed ratio, and after the lapse of the prescribed time, sequentially calculating a speech speed ratio for maintaining the target speech speed ratio, and transmitting the calculated speech speed ratio to the signal synthesis unit. 2. The speech speed converter according to 2.