JP6181921B2 - Voice reproduction apparatus, voice synthesis reproduction apparatus, and programs thereof - Google Patents

Description

  The present invention relates to a sound reproducing device and a program for reproducing a sound by performing signal processing, and a sound synthesizing and reproducing device and a program for reproducing the synthesized sound by signal processing.

  2. Description of the Related Art Conventionally, a voice reproduction method has been proposed in which the voice speed is changed in real time by converting the speech speed of input voice (see Patent Document 1). The speech reproduction method proposed in Patent Document 1 is a method in which speech uttered by a person is processed and reproduced while converting the speech speed in real time, and the input speech data is converted when converting the speech speed of the received speech. By constantly monitoring the length, the output data length calculated in advance by a conversion function relating to the scaling factor given in advance, and the data length of the voice that is actually output in a certain processing unit, missing information A series of processes can be performed without causing them.

Japanese Patent No. 3220043 (Japanese Patent Laid-Open No. 10-301598)

  However, although the speech reproduction method proposed in Patent Document 1 can convert the speech speed in real time with respect to the input speech and reproduce the speech after the speech speed conversion, it takes the speech speed conversion of the input speech. If the time is longer than the time for reproducing the converted sound, the reproduced sound may be interrupted. For example, in the audio reproduction method proposed in Patent Document 1, the voice A is converted when the voice A to be reproduced next to the voice A is concurrently converted during the reproduction of the voice A after the voice speed conversion. If the playback time is 5 seconds and the speech speed conversion time of the voice B is 10 seconds, nothing is played back for 5 seconds after the playback of the voice A is finished.

  The present invention has been made in view of the above points, and provides a highly stable voice reproducing apparatus and voice synthesizing / reproducing apparatus that can reproduce signal-processed voice without interruption, and a program thereof. Let it be an issue.

  In order to solve the above problem, an audio reproducing apparatus according to claim 1 is an audio reproducing apparatus that performs audio signal processing and reproduces the input audio buffering means, audio signal processing means, and output audio buffering means. And an audio reproduction means, an output audio buffer detection means, and an audio signal processing control means.

  An audio reproducing apparatus having such a configuration stores input audio by the input audio buffering means, and the input audio buffer when the control signal is input from the audio signal processing control means by the audio signal processing means. The sound stored in the ring means is signal-processed. That is, the audio signal processing means processes the audio when the amount of the signal-processed audio stored in the output audio buffering means is below the threshold value. As a result, the amount of the sound stored in the output sound buffering means, that is, the amount of the signal processed sound reproduced by the sound reproducing means is always kept constant.

  In addition, the audio reproduction device stores the audio signal-processed by the audio signal processing unit by the output audio buffering unit, and reproduces the audio stored in the output audio buffering unit by the audio reproduction unit. Then, the sound reproducing device detects the amount of sound stored in the output sound buffering means by the output sound buffer detection means, and the amount of sound detected by the sound signal processing control means in the output sound buffer detection means. Is less than a predetermined threshold value, and if it is less than the threshold value, a control signal for signal processing the sound stored in the input sound buffering means is sent to the sound signal processing means. Output.

Further, in the audio reproduction apparatus, the audio signal processing means, and signal processing the test sound predetermined by the output audio buffering means, to save the test speech signal processing in the audio signal processing means, audio reproduction means When the test audio stored in the output audio buffering means is played back and the amount of test audio stored in the output audio buffering means is detected by the output audio buffer detection means, the audio signal processing control means The threshold used in the above is obtained by multiplying the minimum value of the amount of test audio detected by the output audio buffer detection means by a predetermined reliability factor indicating the stability of the processing performance of the audio reproduction device. . This reliability rate is, for example, “1.0” when the CPU of the sound reproducing apparatus is embodied by a computer, for example, when there are few high-performance CPUs or programs operating in parallel, and the CPU with poor performance. When a plurality of programs are operating in parallel, for example, “2.0” or “3.0” is set by the user in advance. That is, the audio reproduction device can use the value obtained from the output of the output audio buffer detection means obtained by processing by inputting the test audio as the threshold value in the audio signal processing control means.

According to a third aspect of the present invention, there is provided an audio reproduction device according to the first aspect, wherein the audio signal processing means performs signal processing for converting the speech speed of the audio stored in the input audio buffering means. . That is, the voice signal processing means converts the voice speed when the amount of the voice speed converted voice stored in the output voice buffering means falls below a threshold value. As a result, the amount of speech stored in the output speech buffering means, that is, the speech speed-converted speech reproduced by the speech reproduction means is always kept constant.

  In order to solve the above-mentioned problem, a speech synthesis / playback device according to claim 4 is a speech synthesis / playback device including the speech playback device according to any one of claims 1 to 3, comprising: speech synthesis means; The input voice buffer detection means and the voice synthesis control means are provided.

  A speech synthesis / playback apparatus having such a configuration synthesizes speech corresponding to an input sentence when a control signal is input from speech synthesis control means by speech synthesis means, and speech synthesis is performed by input speech buffering means. The synthesized voice in the means is saved. That is, the speech synthesis unit synthesizes speech when the amount of synthesized speech stored in the input speech buffering unit is below the threshold value. As a result, the amount of synthesized speech stored in the input speech buffering means, that is, the amount of synthesized speech signal-processed by the speech signal processing means is always kept constant.

  Then, the speech synthesis / playback apparatus detects the amount of speech stored in the input speech buffering means by the input speech buffer detection means, and the amount of speech detected by the speech synthesis control means in the input speech buffer detection means. Is less than a predetermined threshold, and if it is less than the threshold, a control signal for synthesizing the speech corresponding to the input sentence is output to the speech synthesizer.

  In order to solve the above problem, an audio reproduction program according to claim 5 causes a computer to function as the audio reproduction apparatus according to any one of claims 1 to 3.

  In order to solve the above problems, a speech synthesis / playback program according to a sixth aspect causes a computer to function as the speech synthesis / playback apparatus according to the fourth aspect.

  According to the first, second, and fifth aspects of the invention, the signal-processed sound is interrupted in order to process and reproduce the sound while always managing the amount of the signal-processed sound. Can be reproduced stably.

According to the second and third aspects of the invention, since the voice is converted and played back while always managing the amount of the voice whose voice speed has been converted, the voice whose voice speed has been converted is not interrupted. It can be played back stably.

  According to the inventions according to claims 4 and 6, the synthesized speech is synthesized while always managing the amount of synthesized speech, and the synthesized speech is signal-processed and reproduced while always managing the amount of synthesized speech that has undergone signal processing. Therefore, the signal-processed synthesized speech can be stably reproduced without interruption.

It is a block diagram which shows the whole structure of the audio | voice reproduction apparatus which concerns on this invention. It is a flowchart which shows the process sequence of the audio | voice reproduction apparatus which concerns on this invention. 1 is a block diagram showing an overall configuration of a speech synthesis / playback apparatus according to the present invention. It is a flowchart which shows the process sequence of the speech synthesizer / reproducing apparatus concerning this invention.

  A voice playback device, a voice synthesis playback device, and a program thereof according to an embodiment of the present invention will be described with reference to the drawings. In the following description, the same configuration is given the same name and symbol, and detailed description is omitted. Further, in the following description, the voice playback device will be described first, and then the voice synthesis playback device and these programs will be described.

[Configuration of audio playback device]
The configuration of the audio reproducing apparatus according to the present invention will be described with reference to FIG. The audio reproduction device 1 performs signal processing on input audio and reproduces it. Specifically, as shown in FIG. 1, predetermined audio signal processing parameters are set for audio input from the outside. Based on the above, signal processing is performed to reproduce signal-processed sound (hereinafter referred to as signal-processed sound). The audio reproduction device 1 is used when, for example, a viewer who watches a television program in a general home adjusts the speech speed or the voice pitch in real time by operating a remote controller or the like.

  Here, as shown in FIG. 1, the audio reproducing apparatus 1 includes an input audio buffering means 11, an audio signal processing means 12, an output audio buffering means 13, an audio reproducing means 14, and an output audio buffer detecting means 15. And an audio signal processing control means 16.

  The input audio buffering means 11 stores audio before signal processing. As shown in FIG. 1, the input audio buffering means 11 stores externally input audio before signal processing and outputs the audio in response to a request from the audio signal processing means 12 described later. The input audio buffering means 11 is specifically composed of a hard disk or flash memory capable of storing data. Here, as shown in FIG. 1, the input audio buffering means 11 is provided inside the audio reproduction device 1, but may be configured outside the audio reproduction device 1.

  The sound signal processing means 12 performs signal processing on sound. As shown in FIG. 1, the audio signal processing unit 12 processes the audio stored in the input audio buffering unit 11 based on a predetermined audio signal processing parameter. Then, the audio signal processing unit 12 outputs the signal-processed audio to the output audio buffering unit 13 as shown in FIG. Specific examples of signal processing in the audio signal processing means 12 include, for example, speech speed conversion processing, voice quality conversion processing, voice pitch changing processing, and voice emotion component changing processing. .

  When the speech signal processing means 12 performs speech speech speed conversion processing as signal processing, a speech expansion / contraction rate is input from the outside as a speech signal processing parameter. In response to this, the voice signal processing means 12 detects the voice section using the power of the voice data, the number of zero crossings, and the autocorrelation function, and extracts the pitch period for the voice section. Then, the audio signal processing means 12 performs decimation / repetition of the basic period of the audio waveform based on the time length defined by the pitch period and the expansion / contraction rate, and the audio waveforms are overlapped and connected with an appropriate time length. By doing so, speech speed conversion is performed. In addition, about such a speech speed conversion method, a well-known technique can be used (for example, patent 3327936, patent 2955247).

  In addition, when the voice signal processing means 12 performs a process of changing the emotion component of the voice as the signal processing, for example, an emotion parameter corresponding to emotions is input as a voice signal processing parameter from the outside. And the audio | voice signal processing means 12 changes the emotion component of an audio | voice by adjusting the fundamental frequency and spectrum of an audio | voice waveform based on the said emotion parameter.

  Here, the audio signal processing unit 12 does not simply process the audio stored in the input audio buffering unit 11, but performs signal processing according to the amount of audio that has already been subjected to signal processing. That is, as shown in FIG. 1, the audio signal processing unit 12 performs input audio buffering when the amount of signal processed audio already stored in the output audio buffering unit 13 is less than a predetermined threshold. The voice stored in the means 11 is signal-processed.

  More specifically, as shown in FIG. 1, the audio signal processing means 12 includes an audio signal processing control means when the signal processed audio stored in the output audio buffering means 13 is less than the above-mentioned threshold value. 16, a control signal for performing signal processing is input. Then, only when this control signal is input, the audio signal processing means 12 obtains the required number of audios stored in the input audio buffering means 11 in order from the oldest to obtain a signal such as speech speed conversion. Processing is performed, and the signal-processed sound is output to the output sound buffering means 13. Here, the audio signal processing means 12 performs audio signal processing in parallel with the reproduction of the signal processed audio by the audio reproduction means 14 described later.

  The output sound buffering means 13 stores the signal processed sound. As shown in FIG. 1, the output audio buffering means 13 stores the signal processed audio input from the audio signal processing means 12 and also performs the signal processed audio in response to a request from the audio reproduction means 14 to be described later. Is output. The output audio buffering means 13 is specifically composed of a hard disk or flash memory capable of storing data. Here, the output audio buffering means 13 is provided inside the audio reproduction device 1 as shown in FIG. 1, but may be configured outside the audio reproduction device 1.

  The audio reproduction means 14 reproduces the signal processed audio. As shown in FIG. 1, the audio reproduction means 14 obtains the required number of signal processed voices stored in the output audio buffering means 13 in order from the oldest one, and with respect to an audio device (not shown) such as a speaker. Output and play. Here, the sound reproduction means 14 reproduces the signal processed sound in parallel with the signal processing by the sound signal processing means 12 described above.

  The output sound buffer detecting means 15 detects the amount of signal processed sound. As shown in FIG. 1, the output sound buffer detection means 15 always detects the amount of signal processed sound stored in the output sound buffering means 13 at a predetermined sampling period, and determines the amount of the signal processed sound. Output to the audio signal processing control means 16.

  The audio signal processing control means 16 controls signal processing in the audio signal processing means 12. As shown in FIG. 1, the audio signal processing control unit 16 receives the amount of signal processed audio stored in the output audio buffering unit 13 from the output audio buffer detection unit 15. Then, the audio signal processing control unit 16 determines whether the amount of the signal-processed audio is less than a predetermined threshold value, and if it is less than the threshold value, the audio signal processing control unit 16 stores it in the input audio buffering unit 11. A control signal indicating that the sound is processed is output to the sound signal processing means 12. As described above, in the audio reproduction device 1, the audio signal processing control unit 16 determines whether or not the signal processing in the audio signal processing unit 12 is necessary according to the amount of audio stored in the output audio buffering unit 13. Audio signal processing can be controlled.

  Here, the above-described threshold value is a value obtained empirically and experimentally in advance, and is determined according to hardware performance (for example, CPU and data transfer speed) that implements or implements the present invention. Further, the above-described threshold value is determined in advance to indicate the stability of the processing performance of the audio reproduction device 1 to the minimum value of the amount of test audio detected by the output audio buffer detection means 15 under the following conditions. Multiplied by the confidence factor. This reliability rate is "1.0" when the CPU of the sound reproducing apparatus 1 is embodied by, for example, a computer, and there are few programs operating in parallel, and a CPU with poor performance. Or, when a plurality of programs are operating in parallel, “2.0” or “3.0” is set by the user in advance.

  That is, the above-mentioned conditions are that a predetermined test sound is signal-processed by the sound signal processing means 12, and a test sound signal-processed by the sound signal processing means 12 is stored by the output sound buffering means 13, When the test sound stored in the output sound buffering means 13 is reproduced by the sound reproduction means 14 and the amount of the test sound stored in the output sound buffering means 13 is detected by the output sound buffer detection means 15 It is.

  In the audio reproduction device 1 having the above-described configuration, the audio signal processing unit 12 has the amount of signal-processed audio (for example, speech-converted audio) stored in the output audio buffering unit 13 below a threshold value. In some cases, the voice is subjected to signal processing (for example, speech speed conversion). As a result, the amount of sound stored in the output sound buffering means 13, that is, the amount of signal processed sound reproduced by the sound reproducing means 14, is always kept constant. Therefore, according to the audio reproducing device 1, since the audio is signal-processed and reproduced while always managing the amount of the signal-processed audio, the signal-processed audio can be stably reproduced without being interrupted.

  Instead of reproducing the speech speed converted speech by converting the speech speed in real time to the input speech as in the speech playback apparatus 1 according to the present invention, for example, the speech speed is collectively converted in advance. If the speech speed converted speech is played after the speech speed conversion of all the voices is completed, the playback sound will not be interrupted. However, in this case, the time at which the speech speed conversion is performed and the time at which the speech after the speech speed conversion is played back are different, so the expansion / contraction rate specified when the speech speed conversion is performed and the speech after the speech speed conversion is played back. This causes a deviation from the expansion / contraction rate specified at the time. In other words, the voice whose speech speed has been converted with the expansion / contraction rate set to “2 times” in advance cannot be changed to a speech speed other than “2 times” during reproduction.

  In this way, conventionally, if the response of the voice speed conversion ratio (expansion rate) is improved by performing the voice speed conversion and playback in parallel, the playback sound may be interrupted. If the playback sound is not interrupted separately from the playback, there is a correlation that the response for specifying the magnification of the speech speed conversion is deteriorated. However, the audio playback device 1 according to the present invention is configured as described above. By providing the above, it is possible to achieve both good response for specifying the magnification of speech speed conversion and the uninterrupted playback sound.

[Processing procedure of audio playback device]
A processing procedure of the audio reproduction device 1 according to the present invention will be described with reference to FIG. 2 (refer to FIG. 1 as appropriate).

  The audio reproducing device 1 first inputs the audio before signal processing from the outside to the input audio buffering means 11 (step S1). Next, the audio reproduction device 1 detects the amount of signal processed audio stored in the output audio buffering means 13 by the output audio buffer detection means 15 (step S2). Next, the audio reproducing device 1 determines whether or not the amount of the signal processed audio is less than the threshold by the audio signal processing control means 16 (step S3).

  When the amount of the signal processed sound is less than the threshold value (Yes in step S3), the sound reproducing device 1 performs signal processing on the sound stored in the input sound buffering means 11 by the sound signal processing means 12 (step S3). S4) The signal processed voice is stored by the output voice buffering means 13 (step S5). Next, the audio reproducing device 1 reproduces the signal-processed audio stored in the output audio buffering means 13 through the audio device such as a speaker by the audio reproducing means 14 (step S6). Then, the audio playback device 1 proceeds to step S8 when the input of audio from the outside is completed (Yes in step S7), and proceeds to step S8 when the input of audio from the outside is not completed (No in step S7). Returning to S1, the above-described processing is repeated.

  The audio reproduction device 1 ends the process when the reproduction of all the signal-processed sounds is completed (Yes in Step S8), and does not complete the reproduction of all the signal-processed sounds (No in Step S8). Returns to step S2 and repeats the process described above. In step S8, as a method for determining whether or not the reproduction of all the signal-processed sounds has been completed, for example, the sound signal processing means 12 and the sound reproduction means 14 may be used in the input sound buffering means 11 and the output sound. For example, a method for detecting whether or not the sound in the buffering means 13 remains can be used.

  On the other hand, when the amount of the signal processed audio in the output audio buffering means 13 exceeds the threshold value (No in step S3), the audio reproducing device 1 stands by without performing signal processing (step S9), and after step S6 Perform the process. The audio reproduction device 1 reproduces the signal processed audio through the above procedure.

[Configuration of speech synthesis playback device]
The configuration of the speech synthesis / playback apparatus 2 according to the present invention will be described with reference to FIG. The speech synthesis / playback apparatus 2 performs signal processing on the synthesized speech and reproduces it. Specifically, as shown in FIG. 3, the speech synthesis / playback device 2 synthesizes speech according to an input sentence (text) input from the outside, The synthesized voice is subjected to signal processing based on a predetermined voice signal processing parameter, and a signal-processed synthesized voice (hereinafter referred to as a signal-processed synthesized voice) is reproduced. The voice synthesis / playback apparatus 2 is used for the same application as the voice playback apparatus 1 described above.

  Here, as shown in FIG. 3, the speech synthesis / playback apparatus 2 includes an input voice buffering means 11, a voice signal processing means 12, an output voice buffering means 13, and a voice that are the configuration of the voice playback apparatus 1 described above. In addition to the reproduction unit 14, the output audio buffer detection unit 15, and the audio signal processing control unit 16, a voice synthesis unit 21, an input audio buffer detection unit 22, and a voice synthesis control unit 23 are provided. In the following description, detailed description of the same components as those of the audio playback device 1 is omitted.

  The voice synthesizer 21 synthesizes a voice corresponding to an input sentence (text). As shown in FIG. 3, the speech synthesis means 21 performs speech synthesis using speech data and speech synthesis parameters based on text input from the outside. The voice synthesizer 21 can synthesize voice using general methods such as an HMM voice synthesis method, a waveform editing method, and a waveform connection method. Examples of the speech synthesis parameters include parameters such as voice volume, speech rate, pitch (fundamental frequency), and spectrum.

  Here, the speech synthesis means 21 does not simply perform speech synthesis from the input sentence, but performs speech synthesis according to the amount of speech already synthesized by itself. That is, as shown in FIG. 3, the speech synthesizer 21 has a case where the amount of synthesized speech (hereinafter referred to as synthesized speech) already stored in the input speech buffering unit 11 is less than a predetermined threshold. Next, speech synthesis is performed from the input sentence.

  More specifically, as shown in FIG. 3, when the synthesized speech stored in the input speech buffering unit 11 is less than the above-described threshold, the speech synthesis control unit 23 sends the speech synthesis unit 21 a speech. A control signal for performing synthesis is input. The speech synthesizer 21 synthesizes speech from the input sentence only when this control signal is input, and outputs the synthesized speech to the input speech buffering unit 11. Here, the speech synthesizer 21 synthesizes the speech in parallel with the signal processing of the synthesized speech by the speech signal processor 12 and the reproduction of the signal processed synthesized speech by the speech reproducer 14. The unit of speech synthesis in the speech synthesizer 21 is not particularly limited. For example, speech synthesis can be performed in units of phrases and sentences.

  Here, the input sound buffering means 11 has the same configuration as that in the sound reproduction apparatus 1 described above, but here, the sound synthesized by the sound synthesizing means 21 is not the sound inputted from the outside. save.

  The input voice buffer detection means 22 detects the amount of synthesized voice. As shown in FIG. 3, the input voice buffer detection means 22 always detects the amount of synthesized speech stored in the input voice buffering means 11 at a predetermined sampling period, and the synthesized speech volume is detected as voice synthesis control means. Output to 23.

  The voice synthesis control unit 23 controls voice synthesis in the voice synthesis unit 21. As shown in FIG. 3, the amount of synthesized speech stored in the input speech buffering unit 11 is input to the speech synthesis control unit 23 from the input speech buffer detection unit 22. Then, the speech synthesis control unit 23 determines whether the amount of the synthesized speech is less than a predetermined threshold, and if it is less than the threshold, the speech synthesis control unit 23 synthesizes a control signal for synthesizing speech from the input sentence. Output to the means 21. As described above, the speech synthesis / playback apparatus 2 determines whether or not the speech synthesis unit 21 needs to synthesize the speech synthesis control unit 23 according to the amount of synthesized speech stored in the input speech buffering unit 11. Speech synthesis can be controlled.

  Here, the above-described threshold value is a value obtained empirically and experimentally in advance, and is determined according to hardware performance (for example, CPU and data transfer speed) that implements or implements the present invention.

  In the speech synthesis / playback apparatus 2 having the above-described configuration, the speech synthesis unit 21 synthesizes speech when the amount of synthesized speech stored in the input speech buffering unit 11 falls below a threshold value. As a result, the amount of synthesized speech stored in the input speech buffering means 11, that is, the amount of synthesized speech that is subjected to signal processing (for example, speech speed conversion) by the speech signal processing means 12, is always kept constant. Also, the speech synthesis / playback apparatus 2 is configured so that the speech signal processing unit 12 has a signal-processed synthesized speech (for example, synthesized speech that has been subjected to speech speed conversion) stored in the output speech buffering unit 13 below the threshold. Process the synthesized speech. As a result, the amount of the synthesized speech stored in the output speech buffering means 13, that is, the amount of the signal processed synthetic speech reproduced by the speech reproducing means 14 is always kept constant. Therefore, according to the speech synthesis / playback apparatus 2, the speech is synthesized while always managing the amount of the synthesized speech, and the synthesized speech is signal-processed and reproduced while always managing the amount of the synthesized speech that has been signal processed. The signal-processed synthesized speech can be stably reproduced without interruption.

[Processing procedure of speech synthesis playback device]
The processing procedure of the speech synthesis / playback apparatus 2 according to the present invention will be described with reference to FIG. 4 (refer to FIG. 3 as appropriate).

  The voice synthesis / playback apparatus 2 first inputs text from the outside to the voice synthesis means 21 (step S11). Next, the speech synthesis / playback apparatus 2 detects the amount of synthesized speech stored in the input speech buffering means 11 by the input speech buffer detection means 22 (step S12). Next, the speech synthesis / playback apparatus 2 determines whether or not the amount of synthesized speech is less than the threshold by the speech synthesis control means 23 (step S13).

  When the amount of synthesized speech is less than the threshold (Yes in step S13), the speech synthesis / playback apparatus 2 synthesizes speech corresponding to the text by the speech synthesis unit 21 (step S14), and the input speech buffering unit 11 The synthesized speech is saved (step S15). Next, the speech synthesis / playback apparatus 2 detects the amount of the signal-processed synthesized speech stored in the output speech buffering means 13 by the output speech buffer detection means 15 (step S16). Next, the speech synthesis / playback apparatus 2 determines whether or not the amount of the signal-processed synthesized speech is less than the threshold value by the speech signal processing control means 16 (step S17).

  If the amount of the signal-processed synthesized speech is less than the threshold value (Yes in step S17), the speech synthesis / playback apparatus 2 performs signal processing on the synthesized speech stored in the input speech buffering unit 11 by the speech signal processing unit 12. In step S18, the output voice buffering means 13 stores the signal-processed synthesized voice (step S19). Next, the speech synthesis / playback apparatus 2 plays back the signal-processed synthesized speech stored in the output speech buffering means 13 via the speech device such as a speaker by the voice playback means 14 (step S20). Then, the speech synthesis / playback apparatus 2 proceeds to step S22 when the input of the text from the outside is completed (Yes in step S21), and proceeds to step S22 when the input of the text from the outside is not completed (No in step S21). It returns to step S11 and repeats the above-mentioned process.

  When the reproduction of all the signal-processed synthesized sounds is completed (Yes in step S22), the audio reproducing device 1 ends the process, and when the reproduction of all the signal-processed synthesized sounds is not completed (in step S22). No) returns to step S12 and repeats the process described above. In step S22, as a method for determining whether or not the reproduction of all the signal-processed synthesized sounds has been completed, for example, the audio signal processing means 12 and the audio reproduction means 14 are used in the input audio buffering means 11 and the output. For example, a method for detecting whether or not the voice in the voice buffering means 13 remains can be used.

  On the other hand, when the amount of synthesized speech in the input speech buffering means 11 exceeds the threshold (No in step S13), the speech synthesis / playback device 2 stands by without synthesizing speech (step S23), and after step S16. Perform the process. If the amount of the signal-processed synthesized speech in the output speech buffering means 13 exceeds the threshold (No in step S17), the speech synthesis / playback apparatus 2 stands by without performing signal processing (step S24), and step The process after S20 is performed. The voice synthesis / playback apparatus 2 plays the signal-processed synthesized voice through the above-described procedure.

[Specific example of speech synthesis playback processing]
A specific example of processing by the speech synthesis / playback apparatus 2 according to the present invention will be described with reference to FIG. Here, as an example, suppose that speech speed is synthesized by synthesizing an input sentence such as “If you don't bear the cost of yourself, isn't it quite rude?” In the following example, the input audio buffering means 11 and the output audio buffering means 13 are initially in a state where no data is stored.

  First, the input voice buffer detection means 22 detects the amount of synthesized voice existing in the input voice buffering means 11 and outputs “0” to the voice synthesis control means 23. Next, since the speech synthesis control unit 23 receives the output “0” of the input speech buffer detection unit 22 and is below the threshold, the control signal indicating that the speech synthesis unit 21 synthesizes speech from the input sentence. Is output. Next, the speech synthesizer 21 generates synthesized speech corresponding to the input sentence and outputs it to the input speech buffering unit 11. Note that at this time, the speech synthesizer 21 generates and outputs a synthesized speech corresponding to “only me” in the input sentence. Next, the input voice buffering unit 11 stores the synthesized voice input from the voice synthesis unit 21.

  Subsequently, as before, the input voice buffer detection means 22 detects the amount of synthesized speech existing in the input voice buffering means 11, and the synthesized voice length of “only me” inputted earlier is synthesized. Output to the control means 23. Next, the speech synthesis control unit 23 receives the output of the input speech buffer detection unit 22 and outputs a control signal for synthesizing speech from the input sentence to the speech synthesis unit 21 when it is below the threshold value. However, if it is equal to or greater than the threshold, nothing is done.

  On the other hand, the output audio buffer detection means 15 detects the amount of signal-processed synthesized speech existing in the output audio buffering means 13 and outputs “0” to the audio signal processing control means 16. Next, since the audio signal processing control unit 16 receives the output “0” of the output audio buffer detection unit 15 and is below the threshold value, the audio signal processing unit 12 performs control to signal-process the synthesized audio. Output a signal. Next, the voice signal processing means 12 obtains the required number of synthesized voices stored in the input voice buffering means 11 in order from the oldest and performs speech speed conversion to generate signal-processed synthesized voices. And output to the output audio buffering means 13. Next, the audio reproduction means 14 acquires the required number of signal-processed synthesized voices stored in the output audio buffering means 13 in order from the oldest and outputs them to an audio device (not shown) such as a speaker. To play.

  Subsequently, as before, the output audio buffer detection means 15 detects the amount of the signal-processed synthesized voice existing in the output audio buffering means 13 and outputs it to the audio signal processing control means 16. Next, the audio signal processing control means 16 receives the output of the output audio buffer detection means 15 and, if it is below the threshold value, gives a control signal to the audio signal processing means 12 to signal-process the synthesized voice. If it is output and the threshold value is exceeded, nothing is done.

  After a while, since the output of the output audio buffer detection means 15 falls below the threshold value due to the reproduction of the signal processed synthetic voice, the audio signal processing means 12 is stored in the input voice buffering means 11. Is converted to speech speed and output to the output voice buffering means 13. At the same time, since the output of the input speech buffer detection means 22 falls below the threshold value, the speech synthesis means 21 performs speech synthesis of the input sentence and outputs synthesized speech to the input speech buffering means 11. In the above processing, the timing of the operation of speech synthesis and speech speed conversion is important, and the threshold value used in the speech synthesis control means 23 and the speech signal processing control means 16 is the minimum value that stabilizes the operation. By setting to, speech synthesis parameters such as the latest emotion and signal processing parameters such as the expansion / contraction rate of speech speed conversion can be reflected in the speech.

[Voice playback program and voice synthesis playback program]
Here, the voice reproduction device 1 and the voice synthesis reproduction device 2 described above can be realized by causing a general computer to operate according to a program that causes each of the above-described units and units to function. This program can be distributed via a communication line, or can be written on a recording medium such as a CD-ROM for distribution.

  Specifically, the audio reproduction program includes an input audio buffering unit 11 that stores the input audio and an output audio buffering unit 13 that stores the audio signal processed by the audio signal processing unit 12. Can be made to function as the above-described audio signal processing means 12 and the above-described audio reproduction means 14. The speech synthesis / playback program also includes an input speech buffering unit 11 that stores the synthesized speech synthesized by the speech synthesis unit 21 and an output speech buffer that stores the signal processed synthesized speech signal-processed by the speech signal processing unit 12. A voice synthesizing / reproducing apparatus 2 including a ring unit 13 and processing the synthesized speech by signal processing to function as the speech synthesizing unit 21, the audio signal processing unit 12, and the audio reproducing unit 14. be able to.

  As mentioned above, although the audio | voice reproduction apparatus and the audio | voice synthesis | combination reproduction apparatus which concern on this invention, and these programs were demonstrated concretely by the form for inventing, the meaning of this invention is not limited to these description. Should be construed broadly based on the claims. Needless to say, various changes and modifications based on these descriptions are also included in the spirit of the present invention.

DESCRIPTION OF SYMBOLS 1 Audio | voice reproduction apparatus 11 Input audio | voice buffering means 12 Audio | voice signal processing means 13 Output audio | voice buffering means 14 Audio | voice reproduction | regeneration means 15 Output audio | voice buffer detection means 16 Audio | voice signal processing control means 2 Audio | voice synthesis | combination reproduction | regeneration apparatus 21 Audio | voice synthesis means 22 Input audio | voice buffer Detection means 23 Speech synthesis control means

Claims (6)

An audio reproduction device that performs signal processing and reproduces audio,
Input voice buffering means for storing the input voice;
Audio signal processing means for signal-processing audio stored in the input audio buffering means;
Output audio buffering means for storing the audio signal processed by the audio signal processing means;
Audio reproduction means for reproducing audio stored in the output audio buffering means;
Output audio buffer detection means for detecting the amount of audio stored in the output audio buffering means;
It is determined whether or not the amount of sound detected by the output sound buffer detecting means is less than a predetermined threshold value, and if it is less than the threshold value, the sound stored in the input sound buffering means is Audio signal processing control means for outputting a control signal for signal processing to the audio signal processing means,
Signal processing is performed on a predetermined test sound by the sound signal processing means,
The output sound buffering means stores the test sound signal-processed by the sound signal processing means,
The audio reproduction means reproduces the test audio stored in the output audio buffering means,
In the case where the amount of test audio stored in the output audio buffering means is detected by the output audio buffer detection means,
The threshold used in the audio signal processing control means is a predetermined reliability ratio indicating the stability of the processing performance of the audio reproduction apparatus to the minimum value of the amount of test audio detected by the output audio buffer detection means. Multiplied by
The audio signal processing unit performs signal processing on audio stored in the input audio buffering unit when the control signal is input from the audio signal processing control unit. An audio reproduction device that performs signal processing and reproduces audio,
Input voice buffering means for storing the input voice;
Audio signal processing means for signal-processing audio stored in the input audio buffering means;
Output audio buffering means for storing the audio signal processed by the audio signal processing means;
Audio reproduction means for reproducing audio stored in the output audio buffering means;
Output audio buffer detection means for detecting the amount of audio stored in the output audio buffering means;
It is determined whether or not the amount of sound detected by the output sound buffer detecting means is less than a predetermined threshold value, and if it is less than the threshold value, the sound stored in the input sound buffering means is Audio signal processing control means for outputting a control signal for signal processing to the audio signal processing means,
The voice signal processing means performs signal processing for converting a speech speed stored in the input voice buffering means when the control signal is input from the voice signal processing control means. Audio playback device. 2. The audio reproducing apparatus according to claim 1, wherein the audio signal processing means performs signal processing for converting a speech speed of the audio stored in the input audio buffering means. A speech synthesis / playback device comprising the speech playback device according to any one of claims 1 to 3,
Speech synthesis means for synthesizing speech corresponding to the input sentence;
Input audio buffer detection means for detecting the amount of audio stored in the input audio buffering means;
Control whether to determine whether or not the amount of speech detected by the input speech buffer detection means is less than a predetermined threshold, and to synthesize speech corresponding to the input sentence if it is less than the threshold Speech synthesis control means for outputting a signal to the speech synthesis means;
With
The speech synthesizer synthesizes speech corresponding to the input sentence when the control signal is input from the speech synthesis controller.
The speech synthesis / playback apparatus, wherein the input speech buffering means stores the speech synthesized by the speech synthesis means.   An audio reproduction program for causing a computer to function as the audio reproduction apparatus according to any one of claims 1 to 3.   A speech synthesis / playback program for causing a computer to function as the speech synthesis / playback device according to claim 4.

Images