JPH0367278B2 - - Google Patents

Description

[Detailed description of the invention] (a) Industrial application field The present invention converts an input audio signal into a code string with information compression by analyzing it in real time, and converts this code string into a code string in a synthesis mode. The present invention relates to a real-time speech analysis and synthesis device that decodes and synthesizes and outputs speech signals.

(b) Prior Art A real-time speech analysis and synthesis device has a high-speed calculation function and includes a speech analysis section for analyzing and extracting information-compressed parameters that indicate the characteristics of speech in real time, such as Percoll coefficients, and The main components are a memory for storing parameters obtained by the section, and a speech synthesis section for synthesizing speech based on the parameters in the memory, and an operation means for controlling the operations of these components. is added.

Conventionally, in this type of device, the operation means include an analysis mode switch for instructing the voice analysis unit to perform its analysis operation, an analysis start command switch and an analysis stop command switch, and furthermore, a voice synthesis unit for instructing its synthesis operation. A large number of switches such as a composition mode switch, a composition start command switch, a composition stop command switch, etc., were provided to issue commands, making their operation complicated and potentially causing erroneous operation. .

(c) Purpose of the invention The present invention has been made for the purpose of solving the above-mentioned drawbacks,
The present invention provides a real-time speech analysis and synthesis device that prevents erroneous operations.

(d) Structure of the Invention The real-time speech analysis and synthesis device of the present invention includes a display means for distinguishing and displaying an analysis mode and a synthesis mode, and an operation means provided corresponding to the display means, and the display means When the display means is displaying the analysis mode, the audio signal is analyzed by operating the operation means, and when this analysis processing is completed, the display of the display means shifts to the synthesis mode, and the display means When displaying the synthesis mode, the voice signal synthesis process is executed by operating the operating means.

(e) Embodiment FIGS. 1a and b show an external view of an embodiment of the real-time speech analysis and synthesis device of the present invention, and a key arrangement diagram of the operating section of the main body, and FIG. 2 shows the functions of the device of the present invention. 2 is a block diagram showing the configuration, and the functional configuration of FIG. 2 will be explained below with reference to FIG. 1. FIG. Reference numeral 1 denotes a voice analysis circuit, which analyzes and extracts a Percoll coefficient sequence in real time based on a voice signal obtained from an external microphone 2. Reference numeral 3 denotes a voice synthesis circuit, which synthesizes voice based on the Percoll coefficient sequence and outputs the voice from an external speaker 4. 5 is a parameter RAM in which the Parcoal coefficient sequence obtained from the voice analysis circuit 1 is stored;
The percoll coefficient sequence is supplied to the speech synthesis circuit 1. 6 is the above-mentioned speech synthesis circuit 1 and speech analysis circuit 2
This is a processor that controls the operation of the processor, and the control of this processor is performed based on the program ROM 7. Reference numeral 8 denotes an input/output port connected to the processor 6, through which signals are exchanged with the outside. 9 is an external ROM writer connected to the input/output port 8, which writes the percall coefficient string stored in the parameter RAM 5.
Writing to ROM is performed. Reference numeral 10 denotes an operating section for operating the apparatus by acting on the processor 6 through the input/output port 8 or directly. In the operation unit 10, 11 indicates a state in which the speech analysis circuit 1 can perform an analysis operation for each speech unit to be analyzed and synthesized, that is, an analysis mode, and a state in which the speech synthesis circuit 3 can perform a synthesis operation, that is, a synthesis mode. A plurality of LEDs, in this case four LEDs, are provided as display means for identifying and displaying the mode, and for each audio unit, the LED 11... indicates the analysis mode when it is off, the analysis mode when it is lit, and the synthesis mode. There is. 12... are a plurality of, in this case four, press switches (having two states, pressed and not pressed) provided as operating means, each corresponding to the above-mentioned LED 11. When the LED 11 corresponding to each press switch 12 is in the analysis mode with the light off, the analysis process is executed in the voice analysis circuit 1 only while the press switch 12 is being pressed; When the press switch 12 is pressed while the corresponding LED 11 is lit every 12 times and is in the synthesis mode, the voice synthesis circuit 3 executes the synthesis process. 1
3 is a reset switch for resetting the processor 6, and by setting the device to the initial state with the switch 13, all of the LEDs 11 are reset.
... is turned off and the analysis mode is reset for all audio units. Reference numeral 14 denotes a state changeover switch for switching between a stand-alone state in which the device operates alone and a state in which it operates in collaboration with other external devices; when selecting the latter state,
For example, the parcall coefficients stored in the parameter RAM 5 can be sent to the external ROM writer 9. 15 is a bit rate changeover switch for changing the bit allocation amount when extracting Percoll coefficients in the speech analysis circuit 1;
Either 2.4 kbit/sec or 4.8 kbit/sec is selected. In FIG. 1B, 16 is a power switch, and 17 are five light emitting elements constituting a level meter that indicates the level of the audio signal input to the microphone 2.

Next, the operation of the apparatus of the present invention will be explained based on the timing diagram of FIG. First, by pressing the reset switch 13 as shown in FIG.
(light off) state, and all audio units are put into analysis mode. Next, as shown in S1 of the same figure, when the first press switch 12' is turned on (pressed), the operation of the voice analysis circuit 1 is started, and this switch 12' is turned on.
For example, when the microphone 2 is turned on, if you input a voice into the microphone 2, as shown in "Analysis I" in Figure M, the voice analysis circuit 1 performs a percall of the voice "konnichiwa" in real time, as shown in "Analysis I" in Figure M during the ON period. Analyze and extract the coefficient string and set the first parameter of parameter RAM5.
Store in the appropriate area. And this switch 12'
When the switch is turned off (released), the voice analysis circuit 1 stops its operation and turns on the first LED 11 corresponding to the first press switch 12' as shown in L1 in the same figure.
Turn ON (light on), from now on it will be like this.
For audio, the analysis mode cannot be used, and the synthesis mode is automatically entered, and this synthesis mode is maintained. Following this, as shown in FIG. L2, with the second switch 12' turned on,
For example, if the voice ``Goodbye'' is uttered in the same way, during the ON period, as shown in M ``Analysis'' in the same figure,
The speech analysis circuit 1 extracts the Percoll coefficient sequence of the speech "Goodbye" and stores it in the second corresponding area of the parameter RAM 5. And this switch 1
When 2' is turned OFF, the second
Turn on LED11'' and say goodbye from now on.
The synthesis mode is maintained for the voices. After that, as shown in S2 of the same figure, the second
When the second press switch 12'' associated with the LED 11'' is turned on to start the operation of the speech synthesis circuit 3, the second press switch 12'' The associated Percoll coefficient string of the voice "goodbye" is sequentially read out from the second area of the parameter RAM 5 and introduced into the speech synthesis circuit 3, and the synthesized voice "goodbye" is output from the speaker 4. . still,
The end of the operation of the synthesis circuit 3 at this time is the parameter
This is the point in time when all the "goodbye" Percall coefficient strings have been read out from RAM3. After this, every time you turn on the second press switch 12'', it says "goodbye".
A synthesized voice will be output. On the other hand, as shown in FIG. S1, the first press switch 12' associated with the first LED 11' which is in the ON state is turned on
When the voice synthesis circuit 3 starts operating, the voice ``konnichiwa'' associated with the first press switch 12' is synthesized and is emitted from the speaker 4, as shown in ``Synthesis'' in Figure M. It will be output.

Therefore, even if there are a plurality of audio units to be analyzed, the operator of such a device can read the analysis results immediately after the analysis by operating one press switch 12 associated with each audio unit. It is possible to listen to synthesized speech using a certain Percoll coefficient sequence and immediately evaluate it. Furthermore, since an LED 11 and a press switch 12 are provided for each audio unit for a plurality of audio units, analysis and synthesis can be performed at random. In addition, the operator must press one push switch 1 for each audio unit.
Since the only operation required is to press 2, it is possible to effectively analyze the audio in the required time period without any delay in operation.

(f) Effects of the Invention As is clear from the above description, the real-time speech analysis and synthesis device of the present invention includes a display means for distinguishing and displaying the analysis mode and the synthesis mode, and an operation means provided corresponding to the display means. and when the display means is displaying the analysis mode, voice analysis processing is executed by operating the operation means, and when this analysis processing is completed and the display means is displaying the synthesis mode. ,
Since the speech synthesis process is executed by operating the operation means, in other words, at least the start and end of the analysis process and the start of the synthesis process can be operated with the same operation means, so the operability is improved. Similar to the above, it is possible to try to synthesize unanalyzed speech, or to delete the compressed speech code string from the memory obtained from the previous speech analysis and perform a new speech analysis. Misoperation can be resolved.

[Brief explanation of drawings]

1a and 1b are external views of the real-time speech analysis and synthesis device of the present invention, and a key arrangement diagram of its operation section;
The figure is a block diagram showing the configuration of the apparatus of the present invention, and FIG. 3 is a timing diagram showing the operation of the apparatus of the present invention. 1...Speech analysis circuit, 3...Speech synthesis circuit, 5
...Parameter RAM, 10...Operation unit.

Claims (1)

[Claims] 1. In an analysis mode, an input audio signal is analyzed in real time to convert it into a code string with information compression and stored in a memory, and in a synthesis mode, the code string in the memory is decoded to generate audio. A real-time speech analysis and synthesis device for synthesizing and outputting signals includes a display means for distinguishing between an analysis mode and a synthesis mode by a lighting state, and an operation means having a switch configuration provided corresponding to the display means. In preparation, when the display means identifies and displays the analysis mode, by operating a switch on the operating means, analysis processing of the audio signal is executed, and when this analysis processing is completed, the display on the display means is synthesized. A real-time device characterized in that the mode is automatically shifted to a state where the mode is identified and displayed, and when the display means is identifying and displaying the composition mode, the composition processing is executed by operating a switch on the operation means. Speech analysis and synthesis device.