JPH08129397A - Voice synthesis system - Google Patents

Abstract

PURPOSE: To directly input PCM audio data of a host computer to a DA converter by newly providing a path through which data are also received from a host interface by the setting from the host computer. CONSTITUTION: PCM audio signals outputted from a host interface 3 are inputted to a selector 15 and are connected to an audio data register 14, which holds the PCM audio data, through the selector 15. When the host computer is going to input data to a DA converter 5, the selector 15 is switched from the selection side of an audio synthesizer 4 to the selection side of the interface 3. Simultaneously, a sampling period is set on an interval timer 16. Then, the host computer sets PCM audio data in the register 14 through the interface 3. The output of the register 14 is inputted to the converter 5.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice response system having a function of connecting a computer for processing text data to a voice synthesizer having a DA converter and reading a text by voice.

[0002]

2. Description of the Related Art Speech synthesis technology has become widely adopted in computer application fields. Recently, with the advance of semiconductor technology, a voice synthesis LSI has been put on the market, and has come to be standardly equipped in a relatively small information processing apparatus such as a personal computer.

A voice synthesizer is a device that receives text data or word data from a host computer, generates voice data, and outputs it as voice. FIG. 5 shows a configuration example of a general speech synthesizer at this time.

The voice synthesizer mainly comprises a CPU (hereinafter referred to as a sub CPU) 53 for processing voice data, a memory 54,
55, synthesizer 51, and DA converter (DAC) 52
Etc. are built in.

The text data is first analyzed by the processing of the sub CPU 53, and the memory 54 is analyzed by the read data.
Is read-accessed and converted into voice segment data in the memory 54, and further DSP (digital signal processor)
Etc., the speech unit data is arithmetically processed by the
It is converted into PCM voice data expressed in a format and output as an analog signal via the DA converter 52.

[0006]

In the above-mentioned conventional voice synthesis card, the DA converter 52 is directly connected to the synthesizer 51 such as DSP for arithmetically generating the PCM voice signal or to the output buffer such as FIFO. Therefore, the PCM audio signal cannot be input from the host computer and the sub CPU 53 to the DA converter 52.

Therefore, for example, the host computer has PCM audio data for multimedia such as music and sound effects, but a high-quality audio circuit (DA converter / low-pass filter / amplifier etc.) sufficient to reproduce them.
In the case of not including the above, there is a problem in that those functions cannot be effectively utilized due to the above circumstances, although the voice synthesizing device includes a circuit with good sound quality.

The present invention has been made in view of the above circumstances, and by adding a small amount of hardware, a new path for receiving data from the host interface by setting from the host computer is newly provided. DA of voice synthesizer for PCM voice data
The voice P from the host computer can be directly input to the converter and the high quality voice circuit of the voice synthesizer is used.
It is an object of the present invention to provide a voice synthesis system capable of reproducing a CM signal.

[0009]

The speech synthesis system of the present invention receives speech units from a built-in database, and
In addition to a voice synthesizer for calculating and generating CM voice data, a DA converter for converting PCM voice data generated by the voice synthesizer into an analog voice signal, and a PCM voice data generated by the voice synthesizer, a host computer A means for selectively supplying the generated PCM audio data to the DA converter, and the DA when the PCM audio data generated by the host computer is selected.
Means for matching the sampling frequency of the converter with the sampling frequency of the PCM audio data.

In addition to the above structure, a FIFO memory (in which the sub CPU of the voice synthesizer has a sufficient working memory is used as a substitute) in which PCM voice data is continuously written via a host interface unit is provided. It is characterized in that PCM audio data is read out at every sampling cycle by a signal from an interval timer, set in an audio data register, and an analog audio signal is obtained through a DA converter.

[0011]

In the conventional structure, the DA converter receives the data directly from the voice synthesizer, but the present invention is provided with means for switching the signal on this path so that the host interface can receive the data. It is configured to be switched according to the setting from the computer. However, according to this configuration, the host computer needs to prepare the next data for each sampling cycle of the PCM audio signal and set it in the DA converter in real time, which causes a problem that the timing condition of the host computer becomes strict.

Therefore, the load on the host computer can be reduced by providing a FIFO buffer in the above path and transmitting and receiving a large amount of PCM voice data all at once. Alternatively, if the sub CPU with a built-in voice synthesizer has a sufficient working memory, the PCM voice data is collectively transferred from the host computer to the sub CPU by using this memory as a buffer, and the sub CPU performs real time D
Transfer voice data to A converter.

By having such a function, the PCM voice data of the host computer can be transferred to the DA of the voice synthesizer.
A voice PC that can be input directly to the converter and that the host computer uses by using the high-quality voice circuit of the voice synthesizer
It is possible to reproduce the M signal.

[0014]

An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing an embodiment in which the speech synthesis system of the present invention is configured by an IC card.

In FIG. 1, the voice synthesis IC card 1 is
It is configured by connecting the analog circuit portion 11 to the outside. In the embodiment of the present invention, since the voice synthesis system is realized by the IC card, the card interface controller (card I / F controller) 3 substitutes for the host interface (host I / F) during the operation period.

Reference numeral 1 is a voice synthesis IC card, which receives text data from a host computer (not shown) and reads it aloud. Reference numeral 2 is a card connector, which connects a host computer (not shown) to the device according to the present invention.

Reference numeral 3 is the above-mentioned card interface controller, which transmits a card control signal and text data obtained from the host computer to the sub CPU 8. In addition, a signal switching circuit (that is, a circuit configured such that the DA converter can be switched according to a setting from the host computer so that the DA converter can receive data from the host interface) is also added.

Reference numeral 4 is a voice synthesizer, and the sub CPU 8
It receives a speech unit from PCM and calculates PCM speech data, and is mainly composed of a DSP. Reference numeral 5 is a DA converter (DAC), which converts PCM audio data into analog audio. Here, a low pass filter (LPF) is added to the output of the DA converter (DAC) 5 in order to remove the sampling clock signal. The data bus width of the DA converter 5 is 16 bits.

Reference numeral 6 is an input / output controller, which controls bus timing generation of the sub CPU 8 and timer interrupts for requesting data to be passed to the voice synthesizer 4 at fixed time intervals.

Reference numeral 7 is a memory controller, which generates a memory control signal and a refresh control signal required by a built-in memory such as DRAM. Reference numeral 8 is a sub CPU
That is, the control information from the host computer is interpreted to manage the speech synthesizer, and the sentence / word data from the host computer is analyzed and converted into speech unit data, which is supplied to the speech synthesizer 4.

Reference numeral 9 is a ROM memory, which is a parameter for synthesizing pronunciations of dictionary data words for analyzing sentences by the speech synthesizer 4, that is, speech unit data and sub CPU.
The program used by 8 is stored.

Reference numeral 10 is a DRAM, and the sub CPU 8
Used as a working memory. In the embodiment of the present invention, 16
Bit DRAM is used. Reference numeral 11 is a voice analog circuit, and the voice analog circuit 11 performs volume adjustment and amplification. Reference numeral 12 is an earphone in which the user actually listens to the voice of the read sentence. Reference numeral 13 is a connector for connecting the audio circuit to the main body.

The present invention provides a card interface controller (card I / F controller) in a voice synthesis system composed of an IC card having the above-mentioned components.
Host interface (host I / F)
3, a voice synthesizer 4, and a DA converter (DAC) 5
A signal switching circuit for PCM audio data is interposed between the above and.

2 to 4 are block diagrams each showing a configuration example of a signal switching circuit for switching and selecting PCM audio data, which is a feature of the present invention. In FIG. 2, reference numeral 1
Reference numeral 4 is an audio data register for latching audio data input to the DA converter (DAC) 5, and is composed of a 16-bit register.

Reference numeral 15 is a host interface (host I /
F) 3, a voice synthesizer 4, and a DA converter (DA
C) is a selector (or multiplexer) interposed between the PCM voice data and the PCM voice data input from the host interface (card interface controller) 3 and the PCM voice data input from the voice synthesizer 4.

Reference numeral 16 is a timer, which outputs a timing pulse at each sampling cycle of the audio data being reproduced. The timer 16 is configured by using a programmable interval timer whose period can be changed by a command received via the host interface 3.

3 and 4 are block diagrams showing other examples of the configuration of the signal switching circuit. In the second embodiment shown in FIG. 3, reference numeral 17 is a FIFO buffer memory, and here, a 2-port (input / output separation) asynchronous type memory of 16-bit bus × 1K word is used.

In the third embodiment shown in FIG. 4, F shown in FIG.
Sub-CP with built-in voice synthesizer instead of IFO memory
A working memory (DRAM) 10 in the U is used. The operation of the embodiment of the present invention will be described below.

First, the operation of the first embodiment of the present invention will be described with reference to FIG. In this embodiment, the PCM audio signal output from the host interface 3 is input to the selector 15 and further connected via the selector 15 to the audio data register 14 which holds the PCM audio data.

The host computer is a DA converter (D
When you want to input data to AC) 15, first select 1
5 from the selection side of the speech synthesizer 4 to the host interface 3
The sampling period is set in the interval timer 16 at the same time as switching to the selection side.

After that, since the PCM audio data request is issued from the timer 16 every sampling cycle, the host computer sets the PCM audio data in the audio data register 14 via the host interface 3.

The output of the audio data register 14 is input to the DA converter (DAC) 5 and the DA converter (DA)
From C) 5, an analog voice signal according to PCM voice data can be obtained. During the above period, the sub CPU 8 is not functioning at all.

Next, the operation of the second embodiment of the present invention will be described with reference to FIG. In the second embodiment shown in FIG.
A FIFO memory 17 is added to the first embodiment shown in FIG.

In the second embodiment, the host computer sets the sampling period in the interval timer 16 after switching the selector 15 to the FIFO memory 17 side.

After confirming from the status signal that the FIFO 17 is empty, the host computer continuously writes the PCM audio data until the FIFO memory 17 becomes full.

The interval timer 16 reads out data from the FIFO memory 17 at every sampling cycle, inputs the data into the audio data register 14, and outputs it to the DA converter 15.
Get analog voice output from.

Thereafter, the FIFO of the host interface 3
When the status is polled and the data in the FIFO memory 17 becomes small, the operation of continuously writing the data may be repeated. The sub CPU 8 is not functioning at all during the first half period.

Next, the operation of the third embodiment of the present invention will be described with reference to FIG. In the third embodiment shown in FIG. 4, the PCM audio data bus described above is directly connected to the local bus of the sub CPU 8 shown in FIG. Both buses have a width of 1
It is 6 bits.

In the third embodiment, the host computer passes through the host interface 3 and the sub CPU 8
The PCM voice data is output in the same procedure as when the text data is passed to.

The sub CPU 8 stores the received PCM voice data in the work memory (DRAM) 10. Also,
The sub CPU 8 uses the interval timer 16 as the PCM.
The data sampling period is set, and the voice data is set in the voice data register 14 in real time by the interruption from the timer 16.

At this time, the selector 15 is moved from the voice synthesizer 4 to the memory (DRA) according to the setting of the input / output controller 6.
M) Switch to the CPU bus side connected to 10. The PCM audio data set in the audio data register 14 is converted into an analog audio signal by the DA converter (DAC) 5 and output.

During the above period, the sub CPU 8 only delivers PCM voice data, and does not process Japanese sentences or synthesize voice. As described above, in the system according to the embodiment of the present invention, since the path for receiving the data from the host interface 3 is provided by the setting from the host computer, the PCM voice data of the host computer is voice-synthesized. The voice PCM from the host computer can be directly input to the DA converter of the device and the high quality voice circuit of the voice synthesizer is used.
The reproduction of the signal can be enabled.

[0043]

As described above in detail, according to the present invention, a path for receiving data from the host interface can be newly set by setting from the host computer, so that the PCM voice data of the host computer can be transmitted to the D of the voice synthesizer.
It is possible to directly input to the A converter, and it is possible to reproduce the voice PCM signal from the host computer by using the high quality voice circuit of the voice synthesizer.

That is, according to the present invention, a host computer that generates PCM voice data, a voice synthesizer that receives voice units from a built-in database and performs arithmetic operations to generate PCM voice data, and a voice synthesizer that generates voice data. DA converter for converting the generated PCM voice data into an analog voice signal, the PCM voice data generated by the voice synthesizer, and the P generated by the host computer.
Means for selectively supplying CM audio data to the DA converter, and P generated by the host computer
When the CM audio data is selected, a means for matching the sampling frequency of the DA converter with the sampling frequency of the PCM audio data is provided, and a path is newly established so that the data can also be received from the host interface. Since the PCM voice data of the host computer can be directly input to the DA converter of the voice synthesizing device, the high quality voice circuit of the voice synthesizing device can be used to output the voice PCM signal from the host computer. Can be played.

According to the speech synthesis system of the present invention,
A host interface unit that mediates PCM voice signals output from a host computer, a voice synthesizer that receives voice segments from a built-in database and arithmetically generates PCM voice data, and a voice data register that temporarily holds each PCM voice signal. And an interval timer that supplies a timing signal for holding data to the audio data register, a PCM audio signal output from the host interface unit and a PCM audio signal output from the audio synthesizer, A selector for supplying one of the PCM audio signals to the audio data register, and a PC set in the audio data register
A DA converter for converting an M audio signal into an analog signal, a memory unit for storing a PCM audio signal received from the computer via a host interface unit, and a sampling cycle of the PCM audio signal for the interval timer are set, Sub C that sets audio data in the audio data register by timer interrupt
The host computer is provided with a PU and an input / output controller for instructing the selector to switch the connection to a CPU bus to which a synthesizer or a memory is connected, and a new path is provided so that data can be received from the host interface. It has a configuration that can be switched by setting from
In addition, since a large amount of PCM voice data can be transmitted and received at once, the PCM voice data of the host computer can be directly input to the DA converter of the voice synthesizer,
The high-quality voice circuit of the voice synthesizer can be used to reproduce the voice PCM signal from the host computer, and the processing load on the host computer at that time can be greatly reduced.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of a voice synthesis system according to an embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration example of a signal switching circuit according to the first embodiment of the present invention.

FIG. 3 is a block diagram showing a configuration example of a signal switching circuit according to a second embodiment of the present invention.

FIG. 4 is a block diagram showing a configuration example of a signal switching circuit according to a third embodiment of the present invention.

FIG. 5 is a block diagram showing a conventional configuration example.

[Explanation of symbols]

1 ... Voice synthesis IC card, 2 ... Card connector, 3 ... Card interface controller (host interface), 4 ... Voice synthesizer, 5 ... DA converter (DA
C), 8 ... Sub CPU, 14 ... Audio data register, 1
5 ... Selector, 16 ... Interval timer, 17 ... FIF
O memory.

Claims (5)

[Claims] 1. A host computer for generating PCM voice data, and a voice synthesizer for receiving voice segments from a built-in database and performing arithmetic operations to generate PCM voice data.
A DA converter for converting PCM voice data generated by this voice synthesizer into an analog voice signal, a PCM voice data generated by the voice synthesizer, and a PCM voice data generated by a host computer are selected for the DA converter. And a means for matching the sampling frequency of the DA converter with the sampling frequency of the PCM audio data when the PCM audio data generated by the host computer is selected. Synthesis system. 2. A computer for processing text data,
A host interface unit that mediates PCM voice data output from this computer, a voice synthesizer that receives voice units from a built-in database and performs arithmetic operations to generate PCM voice data, and a voice that temporarily holds each PCM voice signal. Receiving a data register, PCM voice data output from the host interface unit and PCM voice data output from the voice synthesizer,
A selector for supplying either one of the PCM audio data to the audio data register by a switching signal obtained via the host interface unit and a DA converter for converting the PCM audio signal set in the audio data register into an analog signal. A voice synthesizing system comprising. 3. The computer has an internally generated PCM.
When supplying the audio signal to the DA converter, it instructs the selector to connect to the host interface unit,
3. The voice synthesis system according to claim 2, wherein a sampling period is set for the interval timer, and voice data is held in the voice data register in synchronization with this period. 4. A P via a host interface unit
It has a FIFO memory in which CM audio data is continuously written, reads PCM audio data at every sampling cycle by a signal from an interval timer, sets it in an audio data register, and obtains an analog audio signal via a DA converter. The voice synthesis system according to claim 3. 5. A host interface unit for mediating PCM voice signals output from a computer, a voice synthesizer for receiving voice segments from a built-in database and calculating and generating PCM voice data, and temporarily holding each of the PCM voice signals. A voice data register, an interval timer for supplying a timing signal for holding data to the voice data register, a PCM voice signal output from the host interface unit and a PCM voice signal output from a voice synthesizer. In response to this, a selector for supplying one of the PCM audio signals to the audio data register, a DA converter for converting the PCM audio signal set in the audio data register into an analog signal, and a host interface unit from the computer. Received through A PCM for the memory unit that stores the received PCM voice signal and the interval timer.
A sub CPU that sets the sampling period of the audio signal and sets the audio data in the audio data register by a timer interrupt, and an input / output that instructs the selector to switch the connection to the CPU bus to which the synthesizer or memory is connected. A voice synthesis system, comprising: a controller.