JPH02716B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a speech synthesis device using a microcomputer as a control section, and is capable of reducing the memory in which speech data is stored as much as possible.

The degree of integration of integrated circuits is increasing at a remarkable rate year by year, and large-capacity memories are now available at relatively low prices. On the other hand, the spread of microcomputers and advances in synthesizer technology are also spectacular, and the new field of speech synthesis is suddenly gaining attention due to these connections. Speech synthesis, which until now was only possible with the help of large-scale computers, has now reached the stage where it can be achieved using just a few integrated circuits, and its application to home appliances has already begun.

However, since voice data generally requires a huge amount of memory and requires a high-speed dedicated control unit to continue processing, it was inevitable that installing a voice synthesizer would significantly increase the system price. .

The present invention relates to a speech synthesis circuit in which the control section is configured with a microcomputer that is inexpensive and easy to develop. The configuration is such that an address specifying means with a counter function is provided for updating addresses in the middle.

Generally, in order to clearly synthesize speech, it is necessary to read one piece of data from a memory in which the speech data is stored every several tens of microseconds to several hundred microseconds and synthesize the data. In other words, the control unit is required to have a processing speed of several kHz to several tens of kHz. However, microcomputers (hereinafter referred to as microcomputers) have slow operating speeds, and it takes several microseconds to several tens of microseconds to execute one instruction.
It takes seconds. To update memory addresses one after another and detect the end address each time requires tens to hundreds of steps, no matter how well programmed. In other words, even if you select a high-speed microcontroller and program it well, the question is whether you can specify memory addresses one after another.

Furthermore, even if a high-speed microcontroller were developed in the future, it would be impossible to do other work for a few seconds to tens of seconds while audio is being synthesized. is missing.

In consideration of these points, the present invention aims to solve the problem by providing a microcomputer with only the responsibility of specifying the starting address and detecting the ending address of the memory, and providing an address specifying means with a counter function for updating the address in the middle. There is.

Furthermore, in the present invention, in order to reduce memory, only predetermined words are synthesized instead of sentences, and in order to prevent the operator from overlooking the synthesized voice information, an alarm sound is synthesized before the voice information. It is configured to do this. The alarm sound, like other audio information, is stored in memory as digital data, and synthesized by the microcomputer outputting its start address. Thereafter, the pause period is counted by an internal timer of the microcomputer, and after a predetermined silent period has elapsed, the start address where predetermined audio information is stored is output.

With this configuration, it is possible to realize a speech synthesis circuit by using a microcomputer with a slow processing speed as the control unit, and also to add a silent period using a timer means while the microcomputer edits words, phrases, and alarm sounds. This allows efficient use of expensive memory. Furthermore, by combining the alarm sounds, it is possible to prevent the operator from overlooking the notification. Additionally, the microcontroller can perform audio editing in parallel while performing tasks such as display.

Hereinafter, one embodiment of the present invention will be described based on the accompanying drawings.

FIG. 1 is a perspective view of a main body showing an example in which the present invention is installed in, for example, a microwave oven, and FIG. 2 is a detailed view of the main parts thereof.

A door 12 is provided on the front surface of the main body 1 so as to be openable and closable, and an operation panel 2 is disposed thereon. At the rear of the door 12 is a heating chamber (not shown) for storing the food to be cooked. The operation panel 2 has an output key 3 for setting the output, a numeric key 4 for setting the heating time, a display section 5 for displaying this, a start key 6 for instructing the start of heating, and a key for erasing the input program. There are provided a clear key 7 for outputting the synthesized sound from the speaker, and a slit 8 for extracting the synthesized sound from the speaker to the outside of the main body.

In this embodiment, the operator of the microwave oven first opens the door 12 and places foods such as potatoes and meat into the heating chamber. Then, a desired output key 3 on the operation panel 2 is selected and pressed, and then a number key 4 is pressed to set a desired heating time. For example, select full power “P1” and set “15.00” (15 minutes)
Set.

3 and 4 show an example of the circuit of this embodiment.

Various keys 3, 4, 6, and 7 on the operation panel 2 are swept with scan signals SC ₄ to _{SC 1} to form a key matrix 10 and are connected to input ports of the microcomputer 11.
Connected to IN ₇ ~ IN ₂ . This microcomputer 11 includes a door switch 13 that detects whether the door 12 is opened or closed.
Information from _a sensor 14 such as a humidity sensor, temperature sensor, or infrared sensor that detects the heating state of the heated object is sent via an amplifier 15.
A power synchronization signal serving as a timer decrement pulse for controlling the timer control unit 16 is input to IN ₀ and 50/60Hz ports, respectively.

Note that the scan signal that sweeps the key matrix
SC ₄ to SC ₁ dynamically light up the 5-digit display section 5 together with SC ₀ . Display data is output to Seg ₇ to Seg ₀ .

The door switch 13' is inserted into the main circuit and directly controls the power supply to the magnetron 17. TC is a timer control signal, and PC is an output control signal. PC
The output control unit 18 is intermittently controlled by the output controller 18 to change the average output of the magnetron 17.

With this configuration, the output keys and numeric keys on the operation panel pressed by the operator of the microwave oven are outputted to the input ports IN ₇ to IN ₂ of the microcomputer 11 by the key matrix 10 . Microcomputer 11
decodes the signals input to IN ₁ to IN ₂ based on a predetermined program stored in internal memory (not shown) and various logic circuits (not shown),
Understand the power and heating time specified by the operator.

Next, when the operator presses the start key 6, this is decoded in the same manner as above, and the timer control signal TC and output control signal PC are changed to H level in order to start heating. As a result, power supply to the magnetron 17 is started. In other words, heating of the food begins. If full power is selected, the PC
The signal remains at H level, but if any other output is selected, the PC signal repeats intermittent cycles at a predetermined period.

When heating is started, the sensor 14 inputs the heating state of the food to the microcomputer 11 every moment.

Next, the configuration of the speech synthesis section will be explained. The voice synthesis section is composed of a microcomputer 11, an address counter 19 and a word counter 20 as address designation means, a memory 21 for storing voice data, a DA converter 22 as a conversion means for converting voice data into an analog quantity, and a speaker 23.

FIG. 6 is a voice data map of the memory 21, and for example, a cooking procedure announcement saying "TORN-OVER" (repeat) is sent from address X'1000' to
In addition to being stored at address X′1FFF′, various types of cooking are also stored. On the other hand, mixed with these audio data, buzzer sound data, that is, pulse sounds of about several kHz, are stored at addresses X'0000' to X'03FF'.

Next, the procedure for synthesizing voices will be explained using the time chart shown in FIG. sensor 14 during heating
When the output from the address counter 19 reaches a certain predetermined value, for example, if it is time to return a batch of food, the microcomputer 11 first transfers 16 bits from the VS ₁₅ to VS ₀ terminals to the VS ₁₅ to _{VS 0} terminals of the address counter 19. The address signal X'0000' is output (1 in FIG. 5).

When a pulse (2 in Figure 5) is input from the microcomputer 11 to the CLA terminal of the address counter 19,
Address counter 19 and word counter 20
clears the previous value.

Next, when the SET signal (3 in Fig. 5) is input to the LOAD terminal of the address counter 19 from the SET terminal of the microcomputer 11, the address counter 19 stores the 16-bit signal from the microcomputer 11 via the VS ₁₅ to VS ₀ terminals. .

The CLK terminal of the AND circuit is always connected to the clock signal.
A 4kHz pulse signal (CLK signal in Figure 5) is being sent. Also, the other end of the AND circuit is of microcomputer 11.
It is connected to the CNT terminal, and this CNT terminal is L.
When changing from level to H level (4 in Figure 5),
A 4kHz pulse signal from the clock is output to the UP terminal of the address counter 19 via the AND circuit (UP signal in Figure 5).

When the address counter 19 receives a 4kHz clock pulse signal from the UP terminal, it updates the stored 16-bit value one by one for each pulse (0000→0001→0002...).

The 16-bit value of address counter 19 is output to terminals A ₁₅ -A ₀ , and this output is connected to terminals A ₁₅ -A ₀ of memory 21 .

The memory 21 transfers the 8-bit digital data stored at the addresses specified by terminals _A15 to _A0 to terminal _D7.
~D Output to ₀ . Therefore, after the audio data at address 0000 is output, the audio data at addresses 0001 and 0002 are sequentially updated.

Such audio data is input to terminals D ₇ to _{D 0} of the DA converter 22, which is a conversion means, and is converted into an analog quantity and output to terminals I out and I out.

Analog data output from the DA converter 22 is appropriately amplified and waveform-shaped via DA out and Audio AMP, and then restored as audio by the speaker 23.

In other words, the buzzer sound stored in addresses 0000 to 0400 is restored, and the buzzer starts to sound "pitsu".

To detect the end address of the audio data, count the UP signal input to the word counter 20, count the carry signal CRY output when a carry occurs by the microcomputer 11, and detect the data end address 03FF of the buzzer sound. It is done by doing.

When such an end address is detected, the voice synthesis is stopped and the various voice control signals return to their original states.
In other words, the buzzer continues to sound for 250ms and then stops.

After this, the pause continues for 750ms due to the internal timer of the microcomputer 11 (FIG. 7). After going through the same buzzer sound playback cycle again, "TORN-OVER" at address 1000 is synthesized. This procedure is exactly the same as when synthesizing buzzer sounds, only the final address is changed to 1FFF address. Note that this final address varies depending on the key first pressed by the operator, that is, the set output and heating time, and a program in the microcomputer 11 extracts audio information corresponding to the content.

When address 1FFF is detected by the microcomputer 11 by counting the pulses of the CRY terminal, the microcomputer 11 determines that all voice synthesis has been completed.
Change the output of the CNT terminal from H level to L level (5 in Figure 5) and finish the work.

With this configuration, “Pitsu, Pitsu, TORN”
The cooking instructions are synthesized into voice, such as "OVER".

In a series of sentences, it is often possible to get the gist from the context even if some words cannot be heard, but in speech synthesis, which requires a large amount of memory, it is difficult to synthesize such long sentences. , you can't help but arrange the words like in this example. Therefore, there is a high possibility that the information will be overlooked, and given the importance of the information, it is necessary to ensure that the system operator overhears it. Prior to speech synthesis, the present invention synthesizes an alarm sound whose meaning does not need to be audible, and synthesizes the sound after calling the operator's attention. This eliminates the possibility of overhearing without requiring a large amount of memory. few.

When the operator restarts cooking after turning the food over based on this notification, the microcomputer 11 starts supplying power to the magnetron 17 again, and the PC signal continues until the remaining time of the set heating time is counted. and controls the TC signal to H level.

When all the set times have been counted, the microcomputer 11 returns the PC signal and the TC signal to the L level, ends the power supply to the magnetron 17, and uses the voice synthesizer to notify the end.

When the cooking is finished, the operator can take out the cooked food from the heating chamber.

As described above, according to the present invention, a speech synthesis circuit can be realized by using a microcomputer with a slow processing speed as a control unit, and while the microcomputer edits words, phrases, and alarm sounds,
Moreover, a silent period can be added by a timer means, and expensive memory can be used effectively. Furthermore, by combining the alarm sounds, it is possible to prevent the operator from overlooking the notification. In addition, the microcomputer can perform audio editing in parallel while performing tasks such as display.

[Brief explanation of drawings]

FIG. 1 is a perspective view of the main body of the device showing an embodiment of the present invention, FIG. 2 is a detailed view of the main parts, FIGS. 3 and 4 are circuit diagrams, and FIG. 5 is a timing chart of each control signal. FIG. 6 is a memory map, and FIG. 7 is a diagram showing speaker output. 19...Address counter, 20...Word counter, 21...Memory, 23...Speaker.

Claims (1)

[Claims] 1. A memory that stores an alarm sound and sounds such as multiple words or phrases as a series of digital data at consecutive addresses, and the start and end addresses of the series of digital data stored in the memory, and the desired address. a microcomputer that controls audio notification; an internal timer formed in the microcomputer; an address specifying means for specifying an arbitrary address in the memory and outputting digital data stored at the address; It consists of a clock signal for automatically updating the address of the designating means, a converting means for converting a series of digital data outputted from the memory into analog audio, and a speaker connected to the converting means. When the computer notifies certain audio information, it first outputs the starting address of the memory where the alarm sound is stored, and when it is set in the address designation means, the clock signal automatically sets the address of the address designation means. A series of digital data of an alarm sound is output one after another from the memory, and the conversion means converts this into an analog alarm sound and notifies the alarm sound from the speaker. The end address of the alarm sound is detected by counting the carry signals, and the address specifying means is stopped, and the internal timer is started to count a predetermined pause period, after which the desired sound is memorized. When the starting address is outputted again and set in the address specifying means, the clock signal automatically updates the address of the address specifying means, and a series of digital data of the desired audio is output one after another from the memory, The converting means converts this into analog audio and broadcasts the desired audio from the speaker, and the microcomputer detects the end address by counting the carry signals of the address specifying means and ends the audio notification. Configured speech synthesizer.