JPH0562755B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a voice response device that responds to humans with voice. Voice response devices have overcome technical difficulties as extremely easy-to-use man-machine interfaces, and are gradually expanding their field of use.

In a voice response device, by outputting in words that are easy for humans to hear, it is possible to reduce confusion caused by mishearing. for example,
It is desirable to add the name of the digits and the unit to be notified to the numeric string to be output, as this will make it much easier to hear.

[Prior Art] Some conventional voice response devices have a function of adding digits and units to a number string. A method of preparation was taken. In addition, in simple versions, units were added after the number string.

[Problems to be Solved by the Invention] In the above-mentioned conventional voice response devices, the former method requires a very large amount of voice parameters to be stored, while the latter method requires a large amount of stored voice parameters to be prepared between the number string and the unit. There was a slight gap between the two, which lacked naturalness. The present invention aims to provide a device that can respond by voice using numbers with digits and units without impairing naturalness while having a relatively small amount of voice parameter storage.

[Means for Solving the Problems] The present invention provides a voice response device that solves the above problems. A voice parameter memory with digits for storing voice parameters;
The number string with unit code inputted from the host computer is inspected to determine the final reading digit, and the number of the final reading digit and the unit code are output, and the number above the final reading digit is used as an index key to read the voice parameter with digit. A conversion function with digits that searches the memory, converts it into audio parameters with digits, and outputs them, and a unit audio parameter memory that stores audio parameters for pronouncing numbers with units such as yen, piece, book, etc. A unit addition function that searches the unit audio parameter memory using the unit code outputted from the conversion function with digits and the number of the final reading digit as an index key, and converts and outputs the audio parameter with units, and output of the conversion function with digits. The voice parameters with digits and the voice parameters with units output from the unit addition function are input, and these are converted into digits and voice parameters with units.
The audio parameter assembly memory assembles and stores the audio parameters of numbers with units, and the audio parameter data of numbers with digits and units stored in the audio parameter assembly memory is read out according to the storage completion signal output by the unit addition function. The voice response device of the present invention is equipped with a voice synthesizer that synthesizes voice from data and utters it, and is configured to read out numeric voice with digits and units in response to input of a numeric string with a unit code from a host computer. achieved by.

[Operation] When a number string with a unit code is input from the host computer by the above means, first, in the conversion function with digits, the digits higher than the final reading digit are converted into voice parameters with digits, and the final reading digit is ,
They are converted into unit speech parameters by the unit addition function, the outputs of both are assembled in the speech parameter assembly memory, and the speech synthesizer converts them into "1".
This is output as a voiced number string with digits and units, such as "12,300,415 yen."

[Example] The gist of the present invention will be specifically explained below with reference to Examples shown in FIGS. 1 to 3.

FIG. 1 is a circuit configuration block diagram of a voice response device according to the present invention. In the figure, 1 is a voice response device, 2 is a memory for storing voice parameters with digits, 3 is a conversion function with digits, 4 is a memory for storing unit voice parameters, 5 is a unit addition function, 6 is a voice parameter assembly memory, 7 is a speech synthesizer,
8 each indicates a number string input into the voice response device.

When a number string 8 with a unit code is input from a host computer (not shown), it is entered into a conversion function 3 with digits, where the number string is examined and the final reading digit ( For example, if the number string is 432000, determine the 2,000 digit), call the audio parameter memory 2 with digits, convert it to audio parameters with digits, and send the output to the audio parameter assembly memory 6. The unit code and the number string below the final reading digit are sent to the unit addition function 5.

The unit addition function 5 compares the input unit code and numeric string with a built-in unit code table, indexes the unit voice parameter, and sends it to the voice parameter assembly memory 6.

The speech parameter assembly memory 6 combines the speech parameter outputs from the digit conversion function 3 and the unit addition function 5 and outputs this to the speech synthesizer 7.

The speech synthesizer 7 generates and outputs voice with digits and units in response to input of the activation signal from the unit additional function 5 based on the input voice parameters.

FIG. 2 shows the contents of voice parameters stored in the voice parameter memory with digits 2 and the unit voice parameter memory 4. As shown in the figure, if the required number of digits is n, the amount of parameters stored in the voice parameter memory 2 with digits is (9×n), and the amount of parameters stored in the unit voice parameter memory 4. is (9×n) per unit type.

Now, the number of digits that can be processed is 7 digits (up to 99999999), and the types of units that can be processed are yen, book, piece, piece, g,
If there are 6 types of Kg, the amount of stored parameters will be 72 voice parameters with digits and (72×6) unit voice parameters.

FIG. 3 is a conceptual diagram showing how a number string is converted into a number string with digits and units.

(a) shows a situation in which a numerical string with a unit code "23452345 yen" is converted to 2,345-52,345-5 yen. 2,304 to 52,304 are converted using the digit conversion function, and 5 yen is converted using the unit addition function. As can be seen in the figure, in numbers 23 and 4, the same audio parameter is used twice.

(b) shows a situation in which the numerical string with unit code "24502000 pieces" is converted to 24502000 pieces. Up to 2,450,000 is converted using the digit conversion function, and up to 2,000 is converted using the unit addition function.

[Effects of the Invention] As explained above, according to the present invention, when a numerical string is input, it is possible to output a voice with digits and units without unnaturalness with a relatively small amount of voice parameter storage. This has a significant effect on improving the man-machine interface.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a circuit configuration of a voice response device according to the present invention, FIG. 2 is a diagram showing how voice parameters are stored, and FIG. 3 is a conceptual diagram of an example of voice conversion with digits and units. In the figure, 1 is a voice response device, 2 is a voice parameter memory with digits, 3 is a conversion function with digits, and 4 is a voice response device.
5 indicates a unit speech parameter memory, 5 indicates a unit additional function, 6 indicates a speech parameter assembly memory, 7 indicates a speech synthesizer, and 8 indicates an input number string.

Claims (1)

[Scope of Claims] 1. A voice response device connected to a host computer and outputting a synthesized voice in response to a unit coded number string output from the host computer, the voice for uttering numbers with digits. Voice parameter memory 2 with digits for storing parameters
and examines the number string with unit code inputted from the host computer, determines the final reading digit from the lower zero, outputs the number of the final reading digit and the unit code, and outputs the number higher than the final reading digit. A conversion function 3 with digits that searches the audio parameter memory 2 with digits as an index key, converts it into audio parameters with digits, and outputs the same; and a unit audio parameter that stores audio parameters for pronouncing numbers with units. memory 4
Then, input the unit code and the number of the final reading digit output from the conversion function 3 with digits, search the unit voice parameter memory 4 using the number and unit as an index key, convert it into a voice parameter with units, and output it. A unit addition function 5 receives the voice parameters with digits output from the conversion function 3 with digits and the voice parameters with units output from the unit addition function 5 as input, and assembles and stores the voice parameters of numbers with digits and units. and a voice parameter assembly memory 6 in accordance with the storage completion signal output from the unit additional function 5.
a voice synthesizer 7 that reads voice parameter data of numbers with digits and units stored in the computer, synthesizes voice from the data, and utters the voice; A voice response device characterized in that it is configured to read aloud numbers with digits and units.