JPH0143320B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a method for audio outputting numerical information of an audio output device (for example, an audio alarm clock, an audio output type calculator, an audio alarm tester, etc.) that outputs the content to be output by audio. The purpose is to minimize the amount of data for speech synthesis and to obtain high-quality synthesized speech.

Generally, when numerical information is outputted as voice, the numerical value as voice can be constructed by combining several phonemes. For example, 2534 ``Nisen gohyaku sanjiyuyon'' is ``ni'' and ``sen''.
It can be constructed by combining seven phonemes: ``go'', ``hyaku'', ``san'', ``jiyu'', and ``yon''. A method is known in which desired numerical information is outputted as voice by storing a required number of basic phonemes in a memory and extracting them in a desired order and performing sound synthesis.

However, as a result of our research, it has become clear that such simple combinations of basic phonemes may cause problems in some cases.

For example, 123 “Hiyaku Ni Jiyu San”
'Hiyaku' in 234 'Ni Hiyaku Sanjiyuu Yon' or 'Hiyaku' in 100 are all pronounced the same, but each has slightly different changes in pitch and amplitude, so they are not treated as one phoneme. If they are shared, the pronunciation may become quite unnatural. Therefore, unless three types of phonemes are used for "hiyaku" depending on the situation, high-quality speech cannot be output. In addition, similar to this, the applicant has already proposed a method for storing different word data for "Jiyuu" for teenagers and "Jiyuu" for people in their 20s to 30s, but the present invention further improves the relationship between these. By carefully considering the following, and by providing not only place value words but also several types of number sounds,
This was done to obtain a higher quality synthesized sound.

FIG. 1 shows each phoneme and an example of its use. For example, "jiyu" in 20, 30, 40 and "jiyu" in 50, 60, 80, etc. use different data. Also 50,5
1,000 “go” and 500, 50,000 “go” and 500 million “go”
each uses different data. Figure 1a shows an example of a phoneme with place value, and n ₁ n ₃ in Figure 1b.
n ₇ n ₁₀ n ₁₂ n ₁₈ n ₂₂ n ₂₅ n ₂₉ n ₃₂ are the voices of single-tone numbers, and are used, for example, when reading the numbers below the decimal point. Furthermore, FIG. 2 shows fewer uses than in FIG. 1, and is effective when it is necessary to minimize the storage capacity of audio data. In the case of FIG. 2, for example, "3" indicates that one data is used in all cases.

An embodiment of the present invention will be described below with reference to the drawings. FIG. 3 is its block diagram, where X is a numerical data storage register that stores numerical values such as calculation results as numerical data, and x is a decimal point that stores the decimal point position of the numerical value stored in the numerical data storage register X. It is a position memory register, and both are configured in RAM (random access memory). Also, OC is an output control unit that extracts the contents of the X register in the order of audio output and outputs it to buffer B for one digit, and J ₁ is the decimal point position stored in x and the value of which digit in the X register is determined by OC. The digit of the numerical value to be outputted as audio is determined based on the signal S ₁ indicating whether it has been output to buffer B, and the signals S _2o and S _2d are outputted to instruct which data to use depending on the contents of buffer B. do. Further, if the content output to B is in the first decimal place, the signal S ₃ is output together with S _2o and S _2d .

Output date GO is CG ₂ output if S ₃ is received
Output the code to VCC in the order of CG ₁ output, and if S ₃ is not received, then CG ₁ output, then CG ₂ output.
Output code to VCC. _J2 is determined by the decimal point position stored in x and the signal _S1 indicating which digit of the X register the OC outputs to the buffer B.
It determines the digit of the numerical value to be output as audio and instructs CG ₂ which scale data to use. CG ₁ and CG ₂ are code generators, and CG ₁ is a numeric voice code generator, which gates the code of the phoneme to be output as voice based on the contents of buffer B and the selection signal _S2o output from _J1 .
Output to VCC via. In addition, CG ₂ is a scale audio code generation unit, and uses the scale signal output from J ₂ and the selection signal S _2d output from J ₁ to generate the code of the phoneme to be output as audio via gate GO.
Output to VCC. VCC is a voice synthesis section, and CC
is code converter, AC is address counter, AD
is an address decoder, and VR is a memory that stores the data that forms the basis of audio. For example, if the code for "Ichi" is input to the CC, the CC converts it to the code of the first address where the voice data of "Ichi" is stored, the AC counts up the address, and the VR inputs the "Ichi". Sequentially output audio data to D/A. D/A is a digital-to-analog converter, which ultimately outputs an audio signal.
Audio is output from the audio output section of the SP. JE is VR
Determine whether or not the END code has been output.
Outputs detection signal to OC and GO.

For example, if the content of register X is 24356 and the content of register x is 2, then
56 is stored in memory, and first the output control section
The OC inputs "2" in the hundreds place into buffer B.
Therefore, the numeric voice code generator _CG1 generates the code _n5 in FIG. 1 based on the contents of the signal _S2o and the buffer B. Further, the scale voice code generating section _CG2 generates the code _d12 in FIG. 1 based on the scale selection signal outputted from the signals _S2d and _J2 . Output gate GO
Since the signal S ₃ is not generated, the output code of CG ₁ is first output to VCC, and "ni" is output as a sound.
When this audio output ends, GO receives the end signal Se, then outputs the output code of CG ₂ to VCC,
Outputs "Hiyaku" aloud. Then the output control section
OC outputs the tens place to buffer B. in this case
CG ₁ generates the code n ₁₁ , CG ₂ generates d ₈ , and since the signal S ₃ is not generated, the output gate outputs the code to VCC in the order of the output of CG ₁ and the output of CG ₂ . ” is output aloud.

Also, if the 1's digit is output to buffer B,
CG ₁ generates n ₈ in Figure 1, but CG ₂ does not generate any code. Therefore, the sound "san" is output.
Next, if the first decimal place is output to buffer B,
CG ₁ outputs the code of n ₁₃ and CG ₂ outputs the code of d ₂₁ . Also, at this time, since _J1 outputs the signal _S3 , the output gate GO outputs the code to VCC in the order of the output of _CG2 and the output of _CG1 . Therefore, "ten" and "go" are output as sounds. If the second decimal place is output to buffer B, CG ₁ generates a code of _n18 , but CG ₂ does not generate a code. As a result, "Roku" is output audibly. As described above, "ni", "hiyaku", "yon", "jiyu", "san", "ten", "go", and "roku" are outputted as voices as a whole.

As is clear from the above explanation, the code generation section,
By storing the codes shown in Figure 1 in CG ₁ and CG ₂ and configuring them to output the code selected by an external selection designation signal, the necessary phonemes will be combined, and even if the same reading is used, subsequent Since different audio codes are selected depending on the numerical conditions, etc., and the optimal audio is output at each time, it is possible to obtain synthesized sounds of very high quality.
Numerical values can be announced using natural sounds.

Furthermore, embodiments of the present invention include the following.

(1) “Hiyaku” in his 100s and “Hiyaku” in his 200s to 900s
A means of storing different data, 100 or 200
An audio output device comprising means for determining whether the person is in the 900s and outputting audio with different sounds according to each of the above data.

(2) A means of storing mere 100 “Hiyaku” with different data from 100 to 900 “Hiyaku”;
An audio output type device comprising means for determining whether the age is 100 or 100 to 900 and outputting audio with different sounds according to each of the above data.

(3) A means to memorize simple 10 “JIU” as data different from “JIU” of 10 to 90s, 10
A voice output type device comprising means for determining whether a person is in their 10s to 90s and outputting a voice with a different sound according to each of the above data.

(4) Means for storing sounds of the number of words with no words added before and after the number as different data, and sounds of the number of words other than that, determining whether the number is a single sound or not, and storing each sound according to each of the above data. An audio output device comprising means for outputting audio with different sounds.

(5) means for storing 4000 "sens" and 1000, 2000, 5000 to 9000 "sens" as different data;
4,000 "sens" or not, and outputs audio with different sounds according to each of the above data.

(6) Means for storing 200, 500 "Hiyaku" and 400, 700, 900 "Hiyaku" as different data, and means for discriminating them and outputting audio with different sounds according to each data. Audio output device.

As explained above, according to the audio output device of the present invention, means for generating a plurality of types of audio code signals respectively corresponding to the same numerical linguistic information and scale linguistic information is provided, and a combination of numerical information and its scale information is generated. Since the code signal generation means is driven based on the determined output to select and output a predetermined code, it is possible to automatically output audio based on the predetermined numerical language information and its scaled language information. This has the advantage of being able to significantly improve the quality of synthesized speech.

[Brief explanation of drawings]

1 and 2 are diagrams showing phonemes and examples of their use, and FIG. 3 is a block diagram of an audio output type device according to the present invention. In the figure, X: X register, x: x register,
OC: Output control circuit, B: 1-digit buffer,
J ₁ : Numerical discrimination section, J ₂ : Scale discrimination section, CG ₁ : Numerical voice code generation section, CG ₂ : Scaled voice code generation section, GO: Output gate, VCC: Voice synthesis section,
CC: Code converter, AC: Address counter,
AD: Address decoder, VR: Audio data storage section, JE: END code detection section, D/A: Digital to analog conversion section, SP: Audio output section.

Claims (1)

[Scope of Claims] 1. An audio output device configured to output numerical values aloud using the numbers constituting the numerical value and place value words, comprising: a pronunciation data storage unit that stores pronunciations for each place of the numerical value; and numerical data that stores the numerical value. a storage register; a decimal point position storage register for storing the decimal point position of the numerical value stored in the numerical data storage register; an output control means for sequentially outputting digits from the numerical data storage register; a discriminating means for discriminating the digit of the number output from the numerical data storage register according to the position of the decimal point; What is claimed is: 1. A voice output device, comprising: a voice synthesizing section that selectively outputs a voice, synthesizes it into voice, and outputs the voice, and outputs the numerical value of the numerical data storage register as voice with scale.