JPS60237495A - Voice recognition equipment - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed description of the invention] Technical fields> The present invention relates to a Japanese speech recognition device.

<Prior art> Conventional speech recognition devices were not equipped with display devices for displaying speech rate and recognition rate, which would be used as a guideline for speakers to improve their recognition rate. I couldn't realize it.

<Object of the Invention> An object of the present invention is to provide a speech recognition device that visually displays the speech rate and recognition rate, and is capable of uttering while trying to improve the recognition rate.

<Example> Hereinafter, the configuration of the present invention will be explained using examples.

FIG. 1 shows a block diagram of the apparatus of the present invention.

A voice signal input to the microphone 1 is amplified and digitally converted by the analog input section 2, and then introduced to the voice analysis section 3 and the syllable segmentation section 4. Voice analysis section 3
divides the input speech signal into frames of about 16m8, performs spectrum analysis, and transfers information such as the characteristic pattern information of the speech and the number of Hower zero crossings to the syllable segmentation unit 4-\ at intervals of about 8m8.

The syllable segmentation section 4 cuts out syllables based on various information sent from the speech analysis section 3.

It also identifies silent sections and sound sections of the input audio,
Time measurement is performed for each section every 8 mS, and the measured data is sent to the CPU 5 as calculation data of the speaking rate.

The characteristic patterns of the syllables extracted by the syllable segmentation unit 4 are stored in the pattern memory 6 through the CPU 5.
The monosyllable recognition unit 9 compares the standard pattern with a standard pattern registered in advance in the standard pattern memory 8. The comparison result is returned to the CPU 5, which stores it in the recognition result storage memory 10 and displays it on the display device 12. The keyboard 11 is used to modify the displayed input voice.

FIG. 2 shows a flowchart of the utterance rate calculation processing program of the syllable segmentation unit 40.

Silent sections and sound sections are repeated alternately. Let Pi and Vi be the count numbers for time measurement in the silent section and the sound section, respectively, P are the contents of the register that cumulatively stores the count number P1 of the silent period, and the contents of the register that cumulatively store the count number ■1 of the sound period. is V, and the threshold value of the silent section P1 is l. This threshold value j is set so as to ignore consonants and word-initial syllables.

Step 81. Each register and counter are cleared in S2. When it is determined that it is a silent section, the counter P1 is counted every 8 mS (S5), and when it is determined that it is a sound section, the counter P1 is counted every 8 mS (S6). That is, P 1←P 1 +1 Vi-Vi-1-1 is executed. Next, P i ) /3 is determined and NO
If so, the accumulation of P←P-1-Pi V←V+V is executed (S9.510), and then the count of cut out syllables C-C+1 is executed (S11). 11':>89 if l is yes. The processing of SlO and Sll is omitted. Finally, in S'12, the average time T of one syllable and the average speech rate M every 32 seconds are calculated using the following equations.

1000 M=- This calculation result is transferred to the CPU.

FIG. 3 shows a flowchart of a program for calculating the identification rate. At S21, counters C1 and C2 are cleared, and the syllables cut out by the above processing are input (
S22), its content is recognized (f924), the recognition result is displayed on the display device 12 (S25), and this process is continued until the end of the word. Here, if the uttered content and the displayed content are different, the recognition result is corrected using the keyboard (EI27). This modification is
This is carried out when it is necessary to correct the recognition result of the candidate ranked lower than the plurality of recognition result candidates. After modification, CPU
A comparison is made for each syllable between the word of the recognition result of the -th candidate and the word of the recognition result corrected by confirmation or correction (S28 to 533), and the recognition rate is calculated l[ =□ 2 is executed, and the speech rate M is calculated and displayed. For example, when you pronounce ``hajime,'' the first character ``ka'' is displayed as ``kajime.''
After correcting it to ``ha'', the identification rate is calculated and displayed as -66.7chi.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of an embodiment of the present invention, and FIGS. 2 and 3 are flowcharts showing programs of the embodiment of the present invention. Patent applicant Nyaap Co., Ltd. Agent Patent attorney Nishi 1) Shin

Claims (1)

[Claims] A monosyllable recognition device for recognizing Japanese speech in units of syllables, comprising a speech rate calculation means, a discrimination rate calculation means, a speech rate display means, and a discrimination rate display means.