JPS6041096A - Voice input unit - 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] "Hachikyenki" Tree 9. Please adjust the level of 1F 7jy
The present invention relates to a voice input device that can be set to a certain level.

In a voice recognition and speech analysis system, the amplitude of the voice signal changes depending on the speaker's psychological condition, environment, and degree of fatigue. The conventional methods for keeping this 1" size at a certain level, and the size of the '4 size, are to use AGC (automatic 1 inch quick adjustment device) and to first register the size and reset it. However, the former method has the disadvantage of generating amplitude distortion in which the sound 111 changes during the voice section because the 111j constant is set for the 111 detection. The rear right side had the drawback of "9." that it was not possible to make detailed adjustments to the individual i'T I, j levels.

The present invention has been made in view of the actual situation as described in -, and in particular, 8: Re-recognition system 1\ or sound J! I-bumasushi'
C・te on the stem, manually powered:'? This was done for the purpose of setting '114'+ji to an appropriate level from 1 to 1.

1-Issei The structure of the present invention will be described below with reference to embodiments.

First, the principle of the present invention will be explained with reference to the waveform diagram of FIG.

FIG. 1(a) shows the power level of the input voice, the horizontal line P in the figure shows the threshold for detecting the end of the voice section, and the circles A, B, and C show the end positions. FIG. 1(b) shows the output waveform of the first peak hold circuit which holds the peak value of the first voice, is reset at point B, and holds the peak value of the third voice. Figure 1 (C) shows the output of the second peak hold circuit which holds the peak value of the previous audio and is reset at point A, and holds the peak value of the second audio and is reset at point C. Shows the waveform. FIG. 1(d) shows the control pulse for extracting the control signal from the peak hold (+r4) in FIG. 1(b), and shows how it is extracted at a high level. FIG. 1(e) , as in Fig. 1(d), a control pulse for extracting a control signal from the peak hold value in Fig. 1(C) is included. Fig. 1(f) shows the termination detection signal. In other words, the tree invention is based on the assumption that the loudness of successive sounds will not change suddenly when they are adjacent, and therefore the loudness of the succeeding sound will be the same as the previous sound. I am trying to control it based on the assumption that it will be the same as the large S.

FIG. 2 is a diagram for explaining an embodiment of the present invention constructed according to the above principle, in which the input audio from the input terminal 1 is
The signal is applied in parallel to the termination detector 2, the first peak hold circuit 3, and the second peak hold circuit 4, and is also input to a volume adjuster 9, such as a voltage-controlled amplifier. In addition, the peak hold circuits 3 and 4 receive control pulses from the termination detection section 2 (FIG. 1(d)) and 7 (FIG. 1(e)).
) is controlled by The output of each peak hold circuit becomes a control signal 10 through a selection changeover means 5 such as a switch, for example, to control a volume adjuster 9. this! The ilJ switching stage 5 is controlled by the control kr''i 8 (usually using FIG. 1(d) or FIG. 1(e)) from the termination detector 2, and the volume adjuster 9 is gradually adjusted according to the volume of the sound. However, although this embodiment performs tone adjustment between consecutive voices, it does not effectively use the dynamic range of the system.

Figure ff53 is a diagram showing an embodiment in which the dynamic range is effectively used. In Figure 1, the same reference numerals as in Figure 2 are used for parts that have the same effect as in Figure 2. The difference between this embodiment and the embodiment shown in FIG. : peak hold value, E: clip level voltage). If controlled in this way, the level can be set effectively using the dynamic range, which is always uniform when the signal is small and becomes large when it is large, and small when it is large. can.

Although each embodiment of the present invention has been described above,
The present invention is not limited to the above-mentioned embodiments; for example, the volume adjuster may be realized by a multiplication type DA converter, using a digital signal by converting the control signal from an audio signal. Needless to say. It goes without saying that the end detector may use the end signal of the voice section detector.

As is clear from the explanation given above, according to the invention, the volume level does not change during the voice section, so the volume level can be adjusted during the voice section like an automatic volume regulator with a time constant. Since so-called varying amplitude distortion does not occur, an audio signal of an appropriate level can be obtained.

[BRIEF DESCRIPTION OF THE DRAWINGS] FIG.
FIGS. 2 and 3 are configuration diagrams for explaining the present invention in detail, respectively. ■...Input end, 2...Termination detection section, 3.4...Peak hold circuit, 5...Switching means, 9...Volume adjuster, 12...Arithmetic circuit. Procedural Amendment (Ki) Kazuo Wakasugi, Commissioner of the Patent Office 1. Indication of the case 1982 Patent Application No. 149384 2. Name of the invention Voice input device; 3. Address of person making the correction Tokyo Metropolitan University 1-ku Nakamagome 1-chome 3-6 Name (first name T*) (674) Ricoh Co., Ltd. Representative: l1
5 Hama ul Hiro 5, Date of amendment order 7, Contents of amendment (1), Details page 4, line 11 [Pulse (Fig. 1 (
d) Correct J to "Pulse 6 (Figure 1 ((1)) J. (2), Correct Figures 2 and 3 as shown in the attached sheet.

Claims (1)

[Claims] H)-"Ft loud input device, which includes an end detection means for detecting the end of an audio signal, and two parallel-connected peak hold means for detecting and holding the peak value of the audio signal. and one control from the peak hold means 4";"j
9. A voice human power device comprising: a selection means for selecting a voice human power signal; and a volume adjustment stage R for adjusting the sound level of the voice human power signal using the control signal. (2) The audio input device according to claim 1, further comprising calculation means for obtaining control (1-1) for effectively utilizing the dynamic range.