JPH02205899A - Voice discriminating system - Google Patents

Abstract

PURPOSE:To easily detect the existence of voice, which is spoken by an actually existing person, with a high detection rate in circumstances of unnecessary ambient sounds by discriminating the existence of this voice on condition that a human body detecting means detects the existence of a human body and an acoustic pressure detecting means which takes a voice detection area as the detection range detects an acoustic pressure higher than a preliminarily determined reference level. CONSTITUTION:A discriminating circuit 30 receives the output of a comparator 14 of a human body detecting circuit 10 and the output of a comparator 24 of an acoustic pressure detecting circuit 20 to discriminate the existence of voice spoken by a person actually existing in the voice detection area on condition that the human body detecting circuit 10 detects the existence of the human body and the acoustic pressure detecting circuit 20 detects an acoustic pressure higher than the reference level. Consequently, the existence of voice is discriminated when the existence of the human body is confirmed by the human body detecting circuit 10. Thus, unnecessary ambient sounds (sounds and noise from acoustic devices) are prevented from being erroneously discriminated as the voice spoken by the actually existing person.

Description

[Detailed description of the invention] [Industrial application field] The present invention relates to a voice determination method.

[Prior Art] Conventionally, methods for detecting the presence of voice in a noisy environment have been proposed in Japanese Patent Publication No. 5614527, Japanese Patent Application Laid-open No. 61-246800, and the like. These methods extract feature parameters from an input signal collected by a microphone.
The presence of a voice is detected only by comparing it with a standard pattern predetermined in the voice detection means.

[The problem that the invention tries to solve! [fi] However, the conventional voice detection method requires a complicated processing device, making it difficult to deploy it for general use.

Further, in the conventional voice detection method, it is not possible to determine whether or not a living human body actually exists in the voice detection area and whether or not a voice is uttered from this human body.

In other words, is there a sound from a tape recorder or TV in the audio detection area?
l! If the surrounding sound has the same characteristic parameters as the standard pattern of sound predetermined by the sound detection means, such as when there is sound based on the recording and playback of these audio equipment or broadcasts, it is detected that the sound is present in the sound detection area. It is mistakenly thought that the sound emitted from the human body exists! This can lead to inconveniences such as malfunctioning of the voice response switch of a hands-free telephone.

The present invention emits sounds from an actual human body in an environment where unnecessary sounds are present in the surroundings. The purpose is to easily detect the presence of voice with a high detection rate.

[Means for Solving the Problems] The present invention as set forth in claim 1 provides a sound pressure detection system in which a human body detection means having a sound detection area as a detection range detects the presence of a human body, and a sound pressure detection unit having a sound detection area as a detection range. Provided that the means detects a sound pressure equal to or higher than a predetermined reference level, it is assumed that there is a sound emitted from an actual human body in the sound detection area↑
It was designed so that the

The present invention according to claim 2 is characterized in that the sound pressure detection means includes a microphone that converts the sound pressure of an input sound into an electrical signal, and a band that selectively passes only audio frequency band from the output of the microphone. The device includes a pass filter and a comparison circuit that outputs a voice detection output when the output of the band pass filter is equal to or higher than a predetermined reference level.

[Function] The present invention according to claim 1 includes a human body detection process, a sound pressure detection process, and a determination circuit. In other words, the existence of a human body is detected in the human body detection process, the presence of sound at a reference level or higher is detected in the sound pressure detection process, and the condition is that a human body exists and sound at a certain level or higher is present in the determination process. ,
It is determined that there is a sound emitted from a real human body.

Therefore, in the present invention as set forth in claim 1, the presence of a voice is determined in a state in which the presence of a human body is confirmed by the human body detection means, ) can be prevented from being incorrectly determined to be the sound emitted by a real human body.

That is, in an environment where unnecessary sounds are present in the surroundings, it is possible to determine with a very high determination rate whether there is a voice emitted from an actual human body.

Further, according to the present invention as set forth in claim 2, in the above-mentioned sound pressure detection process, the sound pressure belonging to the audio frequency band is detected, and as a result, the presence of audio in the audio detection area can be determined with high accuracy. It will be possible.

In other words, even if the sound pressure that does not belong to the audio frequency band is higher than the reference level, it is possible to eliminate the erroneous determination that °C audio is present.

[Example] Fig. 1 is a block diagram showing an example of a sound determination device used for implementing the present invention, Fig. 2 is a schematic diagram showing a sound judgment state, and Fig. 3 is a waveform diagram showing a sound pressure level state. be.

As shown in FIG. 1, the voice determination device of this embodiment is comprised of a human body detection circuit 10, a sound pressure detection port 'IfI 20, and a determination circuit 30, and each circuit is constructed as follows.

(A) The human body detection circuit 10 has a detection range 1 that is a voice detection area (for example, an area within a certain distance of a specific telephone), and an infrared detection element 11 that detects infrared radiation emitted from the human body (see Figure 2). 1) An amplifier 12 that amplifies the output of the infrared detection element 11, a bandpass filter 13 that selectively passes only the operating frequency component of the human body from this output, and an output amplitude of the filter 13 that exceeds a reference level. °
Ctan is generated by the comparison circuit 14 which outputs a human body detection output during the period C.

(The Bl sound pressure detection circuit 20 electrically detects the sound pressure of the input sound [)]
(see FIG. 2), an amplifier 22 that amplifies the output of the microphone 21, and an amplifier 22 that converts the output into
A bandpass filter 23 selectively passes only those in the audio frequency band from among the outputs of the filter 23, and a voice detection output is output when the output of the filter 23 is above a predetermined reference level (threshold value in FIG. 3). The comparator circuit 24 is configured to include an output comparison circuit 24.

(C) ↑ 1/1830 is the output of the comparison circuit 14 of the human body detection circuit 10 and the comparison circuit 24 of the sound pressure detection circuit 20.
1 Human body detection circuit 10 detects the presence of a human body in response to the output of It is determined that the emitted voice is present.

Therefore, according to the above embodiment, the presence of a voice is determined in a state where the existence of a human body is confirmed by the human body detection circuit 10, and unnecessary audible sounds in the surroundings (voices and noises emitted by acoustic audible devices) are detected. It is possible to prevent the sound from being mistakenly judged as a sound emitted by a real human body.

That is, in an environment where unnecessary sounds are present in the surroundings, the presence of voice emitted from an actual human body can be easily determined with a high determination rate.

Further, according to the above embodiment, by using the band pass filter 23 in the sound pressure detection process, the sound pressure belonging to the 11 voice frequency bands is detected, and as a result, the presence of voices in the voice detection area is increased. This means that it can be judged with accuracy. That is, even if the sound pressure that does not belong to the voice frequency band is above a certain level, it is possible to eliminate the false TJI determination that voice is present.

The range of the sound detection area is determined by setting the field of view of, for example, an infrared condenser lens constituting the human body detection means, the amplification factors of the amplifiers of the human body detection means and the sound pressure detection means, and the comparison circuit.
! It can be adjusted by setting the i-level.

Furthermore, in carrying out the present invention, the human body detection means is not limited to the infrared method, but may be based on other detection methods such as a method in which weak radio waves or reflected waves of ultrasonic waves are disturbed by the movement of the human body.

[Effects of the Invention] As described above, according to the present invention, the presence of a voice emitted from an actual human body can be easily detected with a high detection rate in an environment where unnecessary sounds are present in the surroundings.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an example of a sound determination device t used in implementing the present invention, FIG. 2 is a schematic diagram showing a sound fJI constant state, and FIG. 3 is a waveform diagram showing a sound pressure level state. DESCRIPTION OF SYMBOLS 10...Human body detection circuit, 20...Sound pressure detection circuit, 21...Microphone, 23...Band pass filter. 24... Comparison circuit, 3o... Judgment circuit. Steel Diagram 1 Patent Applicant Sekisui Chemical Co., Ltd. Representative Hiroshi 1) Kaoru Diagram 2

Claims (2)

[Claims] (1) The human body detection means whose detection range is the sound detection area detects the presence of a human body, and the sound pressure detection means whose detection range is the sound detection area detects a sound pressure equal to or higher than a predetermined reference level. A voice determination method that determines that there is a voice emitted from an actual human body in the voice detection area. (2) The sound pressure detection means includes a microphone that converts the sound pressure of input sound into an electrical signal, a bandpass filter that selectively passes only audio frequency band from the output of the microphone, and a bandpass filter. 2. The voice determination system according to claim 1, further comprising a comparison circuit that outputs a voice detection output when the output of the voice determination system is equal to or higher than a predetermined reference level.