JP2014055783A - Movement sound detection device and operation method thereof - Google Patents

Abstract

PROBLEM TO BE SOLVED: To easily and accurately detect a faint movement sound emitted by user's movement even in an environment where background sounds and sounds independent of his or her movement exist.SOLUTION: A movement sound detection device includes: a background noise elimination unit 3 which eliminates background sounds by subjecting a sound signal from a microphone amplifier 2A to phase inversion and amplification and combining the resultant signal with a sound signal from a microphone amplifier 2B; an A/D conversion unit 4 which converts a sound signal resulting from background sound elimination into a digital signal; a frequency analysis unit 5 which performs frequency analysis of the digital signal to divide an entire frequency band; a sound pressure extraction unit 6 which extracts sound pressures from a low frequency band lower than an audible band and a high frequency band higher than the audible band; and a movement sound detection unit 7 which detects a movement sound emitted by user's movement on the basis of the extracted sound pressures.

Description

  The present invention relates to an operation sound detection device and an operation method thereof.

  In order to save energy, monitor, and watch over the elderly, various methods for detecting the presence of a user and the presence or absence of movement using an infrared sensor, a camera, a microphone, and the like have been proposed. The infrared sensor detects the intensity of infrared rays emitted from the human body and the intensity of the infrared rays caused by the movement of the human body, thereby controlling illumination and the like. In the case shown in Non-Patent Document 1, the camera is mounted on a television, performs image recognition processing such as face recognition and motion detection, detects not only the presence of a person but also the presence or absence of viewing of the television, and performs energy saving control. Yes. Thus, when using an infrared sensor or a camera, there exists a problem that the range which can be sensed is limited by the direction. Infrared sensors are vulnerable to movement straight ahead toward the sensor, and the camera has a privacy problem, such as taking a picture directly.

  Therefore, proposals have been made to collect ambient sounds using an omnidirectional microphone and to utilize them for user motion detection. In the case shown in Non-Patent Document 2, the presence of a user is detected from the intensity of sound pressure of sound collected by a microphone, and daily life patterns are estimated.

Satoshi Tsurumi, “Image Recognition and Human Interface 3. Human Interface Using Image Recognition 3-1 Intelligent Human Sensor”, Journal of the Institute of Image Information and Television Engineers, Vol. 64 No. 12 Page. 1809-1811 Yoshimitsu Shinagawa, Toshio Kishimoto, Shigeru Ota, “Home Action Monitor Link Using Microphone Sensor”, Journal of Kawasaki Medical Welfare Society, 15 (2), 615-620, 2006

  When the microphone is used as a human sensor, for example, it picks up noise generated from outside the house, vibration or sound generated from the air conditioning or refrigerator in the house, and this may become steady noise. In Non-Patent Document 2, as a countermeasure against such background noise, a sound pressure threshold value is provided in order to distinguish silence and sound, and only an output value equal to or greater than that value is employed. However, there is a problem that weak sound pressure data is removed depending on the threshold value, and it is difficult to detect a fine operation sound.

  In addition, sound generated regardless of the user's movement, for example, sound generated from the speaker of the device, such as sound generated from a television or telephone, may cause the operation sound to be erased, resulting in erroneous detection. It is inconvenient to pick up the sound produced by the speaker of the home device when only the user's operation sound such as footsteps, seating sound, and sound of moving objects is to be detected.

  Furthermore, conventionally, in order to increase the S / N ratio under a noise environment, it has been performed to reduce the sensitivity of the microphone and pick up only the nearby sound. In that case, it was necessary to install a plurality of microphones even in the same room in order to sense a wide range.

  The present invention has been made in view of the above, and an object of the present invention is to easily and accurately detect a minute operation sound generated by a user even in an environment where there is background noise or a sound unrelated to the user's operation. .

  In order to solve the above-described problem, the first aspect of the present invention is a background noise removing unit that removes background noise from a sound signal from a microphone installed in an environment where a user can exist, and the sound after the background noise removal. A sound pressure extraction unit that extracts a sound pressure from a band of a predetermined frequency or less of a signal, and an operation sound detection unit that detects an operation sound by the user based on the extracted sound pressure. .

  For example, the frequency is 100 Hz. For example, the sound pressure extraction unit extracts a sound pressure from a band higher than a predetermined frequency, and the operation sound detection unit detects an operation sound by the user based on the sound pressure. For example, the high frequency is 15 kHz. For example, the background noise removing unit inverts the phase of the audio signal from one of the two microphones and synthesizes the audio signal from the other microphone.

  According to a second aspect of the present invention, the background noise removal unit of the operation sound detection device removes background noise from a sound signal from a microphone installed in an environment where the user may exist, and the sound pressure extraction unit of the operation sound detection device Extracts a sound pressure from a band below a predetermined frequency of the audio signal after the background noise is removed, and the operation sound detection unit of the operation sound detection device operates by the user based on the extracted sound pressure. It is characterized by detecting sound.

  For example, the sound pressure extraction unit extracts a sound pressure from a band higher than a predetermined frequency, and the operation sound detection unit detects an operation sound by the user based on the sound pressure.

  ADVANTAGE OF THE INVENTION According to this invention, the fine operation sound which a user emits can be detected easily and with high precision also in the environment with background noise or a sound irrelevant to a user's operation | movement.

It is a block diagram which shows schematic structure of the operation | movement sound detection apparatus which concerns on this Embodiment. It is a figure for demonstrating the effect | action by the wave of the sound which reaches | attains the microphone 1A and the microphone 1B, and the microphone 1A and the microphone 1B. It is a figure which shows the waveform A of the audio | voice signal from the microphone 1A, the waveform B of the audio | voice signal from the microphone 1B, and the waveform C after a phase inversion and a synthesis | combination. It is a figure which shows a part of frequency spectrum obtained by the frequency analysis part 5. FIG. It is a figure which shows the kind of sound contained in a low frequency band, an audible band, and a high frequency band. It is a figure which shows typically the filtering by an operation | movement sound detection apparatus.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 is a block diagram showing a schematic configuration of an operation sound detection device according to the present embodiment.

  The operation sound detection device includes, for example, two adjacent microphones 1A and 1B installed in an environment where a user such as a person may exist, a microphone amplifier 2A that amplifies an audio signal output from the microphone 1A, and a microphone 1B. Is used together with a microphone amplifier 2B for amplifying an audio signal output from the.

  The operating sound detection device includes a background noise removal unit 3 that removes background noise by inverting and amplifying the sound signal output from the microphone amplifier 2A and synthesizing the sound signal with the sound signal output from the microphone amplifier 2B. A / D converter 4 that digitizes the audio signal (analog signal) after removal, and frequency analysis of the digitized signal, and the entire frequency band is an audible band (for example, 100 Hz to 15 kHz) and a band lower than the audible band (For example, a band of 100 Hz or less, hereinafter referred to as a low frequency band) and a frequency analysis unit 5 that divides the band higher than the audible band (for example, a band of 15 kHz or more, hereinafter referred to as a high frequency band), and a low frequency A sound pressure extraction unit 6 that extracts a sound pressure from the band and the high frequency band, and detects an operation sound by the user based on the extracted sound pressure Comprising a Sakuon detector 7.

  The noise cancellation unit 3 includes a phase inversion unit 31 that inverts the phase of the audio signal from the microphone amplifier 2A, and a variable amplifier 32 that amplifies the audio signal after the phase inversion.

In the operation sound detection device, first, the background noise removing unit 3 removes background noise.
As shown in FIG. 2, for example, the distance between the microphone 1A and the microphone 1B is 3 cm. The background noise from the air conditioner and the outside reaches the microphone 1A and the microphone 1B as a plane wave, so that the phase difference on the same plane is small, and the sound waves having the same phase reach the microphone 1A and the microphone 1B. Therefore, the background noise removal unit 3 can remove the background noise by inverting the sound signal from the microphone 1A and synthesizing it with the sound signal from the microphone 1B.

  On the other hand, for example, when a person or the like moves around the microphone 1A and the microphone 1B to generate a minute operation sound, the operation sound reaches the microphone 1A and the microphone 1B as a spherical wave.

  Further, sound emitted from a speaker such as a television also reaches the microphone 1A and the microphone 1B as a spherical wave.

  Since the phase difference of the spherical wave is not uniform unless it is spherical, a sound wave having a phase difference reaches the microphone 1A and the microphone 1B. Therefore, even when the background noise removing unit 3 inverts the sound signal from the microphone 1A and synthesizes it with the sound signal from the microphone 1B, the operation sound and the sound emitted from the speaker remain, that is, the sound emitted from the speaker thereafter. If it is removed, only the operation sound can be left.

  A waveform A in FIG. 3 is an example of a waveform of an audio signal from the microphone 1A, and shows a relationship between elapsed time and amplitude level. A waveform B is an example of a waveform of an audio signal from the microphone 1B, and shows a relationship between elapsed time and amplitude level.

  Waveforms A and B both include a waveform M due to background noise and are in phase. On the other hand, the waveforms A and B both include the waveform N due to the operating sound, but the phases are not aligned and there is a phase difference. Therefore, in the waveform C after the phase inversion and synthesis, the waveform M is removed, and the waveform C is based on the waveform N.

  Next, the frequency analysis unit 5 divides the entire frequency band into an audible band (for example, 100 Hz to 15 kHz), a low frequency band lower than the audible band, and a high frequency band higher than the audible band.

FIG. 4 is a diagram showing a part of the frequency spectrum obtained by the frequency analysis unit 5.
In a low frequency band of 100 Hz or less, a sound pressure waveform of an operation sound (such as a footstep sound) by a user is included, and a sound pressure waveform emitted from a speaker such as a television is not seen.

  On the other hand, in the audible band exceeding 100 Hz, the operation sound by the user and the sound pressure waveform of the sound emitted from a speaker such as a television are mixed.

  Next, the sound pressure extraction unit 6 extracts the sound pressure from the low frequency band and the high frequency band, not from the audible band where the sound pressure waveforms are mixed.

  As shown in FIG. 5, the frequency of sound emitted from a speaker of a device, such as television sound or telephone sound, is included in an audible band that is easy to hear mainly by humans.

  On the other hand, the low frequency band includes, for example, frequencies such as footsteps and object sounds that are operation sounds by the user. The low frequency band also includes background noise frequencies such as air conditioning. In a low frequency band, the reproduction performance of a general speaker is reduced.

  Further, the high frequency band includes, for example, the frequency of the sound of water flowing in the kitchen and the sound of tableware, which is the operation sound of the user.

  That is, since the sound in the low frequency band and the high frequency band is likely to be an operation sound by the user, the sound pressure extraction unit 6 extracts the sound pressure from the low frequency band and the high frequency band, not from the audible band. .

  Finally, the operation sound detection unit 7 detects an operation sound by the user based on the extracted sound pressure. For example, if the sound pressure extracted from at least one of the low frequency band and the high frequency band is greater than or equal to a predetermined threshold, the operation sound detection unit 7 determines that there is a user in the surroundings, and both sound pressures are If it is less than the threshold, it is determined that there is no user around.

  The operation sound detection unit 7 determines that there is a user in the surroundings if the sound pressure extracted from the low frequency band is equal to or higher than a predetermined threshold, and if the sound pressure is less than the threshold, there is a user in the surroundings. You may decide not to.

  Therefore, as shown in FIG. 6, in the operation sound detection device, among the background noise, the sound emitted from the speaker, and the operation sound by the user, the sound emitted from the speaker and the operation sound by the user pass through the background noise removal unit 3. Only.

  The frequency analysis unit 5 divides the entire frequency region into an audible band including a sound emitted from a speaker, and a low frequency band and a high frequency band including a sound of an operation sound by a user.

  Of the sound pressure of the sound emitted from the speaker included in the audible band and the sound pressure of the operation sound by the user included in the low frequency band and the high frequency band, the sound pressure extraction unit 6 extracts the low frequency band and the high frequency. It is only the sound pressure of the operation sound by the user included in the band. As a result, the operation sound detection unit 7 can detect only the operation sound by the user.

  Therefore, according to the operation sound detection device, the background noise removal unit 3 that removes background noise from the sound signals from the microphones 1A and 1B installed in an environment where the user can exist, and the sound signal after the background noise removal in advance. By providing a sound pressure extraction unit 6 that extracts a sound pressure from a low frequency band of a defined frequency of 100 Hz or less and an operation sound detection unit 7 that detects an operation sound by the user based on the extracted sound pressure, It is possible to easily and accurately detect a slight operation sound generated even in an environment where there is a background noise or a sound such as a speaker unrelated to the user's operation.

  The sound pressure extraction unit 6 extracts sound pressure from a high frequency band of 15 kHz or higher, which is higher than the frequency of 100 Hz, and the operation sound detection unit 7 detects the operation sound by the user based on the sound pressure. The operation sound can be detected with high accuracy.

  The background noise removing unit 3 can remove the background noise having the same phase by inverting the phase of the sound signal from the microphone 1A and synthesizing it with the sound signal from the microphone 1B. Operation sound can be detected under the environment.

  The frequencies 100 Hz and 15 kHz are examples, and other frequencies may be used depending on the environment. Also, the sound pressure threshold is preferably set to a value according to the environment.

  Among the operation sound detection devices, the A / D conversion unit 4, the frequency analysis unit 5, the sound pressure extraction unit 6, and the operation sound detection unit 7 can be configured by a computer, and a computer program for causing the computer to function is as follows. It can be recorded on a computer-readable recording medium such as a semiconductor memory, a magnetic disk, an optical disk, a magneto-optical disk, or a magnetic tape, and can be widely distributed by being transmitted via a communication network such as the Internet.

1A, 1B microphone 2A, 2B microphone amplifier 3 noise canceling unit 3 background noise removing unit 4 A / D conversion unit 5 frequency analyzing unit 6 sound pressure extracting unit 7 operation sound detecting unit 31 phase inversion unit 32 variable amplifier

Claims (8)

A background noise removing unit that removes background noise from an audio signal from a microphone installed in an environment where a user may exist;
A sound pressure extraction unit that extracts a sound pressure from a band of a predetermined frequency or less of the audio signal after the background noise removal;
An operation sound detection unit comprising: an operation sound detection unit that detects an operation sound by the user based on the extracted sound pressure. The operating sound detection device according to claim 1, wherein the frequency is 100 Hz. The sound pressure extraction unit extracts sound pressure from a band of a frequency higher than the predetermined frequency,
The operation sound detection device according to claim 1, wherein the operation sound detection unit detects an operation sound by the user based on the sound pressure. The operation sound detection device according to claim 3, wherein the high frequency is 15 kHz. The said background noise removal part carries out the phase inversion of the audio | voice signal from one microphone of two said microphones, and synthesize | combines with the audio | voice signal from the other said microphone. Operation sound detection device. The background noise removal unit of the operation sound detection device removes background noise from the sound signal from the microphone installed in the environment where the user can exist,
The sound pressure extraction unit of the operation sound detection device extracts sound pressure from a band below a predetermined frequency of the sound signal after the background noise removal,
The operation sound detection unit of the operation sound detection device detects an operation sound by the user based on the extracted sound pressure. The operation method of the operation sound detection device. The sound pressure extraction unit extracts sound pressure from a band of a frequency higher than the predetermined frequency,
The operation method of the operation sound detection device according to claim 6, wherein the operation sound detection unit detects an operation sound by the user based on the sound pressure.   A computer program for causing a computer to function as the operation sound detection device according to claim 1.

Images