JP2008301236A - Audio signal processing apparatus, method, program, and computer-readable recording medium storing the program - Google Patents

Abstract

For example, sudden noise generated when an object collides with a microphone or when a person directly takes a breath can be reliably removed or greatly reduced without requiring complicated processing.
An output control unit (42a) for generating and outputting one output audio signal based on three or more audio signals input by three or more input units (10, 20, 30), and three or more audio signals. In comparison, when a difference occurs between three or more audio signals, a noise detection unit 41a that detects a difference between the minority audio signal and the majority audio signal as noise by majority vote between the three or more audio signals is provided. When the noise is detected by the noise detection unit 41a, the output control unit 42a reduces or eliminates the contribution rate of the audio signal including the noise to the generation of the output audio signal.
[Selection] Figure 1

Description

  The present invention relates to a technique for converting voice such as a conversation into an electric signal using a microphone, for example, and transmitting the signal to voice, recording, or a remote place.

Traditionally, when a single microphone is used to convert sound into an electrical signal, for example, when a speaker hits the microphone or when a breath is directly applied to the microphone, it is an undesirable overload that is not part of the original conversation. Sudden noise is also converted into an electric signal, and as a result, excessive noise is included in the electric signal (audio signal) as it is.
At this time, if technology such as ALC (Automatic Level Control) is used, it is possible to automatically squeeze the volume of the part with excessive noise and reduce unpleasantness. In the narrowed part, the volume of the original signal (conversation) is also reduced, and it may be difficult to hear the conversation.

In addition, when using a mixture of electrical signals from a plurality of microphones, the audio signal from the microphone physically hit by the speaker is alleviated to a fraction of the number of microphones, but the noise is eliminated. However, it was unavoidable that noise entered the audio signal.
Conventionally, a technique for enhancing a target sound component by adjusting a combined sound signal of a plurality of sound signals in accordance with a cross-correlation coefficient of sound signals of a plurality of microphones (for example, Patent Document 1 below) or provided in a vehicle There is a technique (for example, Patent Document 2 below) that constantly monitors the ambient noise situation for each of a plurality of microphones and selects a microphone to be used according to the detected ambient noise situation.
JP 2004-289762 A JP 2006-33700 A

However, the technique of Patent Document 1 requires very complicated processing such as calculation of cross-correlation coefficients and component adjustment.
Further, since the technique of Patent Document 2 is intended for a stationary ambient noise situation, for example, a noise according to a vehicle engine sound, an air conditioner blower sound, or a vehicle window opening / closing situation, for example, a microphone. Sudden noise that occurs when an object (such as a speaker) collides with a person or when a person directly takes a breath cannot be removed.

  The present invention has been devised in view of such problems. For example, sudden noise that occurs when an object collides with a microphone or when a person directly takes a breath can be reliably obtained without requiring complicated processing. It is intended to be able to be removed or greatly reduced.

  In order to achieve the above object, an audio signal processing device according to the present invention is based on three or more input units for inputting an audio signal, and three or more audio signals input by the three or more input units. When an output control unit that generates and outputs an output audio signal is compared with the three or more audio signals input by the three or more input units, and there is a difference between the three or more audio signals, A noise detection unit that detects, as noise, a difference between a minority audio signal and a majority audio signal by a majority vote between the above audio signals, and the output control unit is configured to detect when noise is detected by the noise detection unit. Is characterized in that the contribution rate of the audio signal including noise to the generation of the output audio signal is reduced or made zero (claim 1).

  In order to achieve the above object, an audio signal processing device according to the present invention has a plurality of input units for inputting audio signals and one output based on the plurality of audio signals input by the plurality of input units. An output control unit that generates and outputs an audio signal; and a noise detection unit that detects, as noise, a portion of the plurality of audio signals that has a signal level exceeding a threshold value. When noise is detected, the contribution ratio of the audio signal including noise to the generation of the output audio signal is reduced or made zero (claim 2).

The output control unit generates and outputs the output audio signal by combining the audio signals at a predetermined ratio unless noise is detected by the noise detection unit, while the noise detection unit generates noise. When detected, it is preferable to reduce the proportion of the audio signal in which the noise is detected.
The output control unit outputs an average value of the plurality of audio signals as the output audio signal unless noise is detected by the noise detection unit, and when noise is detected by the noise detection unit, It is preferable to output an average value of audio signals other than the audio signal in which noise is detected as the output audio signal.

Furthermore, a holding unit that temporarily holds the audio signal input by the input unit is provided, and the output control unit generates the output audio signal based on the audio signal held in the holding unit. It is preferable to output (Claim 5).
In order to achieve the above object, the audio signal processing method of the present invention includes an input step of inputting three or more audio signals, and one output based on the three or more audio signals input in the input step. When an output control step for generating and outputting an audio signal is compared with three or more audio signals input in the input step, and there is a difference between the three or more audio signals, the three or more audio signals are A noise detecting step of detecting, as noise, a difference portion of the minority audio signal with respect to the majority audio signal by the majority vote in the method, and when the noise is detected in the noise detection step, The contribution ratio of the audio signal including the output signal to the generation of the output audio signal is reduced or made zero (claim 6).

  In order to achieve the above object, the audio signal processing program of the present invention causes a computer to realize a function of outputting one output audio signal based on three or more audio signals input from three or more input units. An output control unit for generating and outputting the one output audio signal based on the three or more audio signals input by the three or more input units, and three or more input units When the three or more audio signals input by the above are compared, and a difference occurs between the three or more audio signals, the majority audio signal of the minority audio signal with respect to the majority audio signal is determined by a majority vote between the three or more audio signals. The computer functions as a noise detection unit that detects a difference portion as noise, and noise is included when the output control unit detects noise by the noise detection unit. So as to reduce or zero the contribution ratio with respect to generation of the output audio signal that the audio signal is characterized in that to function the computer (claim 7).

  In order to achieve the above object, the computer-readable recording medium of the present invention records the above-described audio signal processing program (claim 8).

  As described above, according to the present invention, for example, the noise detection unit performs complicated processing on sudden noise generated when an object (for example, a speaker) collides with the microphone or when the speaker's breath blows directly and strongly. In addition, the output control unit can reliably remove or significantly reduce the noise from the output audio signal (claims 1 to 8).

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
[1] First Embodiment of the Present Invention First, the configuration of an audio signal processing apparatus (hereinafter referred to as the present audio signal processing apparatus) 1a as a first embodiment of the present invention with reference to the block diagram shown in FIG. Will be described. The audio signal processing device 1a has three or more (more preferably three or more odd numbers; here three) input units 10, 20, and 30 (hereinafter referred to as “10-30” unless otherwise distinguished). And a DSP (Digital Signal Processor) 40a.

  The input units 10 to 30 input external audio as audio signals, and are microphones 11, 21, and 31 (hereinafter, indicated by “11 to 31” unless otherwise distinguished). , Microphone amplifiers 12, 22, and 32 (hereinafter referred to as “12 to 32” unless otherwise distinguished) and ADCs (Analog Digital Converters) 13, 23, and 33 (hereinafter referred to as “analog digital converter”). If there is no particular distinction, the reference numerals “13 to 33” are used).

In the following description, a voice mainly means a voice uttered by a speaker (user). However, in the present invention, a voice is not limited to a voice uttered by a speaker, and is emitted from a musical instrument, for example. Any sound such as sound may be used.
Here, an application example of the audio signal processing apparatus 1a to the mobile phone 2 is shown in FIG. In this case, the microphones 11 to 31 of the input units 10 to 30 are provided with a predetermined interval therebetween. The audio signal processing device 1a uses, for example, the microphone 11 (input unit 10) in a normal call, and uses the microphones 11 to 31 (input units 10 to 30) when performing a videophone call or a hands-free call. Configured to do. The microphone to be used is switched by the DSP 40a.

In FIG. 2, reference numeral “3” indicates an operation button of the mobile phone 2, and reference numeral “4” indicates a display unit of the mobile phone 2.
FIG. 3 shows an example in which the microphones 11 to 31 of the audio signal processing apparatus 1 a are mounted on the headset 5. Also in this case, the microphones 11 to 31 are provided at a predetermined interval. In FIG. 3, reference numeral “6” indicates a headphone portion of the headset 5.

The microphones 11 to 31 convert the voice uttered by the speaker into an electrical signal and output it.
The microphone amplifiers 12 to 32 amplify the electrical signal applied to the sound output from the corresponding microphones 11 to 31 and output the amplified electrical signals to the ADCs 13 to 33 in the subsequent stage.
The ADCs 13 to 33 convert the electrical signals (analog signals) output from the microphone amplifiers 12 to 32 into digital signals (PCM: Pulse Code Modulation), and input them as audio signals to the subsequent DSP 40a.

  The DSP 40a performs processing on audio signals input from the plurality of input units 10 to 30, and includes a noise detection unit 41a and an output control unit 42a. The noise detection unit 41a and the output control unit 42a are realized by the DSP 40a executing a predetermined application program (for example, an audio signal processing program described later).

The noise detection unit 41a compares the three audio signals input by the input units 10 to 30, and when a difference occurs between the three audio signals, the majority of the minority audio signals are determined by a majority vote between the three audio signals. The difference portion with respect to the voice signal of the group is detected as noise.
Specifically, as shown in FIGS. 4A to 4C, the noise detection unit 41a monitors the audio waveforms of the three input audio signals in real time, and a predetermined value or more between these three audio signals. When the difference is generated, noise detection processing is performed.

  4A shows the speech waveform of the speech signal input from the input unit 10, FIG. 4B shows the speech waveform of the speech signal input from the input unit 20, and FIG. The voice waveform of the voice signal input from the input unit 30 is shown. 4A to 4C and FIGS. 5 and 8A and 8B described later, the horizontal axis indicates time, and the vertical axis indicates voice waveform level (for example, voltage value). .

  The noise detection unit 41a compares the audio waveform levels of the three audio signals in real time, and as a result, if a difference of a predetermined value or more occurs in the audio waveform levels of the three audio signals, the noise detection unit 41a generates a difference between the three audio signals. Is determined to have occurred. Therefore, the noise detection unit 41a does not determine that a difference has occurred between the audio signals as soon as the audio waveform levels between the audio signals are not exactly the same, but the difference in the audio waveform levels between the plurality of audio signals is not. If it is within the predetermined value, it is determined that the plurality of audio signals are at the same level. On the other hand, if there is a difference between the audio signals that is recognized as noise, a difference occurs between the audio signals. It is comprised so that it may be judged.

  For example, as shown in FIGS. 4A to 4C, the noise detection unit 41a determines that a difference has occurred between the audio signals from the time when three audio signals are input at time t0 to the time t1. However, when noise suddenly occurs in the audio signal in FIG. 4B at time t1 and a difference of a predetermined value or more occurs between other audio signals, the noise detection unit 41a detects the difference between the audio signals. It is judged that a difference has occurred.

  When the noise detection unit 41a determines that a difference has occurred between the audio signals, the three audio signals are grouped into groups having the same audio waveform level according to the audio waveform level according to the majority logic. A difference between the minority group (here, the audio signal from the input unit 20) and the majority group (here, the audio signal from the input units 10 and 30) is detected as noise.

That is, the noise detection unit 41a specifies an abnormal audio signal including noise by majority decision and specifies the noise.
As described above, since the microphones 11 to 31 are arranged at a predetermined interval (see FIGS. 2 and 3), for example, an object (speaker or the like) hits one of the microphones 11 to 31. In the case where the air is blown or the air is blown directly, noise is generated only in one of the microphones 11 to 31. In other words, if there is a distance between the microphones 11 to 31, for example, when an object collides, the audio signal from the minority microphones 11 to 31 that the object collided with and the other majority microphones 11 to 31. A clear difference appears between the voice signal and the voice signal from, and this difference appears in the minority voice signal (waveform) as a noise waveform, as shown in FIGS. 4 (a) and 4 (b).

Therefore, the noise detection unit 41a can specify the difference between the minority voice signal (waveform) and the majority voice signal (waveform) as noise by majority vote.
Furthermore, as shown at time t2 in FIG. 4 (b), the noise detection unit 41a determines the voice waveform level of the voice signal from the input unit 20 determined to be an abnormal voice signal including noise, and the majority input. When the difference between the sound waveform levels of the sound signals from the units 10 and 30 disappears continuously for a predetermined time, it is determined that the noise of the sound signal from the input unit 20 has disappeared.

As described above, the noise detection unit 41a has a difference (t1) between the audio signal from the input unit 20 in FIG. 4B and the audio signal from the input units 10 and 30 in FIGS. 4A and 4C. ˜t2) is detected as noise.
Similarly, the noise detection unit 41a has a predetermined value or more between the audio signal from the input unit 10 in FIG. 4A and the audio signals from the input units 20 and 30 in FIGS. 4B and 4C. Part of the difference in the audio signal from the input unit 10 from the time point when the difference occurs (time t3 in FIG. 4A) until the difference disappears continuously for a predetermined time (time t4 in FIG. 4A) Is detected as noise.

  The output control unit 42a generates one output audio signal based on the three audio signals input from the input units 10 to 30 and outputs the same to the outside, and noise must be detected by the noise detection unit 41a. For example, the audio signals from the input units 10 to 30 are combined (mixed) at a predetermined ratio (for example, the audio signal of the input unit 10 is 50% and the audio signals of the input units 20 and 30 are 25%). An output audio signal is generated, or an average value of three audio signals is output as an output audio signal.

  On the other hand, when the noise is detected by the noise detection unit 41a, the output control unit 42a reduces or eliminates the ratio of the audio signal in which the noise is detected (contribution rate to the generation of the output audio signal). An output audio signal with reduced or eliminated is generated and output. For example, when the output audio signal is generated by combining the audio signals at a predetermined ratio, the noise is detected as much as possible in the output audio signal by reducing or reducing the combination ratio of the audio signals in which noise is detected. Do not include.

  Therefore, in the output control unit 42a, in the case shown in FIGS. 4A to 4C, the noise detection unit 41a includes noise in the audio signal from the input unit 20 in FIG. 4B. During the time t1 to t2 when the signal is detected, the coupling ratio of the audio signal from the input unit 20 to the output audio signal is changed from 25% to 0%, and the audio signal from the input units 10 and 30 is combined with the output audio signal. Both percentages are changed to 50%. Furthermore, during the time t3 to t4 when the noise detection unit 41a detects that the audio signal in FIG. 4A includes noise, the coupling ratio of the audio signal from the input unit 10 to the output audio signal is changed. 50% is changed to 0%, and the coupling ratio of the audio signal from the input units 10 and 30 to the output audio signal is changed to 50%.

Further, when the average value of the three audio signals is output as the output audio signal, the output control unit 42a uses the average value of other audio signals that do not include the audio signal from which noise is detected as the audio output signal. Output.
In the cases shown in FIGS. 4A to 4C, the noise detection unit 41a detects that the audio signal from the input unit 20 in FIG. 4B includes noise for the time t1 to t2. Then, the output control unit 42a does not include the audio signal from the input unit 20 during that period, and outputs the average value of the audio signals from the input units 10 and 30 as the output audio signal. Similarly, when the noise detection unit 41a detects that the audio signal from the input unit 10 contains noise during the time t3 to t4, the output control unit 42a receives from the input unit 10 during that time. The average value of audio signals from the input units 20 and 30 other than the audio signal is output to the outside as an output audio signal.

Thereby, the output control part 42a outputs the output audio | voice signal which does not contain noise, as shown in FIG.
Here, the operation procedure of the audio signal processing device 1a (that is, the audio signal processing method of the present invention) will be described with reference to the flowchart (steps S1 to S5) shown in FIG. When each signal is input (step S1; input step), the noise detection unit 41a of the DSP 40a compares the three audio signals input by the input units 10 to 30 as described above, and a difference occurs between the three audio signals. It is determined whether or not (step S2).

When the noise detection unit 41a determines that there is no difference between the three audio signals (No route in step S2), the output control unit 42a performs a predetermined operation on the three audio signals as a normal operation (output control). The ratio combination or average value is output to the outside as an output audio signal (step S3; output control step), and the process ends.
On the other hand, when the noise detection unit 41a determines that a difference has occurred between the three audio signals (Yes route in step S2), the difference between the minority audio signal and the majority audio signal is determined by the majority vote between the three audio signals. (Step S4; noise detection step).

Next, the output control unit 42a generates an output audio signal by reducing or eliminating the contribution ratio of the audio signal including the detected noise to the generation of the output audio signal as output control at the time of the noise detection unit. Output to the outside (step S5; output control step), and the process ends.
As described above, according to the audio signal processing device 1a (audio signal processing method) as the first embodiment of the present invention, the noise detection unit 41a detects noise by the majority vote between the audio signals from the input units 10 to 30. (Noise detection step) When the output control unit 42a detects noise by the noise detection unit 41a, the contribution rate of the output audio signal of the audio signal including the noise to the generation of the output audio signal is reduced or zero ( Output control step), for example, the noise detection unit 41a performs complicated processing for sudden noise generated when an object (for example, a speaker) collides with the microphones 11 to 31 or when the speaker's breath is blown directly and strongly. The output control unit 42a can reliably remove or significantly reduce the noise from the output audio signal.

Of course, the noise that can be removed or reduced by the audio signal processing apparatus 1a is not only noise caused when an object collides with the microphones 11 to 31 or when breath is blown strongly, but also spike noise caused by discharge or a network line. The same effect can be obtained even when electrical noise due to abnormalities is added.
Furthermore, since the noise detection unit 41a detects noise using the majority logic, not only a loud sound but also a sound with a small sound volume is input due to, for example, a failure of the microphones 11 to 31. Even in such a case, the noise of the small volume can be detected as noise, and the audio signal from the failed microphones 11 to 31 can be prevented from being included in the output audio signal.

In addition, when the noise is detected by the noise detection unit 41a, the output control unit 42a normally combines the audio signals at a predetermined ratio, and the combination ratio of the audio signals from which the noise is detected Therefore, the noise can be surely removed or greatly reduced from the output audio signal.
Further, when the output control unit 42a normally outputs an average value of three audio signals as an output audio signal, when the noise is detected by the noise detection unit 41a, the audio from which the noise is detected is output. Since the average value of the audio signals other than the signal is output as the output audio signal, the noise can be reliably removed from the output audio signal.

[2] Second Embodiment of the Present Invention Next, an audio signal processing device 1b as a second embodiment of the present invention will be described with reference to the block diagram shown in FIG. In FIG. 7, the same reference numerals as those described above indicate the same or substantially the same parts.
The audio signal processing apparatus 1b is configured to include a plurality (here, two) of input units 10 and 20 and a DSP 40b.

Here, the input units 10 and 20 are the same as those in the first embodiment described above, and are configured to include microphones 11 and 21, microphone amplifiers 12 and 22, and ADCs 13 and 23, respectively.
For example, an audio signal as shown in FIG. 8A is input from the input unit 10, and an audio signal as shown in FIG. 8B is input from the input unit 20.

The DSP 40b performs processing on audio signals input from the plurality of input units 10 and 20, and includes a noise detection unit 41b and an output control unit 42b.
The noise detection unit 41b detects a portion where the signal level exceeds a threshold in a plurality of audio signals as noise. For example, as shown in FIGS. 8A and 8B, the audio waveform level of the input signal is detected. Is detected as noise when the threshold value is less than or equal to a preset threshold value s1 or greater than or equal to the threshold value s2.

  Specifically, as shown in FIG. 8B, the noise detection unit 41b, when the audio waveform level of the audio signal from the input unit 20 becomes equal to or less than the threshold s1 (see time t5), the audio signal includes noise. The audio signal from the input unit 20 contains noise until the audio waveform level is continuously higher than the threshold value s1 and lower than the threshold value s2 for a predetermined time (Δt) thereafter (see time t7). Judge.

  That is, the noise detection unit 41b detects noise at time t5 and then continues until a time t7 when the speech waveform level is greater than the threshold value s1 and smaller than the threshold value s2 for a predetermined time Δt (that is, the speech waveform level is at time t6). It is determined that the audio signal from the input unit 20 contains noise after the predetermined time Δt has elapsed from the time when the value exceeds the threshold s1 to the time t7.

In FIG. 8B, the speech waveform level is larger than the threshold value s1 and smaller than the threshold value s2 until the time t6 after the time t5. However, since the time is shorter than the predetermined time Δt, noise is used here. The detection unit 41b does not determine that the noise has been eliminated.
Similarly, as shown in FIG. 8A, the noise detection unit 41b detects noise from time t8 to time t9 in the audio signal from the input unit 10.

When noise is detected by the noise detection unit 41b, the output control unit 42b reduces or reduces the contribution ratio of the audio signal including the noise to the generation of the output audio signal (here, zero). Thus, the volume ratio of the audio signal on the side containing noise included in the output audio signal is reduced, and the noise in the output audio signal is reduced or eliminated.
Specifically, the output control unit 42b outputs an average value of the audio signals of the input units 10 and 20 as an output audio signal during normal times (that is, when no noise is detected by the noise detection unit 41b), When noise is detected by the noise detection unit 41b, the contribution ratio of the audio signal including noise is set to zero, and the audio signal not including noise is output as an output audio signal.

  Here, when the noise is detected by the noise detection unit 41b as described above, the output control unit 42b switches the output audio signal to an audio signal not including the noise, as shown in FIG. An adjustment signal (panpot signal) for adjusting (panpotting) the audio signal to be included in the audio signal is generated by proportional distribution, and output control of the audio signal is performed based on the adjustment signal. In FIG. 9, the horizontal axis indicates time, and the vertical axis indicates the value (adjustment value) of the adjustment signal for the pan pot.

In the example shown in FIG. 9, when the adjustment value is “0”, the output control unit 42b outputs only the audio signal from the input unit 20 as the output audio signal, and when the adjustment value is “0.5”. The audio signals from the input units 10 and 20 are combined at the same ratio and output as an output audio signal. When the adjustment value is “1”, only the audio signal from the input unit 10 is output as the output audio signal.
Then, the output control unit 42b adjusts a signal to output the audio signal from the input unit 10 as an output audio signal during the time t5 to t7 when the noise detection unit 41b detects the noise of the audio signal from the input unit 20. Is set to “1” to perform output control.

Further, the output control unit 42b is an adjustment signal for outputting the audio signal from the input unit 20 as an output audio signal during the time t8 to t9 when the noise detection unit 41b detects the noise of the audio signal from the input unit 10. Is set to “0” to perform output control. As a result, the output control unit 42b outputs an output audio signal with reduced noise as shown in FIG.
The output control unit 42b gradually adjusts the adjustment value “1” or “within a predetermined time without changing the adjustment signal from the adjustment value“ 0.5 ”to the adjustment value“ 1 ”or“ 0 ”. This is changed to “0”. This is because when the output audio signal is directly switched from the state in which the audio signals of the input units 10 and 20 are combined in half to any one of the audio signals, this switching itself causes noise generation. Because it becomes.

  As described above, according to the audio signal processing device 1b as the second embodiment of the present invention, the noise detection unit 41b performs noise detection on a portion of the plurality of audio signals from the input units 10 and 20 where the signal level exceeds the threshold value. When the noise is detected by the noise detection unit 41b, the output control unit 42b zeroes (or reduces) the contribution rate of the audio signal including the noise to the generation of the output audio signal. For example, the microphone 11 , 21 can be detected reliably without the need for complicated processing, and the noise detection unit 41a can detect sudden noise generated when an object collides with the speaker 21 or when the speaker's breath blows directly. 42b can reliably remove or significantly reduce the noise from the output audio signal.

[3] Others The present invention is not limited to the above-described embodiments, and various modifications can be made without departing from the spirit of the present invention.
For example, in the first embodiment described above, the example in which the audio signal processing device 1a is configured to include the input units 10 to 30 and the DSP 40a has been described. However, the present invention is not limited to this, for example, As shown in FIG. 10, a DAC (Digital Analog Converter) 50 may be provided. The DAC 50 converts an output audio signal output from the DSP 40c into an analog signal and outputs the analog signal to the outside. May be.

FIG. 10 is a block diagram showing a configuration of an audio signal processing apparatus 1c as a modification of the present invention. In FIG. 10, the same reference numerals as those already described indicate the same or substantially the same parts. .
In the first embodiment described above, the case where the audio signal input from the input units 10 to 30 is processed by the DSP 40a in real time and output to the outside has been described as an example. However, the present invention is not limited to this. If the output audio signal to be output is allowed to have a slight delay, the DSP 40c holds the audio signal input from the input units 10 to 30 as shown in the audio signal processing device 1c in FIG. When the noise is detected by the noise detection unit 41a, the output control unit 42a uses the audio signal held in the holding unit 43a slightly before the detection timing (predetermined time) as an audio signal. You may comprise so that it may mix.

  That is, as illustrated in FIG. 4B, when the noise detection unit 41 a detects noise in the audio signal from the input unit 20 during the time t <b> 1 to t <b> 2, the output control unit 42 a is held by the holding unit 20. When using the audio signal, the audio signal from the input unit 20 is configured to generate the output audio signal so that the audio signal from the input unit 20 is not included in the output audio signal slightly before the time t1. It is desirable. Thereby, after the noise detection unit 41a detects the noise, it is ensured that the output audio signal includes the noise in a short time until the output control unit 42a performs the output control so as not to include the noise. Therefore, it is possible to more reliably realize noise removal from the output audio signal.

  Note that this modification can naturally be applied to the second embodiment described above, and in this case, the noise detection unit 41b performs time in an audio signal from the input unit 20 as shown in FIG. 8B, for example. When noise generation is detected at t5, the output control unit 42b uses the audio signal held in the holding unit to convert the audio signal from the input unit 20 into the output audio signal from a predetermined time before the time t5. Control so that it is not included. As a result, the noise from when the noise is generated until the noise is detected by the noise detection unit 41b, and the output control unit 42b outputs the output audio signal using the adjustment signal (panpot signal; see FIG. 9). The noise contained in the output audio signal can be more reliably removed from the output audio signal until it is completely switched to the audio signal from the input unit 10.

  Furthermore, in the first embodiment described above, since the noise detection unit 41a detects noise based on the majority logic, it detects not only high-volume noise but also low-volume noise that is not included in other input audio signals. However, as shown in FIG. 10, for example, the noise detection unit 41a of the DSP 40c makes noise from a specific audio signal steadily (that is, continuously for a predetermined time or more). When the signal is detected, the DSP 40c may be configured to include an abnormality detection unit 44a that performs abnormality detection by determining that there is an abnormality in the input units 10 to 30 related to the specific audio signal. When the detection unit 44a detects an abnormality in the input units 10 to 30, the DSP 40c may be configured to include warning means 45 for warning the outside (user) to that effect.

  The warning unit 45 is realized by, for example, an LED (Light Emitting Diode), and is installed in the vicinity of each of the microphones 11 to 31. And if the abnormality detection part 44a detects abnormality in the input parts 10-30, it will light the LED corresponding to the microphones 11-31 of the input part 10-30 which detected abnormality, and will make a speaker the microphones 11-31 ( Warning of abnormalities in the input units 10 to 30).

Thus, the speaker can know the microphones 11 to 31 that are not operating in a complete state, and in some cases, preventive maintenance can be performed before the input units 10 to 30 completely fail.
In the above-described embodiment, the number of input units 10, 20, and 30 is not limited.
Furthermore, in the embodiment described above, the number of microphones 11 to 31 included in one input unit 10 to 30 is not limited. That is, in the above-described embodiment, an example in which each input unit 10-30 is configured to include one microphone 11-31, that is, a case where the input audio signal is monaural audio has been described as an example. The present invention is not limited to this. For example, each of the input units 10 to 30 may include a plurality of microphones 11 to 31 and the input sound may be stereo sound (multi-channel sound).

  In the above-described embodiment, the case where the output control units 42a and 42b generate the output audio signal by mixing the input audio signals from the plurality of input units 10 to 30 or the input units 10 and 20 will be described as an example. However, the present invention is not limited to this, and the output control units 42a and 42b may be configured to output one of the plurality of input audio signals as an output audio signal. In this case, when noise is detected by the noise detection units 41a and 41b, the output control units 42a and 42b output an input audio signal other than the input audio signal from which the noise has been detected to the outside as an output audio signal. Thereby, the effect similar to embodiment mentioned above can be acquired.

In the above-described embodiment, the noise detection units 41a and 41b and the output control units 42a and 42b have been described as examples realized by the DSPs 40a and 40b. However, the present invention is not limited to this, and the noises are not limited thereto. The detection units 41a and 41b and the output control units 42a and 42b may be realized by hardware such as a gate array.
The noise detection units 41a and 41b and the output control units 42a and 42b may be realized by, for example, a CPU (Central Processing Unit) of a personal computer (personal computer) instead of the DSPs 40a and 40b.

In other words, if a personal computer having a high processing capability is used, a predetermined application program (an audio signal processing program to be described later) is executed by the CPU without using the dedicated DSPs 40a and 40b. 41b and output control units 42a and 42b can realize DSPs 40a and 40b.
The functions of the noise detection units 41a and 41b, the output control units 42a and 42b, and the abnormality detection unit 44a described above are performed by a computer (including a CPU, an information processing device, and various terminals) by a predetermined application program (audio signal processing). It may be realized by executing a program.

  The program is, for example, a computer such as a flexible disk, CD (CD-ROM, CD-R, CD-RW, etc.), DVD (DVD-ROM, DVD-RAM, DVD-R, DVD-RW, DVD + R, DVD + RW, etc.). It is provided in a form recorded on a readable recording medium. In this case, the computer reads the audio signal processing program from the recording medium, transfers it to an internal storage device or an external storage device, and uses it. Further, the program may be recorded in a storage device (recording medium) such as a magnetic disk, an optical disk, or a magneto-optical disk, and provided from the storage device to a computer via a communication line.

  Here, the computer is a concept including hardware and an OS (operating system) and means hardware that operates under the control of the OS. Further, when the OS is unnecessary and the hardware is operated by the application program alone, the hardware itself corresponds to the computer. The hardware includes at least a microprocessor such as a CPU and means for reading a computer program recorded on a recording medium.

The application program as the audio signal processing program includes a program code that causes the above-described computer to realize the functions as the noise detection units 41a and 41b, the output control units 42a and 42b, and the abnormality detection unit 44a. Also, some of the functions may be realized by the OS instead of the application program.
In addition to the above-described flexible disk, CD, DVD, magnetic disk, optical disk, and magneto-optical disk, the recording medium according to this embodiment includes an IC card, ROM cartridge, magnetic tape, punch card, and internal storage device of a computer ( It is also possible to use various computer-readable media such as a memory such as a RAM or a ROM, an external storage device, or a printed matter on which a code such as a barcode is printed.

It is a block diagram which shows the structure of the audio | voice signal processing apparatus as 1st Embodiment of this invention. It is a figure for demonstrating the example of application to the mobile telephone of the audio | voice signal processing apparatus as 1st Embodiment of this invention. It is a figure for demonstrating the example which mounted the microphone of the input part of the audio | voice signal processing apparatus as 1st Embodiment of this invention in the headset. (A)-(c) is a figure which each shows the audio | voice waveform of the audio | voice signal input from three input parts of the audio | voice signal processing apparatus as 1st Embodiment of this invention. It is a figure which shows the audio | voice waveform of the audio | voice signal output from the output control part of the audio | voice signal processing apparatus as 1st Embodiment of this invention. It is a flowchart which shows the operation | movement procedure of the audio | voice signal processing method as 1st Embodiment of this invention. It is a block diagram which shows the structure of the audio | voice signal processing apparatus as 2nd Embodiment of this invention. (A), (b) is a figure which respectively shows the audio | voice waveform of the audio | voice signal input from the several input part of the audio | voice signal processing apparatus as 2nd Embodiment of this invention. It is a figure for demonstrating the adjustment signal produced | generated by the output control part of the audio | voice signal processing apparatus as 2nd Embodiment of this invention. It is a block diagram which shows the structure of the audio | voice signal processing apparatus as a modification of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1a-1c Audio | voice signal processing apparatus 2 Mobile phone 3 Operation button 4 Display part 5 Headset 6 Headphone part 10,20,30 Input part 11,21,31 Microphone 12,22,32 Microphone amplifier 13,23,33 ADC
40a-40c DSP
41a, 41b Noise detection unit 42a, 42b Output control unit 43a Holding unit 44a Abnormality detection unit 45 Warning means 50 DAC
S1 input step S3, S5 output control step S4 noise detection step

Claims (8)

Three or more input units for inputting audio signals;
An output control unit that generates and outputs one output audio signal based on three or more audio signals input by the three or more input units;
When the three or more audio signals input by the three or more input units are compared, and there is a difference between the three or more audio signals, the majority of the three or more audio signals are subjected to a majority vote. A noise detector that detects the difference between the majority voice signal as noise,
When the noise is detected by the noise detection unit, the output control unit reduces or eliminates a contribution rate of the audio signal including the noise to the generation of the output audio signal. Processing equipment. A plurality of input units for inputting audio signals;
An output control unit that generates and outputs one output audio signal based on the plurality of audio signals input by the plurality of input units;
A noise detector that detects a portion of the plurality of audio signals whose signal level exceeds a threshold as noise;
When the noise is detected by the noise detection unit, the output control unit reduces or eliminates a contribution rate of the audio signal including the noise to the generation of the output audio signal. Processing equipment.   If no noise is detected by the noise detection unit, the output control unit generates and outputs the output audio signal by combining the audio signals at a predetermined ratio, while the noise detection unit detects the noise. The audio signal processing apparatus according to claim 1, wherein the ratio of the audio signal in which the noise is detected is reduced.   The output control unit outputs an average value of a plurality of the audio signals as the output audio signal if no noise is detected by the noise detection unit. On the other hand, when noise is detected by the noise detection unit, the noise is 3. The audio signal processing apparatus according to claim 1, wherein an average value of audio signals other than the detected audio signal is output as the output audio signal. A holding unit that temporarily holds the audio signal input by the input unit;
5. The audio signal according to claim 1, wherein the output control unit generates and outputs the output audio signal based on the audio signal held in the holding unit. 6. Processing equipment. An input step for inputting three or more audio signals;
An output control step of generating and outputting one output audio signal based on the three or more audio signals input in the input step;
When three or more audio signals input in the input step are compared and a difference occurs between the three or more audio signals, the majority voice of the minority audio signal is determined by a majority vote between the three or more audio signals. A noise detection step of detecting a difference portion with respect to the signal as noise,
In the output control step, when noise is detected in the noise detection step, the contribution rate of the audio signal including noise to the generation of the output audio signal is reduced or made zero. Processing method. An audio signal processing program for causing a computer to realize a function of outputting one output audio signal based on three or more audio signals input from three or more input units,
An output control unit that generates and outputs the one output audio signal based on the three or more audio signals input by the three or more input units; and
When the three or more audio signals input by the three or more input units are compared, and there is a difference between the three or more audio signals, the majority of the three or more audio signals are subjected to a majority vote. As a noise detection unit that detects a difference portion with respect to a majority voice signal as noise, the computer functions,
When the noise is detected by the noise detection unit, the output control unit causes the computer to function so as to reduce or reduce the contribution ratio of the audio signal including noise to the generation of the output audio signal. An audio signal processing program. A computer-readable recording medium on which an audio signal processing program for causing a computer to realize a function of outputting one output audio signal based on three or more audio signals input from three or more input units,
The audio signal processing program is
An output control unit that generates and outputs the one output audio signal based on the three or more audio signals input by the three or more input units; and
When the three or more audio signals input by the three or more input units are compared, and there is a difference between the three or more audio signals, the majority of the three or more audio signals are subjected to a majority vote. As a noise detection unit that detects a difference portion with respect to a majority voice signal as noise, the computer functions,
When the noise is detected by the noise detection unit, the output control unit causes the computer to function so as to reduce or reduce the contribution ratio of the audio signal including noise to the generation of the output audio signal. The computer-readable recording medium which recorded the audio | voice signal processing program characterized by the above-mentioned.

Images