WO2021085174A1

WO2021085174A1 - Voice processing device and voice processing method

Info

Publication number: WO2021085174A1
Application number: PCT/JP2020/039054
Authority: WO
Inventors: 洋平櫻庭
Original assignee: ソニー株式会社
Priority date: 2019-10-30
Filing date: 2020-10-16
Publication date: 2021-05-06

Abstract

The present invention pertains to a voice processing device and a voice processing method with which it is possible to minimize acoustic feedback more assuredly. Provided is a voice processing device equipped with a signal processing unit that processes a voice signal picked up by a microphone, that compares, with a given threshold, an index relating to the amount of wraparound of a sound going into the microphone according to a voice signal outputted from a speaker, and that performs a threshold value determination. Accordingly, it is possible to minimize acoustic feedback more assuredly. The present invention is applicable to, for example, a sound amplification system that performs off-microphone sound amplification.

Description

Voice processing device and voice processing method

The present technology relates to a voice processing device and a voice processing method, and more particularly to a voice processing device and a voice processing method capable of suppressing howling more reliably.

In a system composed of a microphone and a speaker, a method of detecting the frequency at which howling occurs by prior measurement and providing a notch filter when suppressing howling is used in many howling suppressors (for example, patents). Reference 1).

Japanese Unexamined Patent Publication No. 2008-224816

However, the method of providing a notch filter is not sufficient as a countermeasure when the parameters in the signal processing unit that processes the audio signal are changed, and a technique for surely suppressing howling has been required.

This technology was made in view of such a situation, and makes it possible to suppress howling more reliably.

The voice processing device of one aspect of the present technology processes the voice signal picked up by the microphone, and sets an index regarding the amount of wraparound of the sound corresponding to the voice signal output from the speaker to the microphone as a predetermined threshold value. It is a voice processing device including a signal processing unit that compares and determines a threshold.

The voice processing method of one aspect of the present technology is an index relating to the amount of sound wrapping around the microphone according to the voice signal output from the speaker by the voice processing device processing the voice signal picked up by the microphone. Is a voice processing method for determining a threshold by comparing with a predetermined threshold.

In the audio processing device and the audio processing method of one aspect of the present technology, the audio signal picked up by the microphone is processed, and the amount of the sound corresponding to the audio signal output from the speaker wraps around the microphone. The index is compared with a predetermined threshold, and the threshold determination is performed.

The voice processing device on one aspect of the present technology may be an independent device or an internal block constituting one device.

It is a figure which shows the example of the installation of the microphone and the speaker to which this technology is applied. It is a block diagram which shows the example of the structure of one Embodiment of the voice processing apparatus to which this technique is applied. It is a flowchart explaining the flow of a feedback rate setting process. It is a flowchart explaining the flow of a feedback rate calculation process. It is a flowchart explaining the flow of a feedback rate control process. It is a figure which shows the example of feedback rate control. It is a flowchart explaining the flow of a feedback rate update process. It is a figure which shows the example of feedback rate update. It is a block diagram which shows the example of another configuration of one Embodiment of the voice processing apparatus to which this technique is applied. It is a block diagram which shows the example of still another structure of one Embodiment of the voice processing apparatus to which this technique is applied. It is a block diagram which shows the example of the structure of one Embodiment of the information processing apparatus to which this technique is applied. It is a figure which shows the 1st example of the presentation of evaluation information. It is a figure which shows the 2nd example of the presentation of evaluation information. It is a block diagram which shows the example of the hardware configuration of a computer.

<1. First Embodiment>

Generally, a hand microphone or pin microphone is used for loudspeaking (sound picked up by a microphone is reproduced from a speaker installed in the same room). The reason for this is that it is necessary to reduce the sensitivity of the microphone in order to reduce the amount of wraparound to the speaker or microphone, and it is necessary to install the microphone near the speaker's mouth so that the volume can be increased. is there.

On the other hand, as shown in FIG. 1, instead of a hand-held microphone or a pin microphone, installing a microphone at a position away from the speaker's mouth, such as a microphone 10 mounted on the ceiling, to louden the sound is called off-mic loudspeaker. I'm out. For example, in FIG. 1, the voice spoken by the teacher is picked up by a microphone 10 mounted on the ceiling and loudened throughout the classroom so that the students can hear it.

However, if you actually make an off-mic loudspeaker in a classroom or conference room, howling will occur. The reason for this is that the ceiling-mounted microphone 10 needs to be more sensitive than a handheld microphone or pin microphone, so that the amount of self-voice wraparound from the speaker 20 to the microphone 10 is large, that is, acoustic coupling is achieved. Because it is big.

The device that suppresses howling is called a howling suppressor or feedback reducer. There are two types of feedback reducers: pre-measurement type and follow-up type.

In a pre-measurement type feedback reducer, it is common to measure whether howling occurs in advance, and if howling occurs, insert a notch filter at that frequency. Further, instead of the notch filter, a graphic equalizer or the like may be used to reduce the gain of the frequency at which howling occurs.

The follow-up type feedback reducer automatically detects howling and dynamically adds a notch filter to the frequency at which howling occurs, or lowers the gain of the frequency at which howling occurs with a graphic equalizer instead of the notch filter. It is a device of.

The pre-measurement type feedback reducer has a problem that new howling occurs when internal parameters such as volume and equalizer are changed.

On the other hand, in the follow-up type feedback reducer, when new howling occurs due to changes in internal parameters such as volume and equalizer, suppression processing that automatically detects it and prevents howling from occurring will work. However, since it takes a certain amount of time for howling to be detected and the suppression filter to operate, there is a problem that howling is heard during that time period.

In view of this situation, this technology makes it possible to suppress howling more reliably. In particular, in this technology, howling can be reliably suppressed even when internal parameters such as volume and equalizer are changed during off-mic loudspeaker.

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

(Configuration of audio processing device)
FIG. 2 shows an example of the configuration of an embodiment of a voice processing device to which the present technology is applied.

In FIG. 2, the voice processing device 1 includes an A / D conversion unit 12, a signal processing unit 13, and a signal output unit 14. The voice processing device 1 is electrically connected to each of the microphone 10 and the speaker 20.

The microphone 10 is composed of a microphone unit 11-1 and a microphone unit 11-2. An A / D conversion unit 12 is provided after the two microphone units 11.

Note that the microphone 10 may be provided with one or more microphone units 11. Further, the microphone unit 11 may be an omnidirectional microphone or a unidirectional microphone.

The microphone unit 11-1 collects voice (sound) and supplies a voice signal as an analog signal to the A / D conversion unit 12. The microphone unit 11-2 collects voice (sound) and supplies a voice signal as an analog signal to the A / D conversion unit 12.

The A / D conversion unit 12 converts the audio signals supplied from the microphone unit 11-1 and the microphone unit 11-2 from analog signals to digital signals and supplies them to the signal processing unit 13.

The signal processing unit 13 is configured as a digital signal processor (DSP: Digital Signal Processor) or the like. The signal processing unit 13 performs predetermined signal processing on the audio signal supplied from the A / D conversion unit 12, and supplies the resulting audio signal (audio signal with suppressed howling) to the signal output unit 14. To do.

The signal output unit 14 includes an audio output terminal. The signal output unit 14 outputs the audio signal supplied from the signal processing unit 13 to the speaker 20 connected to the audio output terminal.

The speaker 20 processes the audio signal output from the audio processing device 1 (signal output unit 14), and outputs the audio (sound) corresponding to the audio signal.

Note that the voice processing device 1 may include at least one of the microphone 10 and the speaker 20. Further, the microphone 10 may include all or at least a part of the A / D conversion unit 12, the signal processing unit 13, and the signal output unit 14. That is, the signal processing unit 13 may be provided inside the housing of the microphone 10 or may be provided inside the housing of the microphone 10 and externally attached to the housing of the microphone 10.

Further, the signal processing unit 13 includes an auto equalizer unit 101, a volume unit 102, a calibration signal generation unit 103, an output sound power value calculation unit 104, an input sound power value calculation unit 105, and a feedback rate calculation unit 106. Will be done.

The auto equalizer unit 101 automatically changes the frequency characteristics of the audio signal input therein and supplies it to the volume unit 102.

The volume unit 102 adjusts the volume of the audio signal supplied from the auto equalizer unit 101 and supplies it to the signal output unit 14.

The calibration signal generation unit 103 generates a calibration signal such as a white noise signal or a pink noise signal during a calibration period such as during setting, and supplies the calibration signal to the signal output unit 14. The signal output unit 14 outputs the calibration signal supplied from the calibration signal generation unit 103 to the speaker 20.

As a result, during the calibration period, the speaker 20 outputs a calibration sound corresponding to the calibration signal input from the signal output unit 14. The calibration sound is picked up by the microphone 10, and the audio signal is input to the signal processing unit 13.

At this time, in the signal processing unit 13, each of the output sound power and the input sound power is calculated by the output sound power value calculation unit 104 and the input sound power value calculation unit 105.

The output sound power value calculation unit 104 calculates the output sound power value based on the calibration signal input therein, that is, the audio signal output from the signal processing unit 13, and supplies the output sound power value to the feedback rate calculation unit 106.

The input sound power value calculation unit 105 calculates the input sound power value based on the audio signal input therein, that is, the audio signal input to the signal processing unit 13, and supplies the input sound power value to the feedback rate calculation unit 106.

The feedback rate calculation unit 106 is input with the output sound power value from the output sound power value calculation unit 104 and the input sound power value from the input sound power value calculation unit 105. The feedback rate calculation unit 106 calculates the feedback rate by performing a predetermined calculation using the output sound power value and the input sound power value.

Here, the feedback rate is an index related to the amount of sound wraparound (wraparound rate) that quantitatively expresses how much the sound output from the speaker 20 wraps around the microphone 10. Hereinafter, a case where the feedback rate is obtained for each frequency band will be described.

The feedback rate calculated by the feedback rate calculation unit 106 is supplied to the auto equalizer unit 101. As a result, the feedback rate is set in the auto equalizer unit 101 during the calibration period.

After that, at the time of off-mic loudspeaker, the speaker 20 outputs a sound corresponding to the audio signal input from the signal output unit 14. This sound is picked up by the microphone 10, and the audio signal is input to the signal processing unit 13.

At this time, the auto equalizer unit 101 performs threshold determination using the feedback rate of each frequency band. When the feedback rate exceeds the threshold value, the auto equalizer unit 101 calculates a gain for lowering the feedback rate to the threshold value or less, and applies the gain to the audio signal.

Further, when the volume is changed by the volume unit 102 at the time of off-mic loudspeaker, the auto equalizer unit 101 resets (updates) the feedback rate in conjunction with the changed volume.

The auto equalizer unit 101 is corrected by updating the feedback rate, and the threshold value is determined using the corrected feedback rate of each frequency band. In the auto equalizer unit 101, when the corrected feedback rate exceeds the threshold value, a gain for lowering the corrected feedback rate to the threshold value or less is calculated, and the gain is applied to the audio signal.

In the voice processing device 1 configured as described above, the signal processing unit 13 processes the voice signal picked up by the microphone 10 and transmits the voice according to the voice signal output from the speaker 20 to the microphone 10. A threshold judgment is performed by comparing the index (feedback rate) related to the amount of wraparound with a predetermined threshold, and the gain (suppression gain) corresponding to the threshold judgment is applied to the voice signal to reduce howling that occurs during off-microphone loudspeaker. At the same time, the loudspeaker sound quality can be improved.

Further, when the volume is changed, the signal processing unit 13 corrects the feedback rate according to the volume change and adjusts the gain (suppression gain) for howling countermeasures. Even if it is changed, the occurrence of howling can be suppressed.

In particular, compared to the current pre-measurement type feedback reducer, it is possible to suppress the occurrence of new howling when changing internal parameters such as volume. Further, as compared with the current follow-up type feedback reducer, howling does not occur because there is no delay until howling is detected.

Next, the details of signal processing by the signal processing unit 13 will be described with reference to FIGS. 3 to 8.

(Feedback rate setting process)
First, the flow of the feedback rate setting process performed at the time of setting will be described with reference to the flowchart of FIG.

In step S11, it is determined whether or not it is at the time of setting. If it is determined in the determination process of step S11 that it is the time of setting, the process proceeds to step S12, and the processes of steps S12 to S15 are performed in order to set the feedback rate.

In step S12, the calibration signal generation unit 103 generates a calibration signal. As this calibration signal, a white noise signal or the like is generated.

In step S13, the signal output unit 14 outputs the generated calibration signal to the speaker 20. As a result, white noise or the like is output from the speaker 20 as a calibration sound.

In step S14, the feedback rate calculation process is performed. Here, the details of the feedback rate calculation process will be described with reference to the flowchart of FIG.

In step S31, the output sound power value calculation unit 104 calculates the output sound power value based on the calibration signal output from the signal processing unit 13.

In step S32, the input sound power value calculation unit 105 calculates the input sound power value based on the voice signal input to the signal processing unit 13.

In step S33, the feedback rate calculation unit 106 calculates the feedback rate by performing a predetermined calculation using the calculated output sound power value and input sound power value.

For example, when the feedback rate is calculated for each frequency band, if the output sound for each frequency band is Y (ω), the input sound is M (ω), and the feedback rate for each frequency band is F0 (ω), then F0 ( ω) is represented by the following equation (1).

F0 (ω) = ｜ M (ω) ｜ / ｜ Y (ω) ｜・・・ (1)

That is, in the equation (1), at frequencies where the value of F0 (ω) exceeds 1.0 times, howling occurs because the sound output from the speaker 20 is input to the microphone 10 more loudly. Means.

Further, the equation (1) is wide because even if the value of F0 (ω) is less than 1.0, howling does not occur when the value exceeds a certain value (for example, about 0.5), but the sound loop is large. It means that the reverberation will occur as if you were talking in the hall. For example, this approximately 0.5 can be set to the threshold value T as a criterion for determining sound quality with a reverberant feeling.

When the process of step S33 is completed, the process returns to step S14 of FIG. 3, and the subsequent processes are performed.

In step S15, the auto equalizer unit 101 sets the feedback rate F0 (ω) for each frequency band calculated by the feedback rate calculation process.

If the process of step S15 is completed, or if it is determined in the determination process of step S11 that it is not the time of setting, the process is terminated.

The flow of the feedback rate setting process has been explained above. In this feedback rate setting process, the feedback rate F0 (ω) for each frequency band is calculated using the calibration sound at the time of setting and is set as an initial value.

Note that the feedback rate setting process is not limited to the setting time, and may be executed when a predetermined operation such as pressing the start button is performed at the start of use such as at the start of a class or a conference.

(Feedback rate control processing)
Next, the flow of the feedback rate control process performed at the time of off-mic loudspeaker after the setting is completed will be described with reference to the flowchart of FIG.

In step S51, the auto equalizer unit 101 determines whether or not the feedback rate F0 (ω) of each frequency band is equal to or less than the threshold value T for the audio signal input therein. This feedback rate F0 (ω) is set at the time of setting or the like.

If it is determined in the determination process of step S51 that the feedback rate F0 of each frequency band is equal to or less than the threshold value T, the process proceeds to step S52. In step S52, the auto-equalizer unit 101 sets the suppression gain G (ω) of the frequency band to 1.0.

On the other hand, if it is determined in the determination process of step S51 that the feedback rate F0 of each frequency band exceeds the threshold value T, the process proceeds to step S53. In step S53, the auto-equalizer unit 101 calculates and sets the suppression gain G (ω) of the frequency band by the following equation (2).

G (ω) = T / F0 (ω) ・・・ (2)

That is, the suppression gain G (ω) is calculated by the relationship shown in the following equation (3) according to the result of the threshold value determination of the feedback rate F0 (ω).

F0 (ω) ≤ T: G (ω) = 1.0
F0 (ω)> T: G (ω) = T / F0 (ω) ・・・ (3)

When the process of step S52 or S53 is completed, the process proceeds to step S54.

In step S54, the auto-equalizer unit 101 applies the calculated suppression gain G (ω) for each frequency band to the audio signal input therein. That is, in the auto equalizer unit 101, the suppression gain G (ω) for limiting the actual feedback rate F0 (ω) to be equal to or less than the threshold value T is calculated, and the calculated suppression gain G (ω) is used as an audio signal. Applying.

The flow of feedback rate control processing has been explained above. In this feedback rate control process, a threshold determination is performed using the feedback rate F0 (ω) of each frequency band at the time of off-mic loudspeaker, and when the feedback rate F0 (ω) exceeds the threshold T, the feedback rate F0 (ω) is used. ) Is calculated so that the suppression gain G (ω) is equal to or less than the threshold value T, and the calculated suppression gain G (ω) is multiplied by the voice signal. As a result, the occurrence of howling can be suppressed in advance, and high-quality sound with little reverberation can be realized.

FIG. 6 shows an example of feedback rate control in the auto equalizer unit 101.

A in FIG. 6 is a bar graph showing the feedback rate (unit: dB) for each frequency band by vertical bars with the horizontal axis as the frequency. This feedback rate is obtained from the relationship between the output sound and the input sound by the above-mentioned equation (1) at the time of setting.

In A of FIG. 6, the numbers # 1 to # 32 corresponding to the vertical bars (only odd numbers are shown) are described as the values on the horizontal axis, and these numbers indicate the numbers that identify the frequency band. ing. In the auto equalizer unit 101, the feedback rate F0 (ω) for each frequency band # 1 to # 32 is set.

B in FIG. 6 is a bar graph showing the auto-equalizer value (unit: dB) for each frequency band by a vertical bar with the horizontal axis as the frequency. This auto equalizer value is an audio signal obtained by the processing of the auto equalizer unit 101 at the time of off-mic loudspeaker.

In B of FIG. 6, the auto equalizer value is shown for each frequency band # 1 to # 32. In this example, the threshold T is set to -6 dB, and the auto-equalizer value is processed so that the feedback rate F0 (ω) becomes -6 dB.

C in FIG. 6 is a bar graph showing the corrected feedback rate (unit: dB) for each frequency band by a vertical bar with the horizontal axis as the frequency. The feedback rate F0 (ω) for each frequency band # 1 to # 32 is adjusted so as to be equal to or less than the threshold value T of -6 dB.

That is, in the auto equalizer unit 101, the suppression gain G (ω) corresponding to the result of the threshold determination of the feedback rate F0 (ω) for each frequency band is applied to the audio signal in the corresponding frequency band, resulting in the result. The corrected feedback rate F0 (ω) for each frequency band can be set to the threshold value T or less.

In this way, when the positions of the microphone 10 and the speaker 20 are fixed in a classroom or the like, if the volume of the amplifier is fixed, the calibration sound is output from the speaker 20 and the sound picked up by the microphone 10 is output. By measuring, the frequency at which howling occurs can be recognized in advance.

If the feedback rate F0 (ω) exceeds 1.0 times, howling will occur, and if it exceeds approximately 0.5 times, the sound quality will have a reverberant feeling. By applying the suppression gain G (ω) according to the comparison result with the threshold value T (approximately 0.5), the actual feedback rate F0 (ω) is controlled to be 0.5 or less (-6 dB or less).

This makes it possible to suppress howling more reliably. In addition, the suppression gain G (ω) according to the result of the threshold determination of the feedback rate F0 (ω) for each frequency band is applied, and the required frequency band is maintained while maintaining the sound quality by the minimum necessary reduction of the auto equalizer value. Since it controls howling, it is possible to suppress howling and not sacrifice sound quality more than necessary.

(Feedback rate update process)
Next, the flow of the feedback rate update process executed when the volume is changed during the off-mic loudspeaker will be described with reference to the flowchart of FIG. 7.

In step S71, it is determined whether or not the volume has been changed by a user operation or the like. If it is determined in the determination process of step S71 that the volume has been changed, the process proceeds to step S72.

In step S72, the auto equalizer unit 101 updates (corrects) the value (initial value, etc.) of the feedback rate F0 (ω) set at the time of setting or the like in conjunction with the change in volume.

Here, if the amplification amount corresponding to the volume change is multiplied by P, the feedback rate F (ω) linked to the volume change can be expressed by the following equation (4). The amplification amount corresponding to the volume change is input to the auto equalizer unit 101 when the volume is changed.

F (ω) = P × F0 (ω) ・・・ (4)

When the process of step S72 is completed or it is determined in the determination process of step S71 that the volume has not been changed, the process proceeds to step S73.

In step S73, the feedback rate control process is performed. This feedback rate control process is as described with reference to FIG. 5, but here, since the feedback rate is updated (corrected) to the feedback rate F (ω) in conjunction with the volume change, the corrected feedback Using the rate F (ω), the suppression gain G (ω) for each frequency band is recalculated.

That is, the suppression gain G (ω) is calculated by the relationship shown in the following equation (5) according to the result of the threshold value determination of the corrected feedback rate F (ω).

F (ω) ≤ T: G (ω) = 1.0
F (ω)> T: G (ω) = T / F (ω) ・・・ (5)

In the auto equalizer unit 101, the suppression gain G (ω) for each frequency band calculated according to the threshold value determination of the corrected feedback rate F (ω) according to the equation (5) is applied to the audio signal.

The flow of the feedback rate update process has been explained above. In this feedback rate update process, when the volume is changed, the feedback rate F (ω) is corrected in conjunction with the volume change, and when the corrected feedback rate F (ω) exceeds the threshold value T, the feedback rate F is corrected. The suppression gain G (ω) is calculated so that (ω) is equal to or less than the threshold value T, and the calculated suppression gain G (ω) is multiplied by the voice signal.

This makes it possible to suppress the occurrence of howling and maintain high sound quality. In particular, by changing the suppression gain G (ω) for howling countermeasures in accordance with the change in volume, it is possible to suppress the occurrence of howling even if the volume is changed by the user's operation.

FIG. 8 shows an example of updating the feedback rate in the auto equalizer unit 101.

A1 and A2 in FIG. 8 represent the feedback rate F0 (ω) and the auto-equalizer value before the volume change by the vertical bars when the horizontal axis is the frequency. B1 and B2 in FIG. 8 and C1 and C2 in FIG. 8 represent the feedback rate F (ω) and the auto-equalizer value after the volume change by vertical bars when the horizontal axis is the frequency.

In A1 and A2 of FIG. 8, B1 and B2 of FIG. 8, and C1 and C2 of FIG. 8, as shown by the transition of "Gain 0 dB", "Gain + 6 dB", and "Gain + 12 dB" at the time of setting. The relationship between the feedback rate and the auto-equalizer value corresponding to the volume is shown in A of FIG. 8, and the relationship between the feedback rate and the auto-equalizer value when the volume is gradually increased during off-mic loudspeaker is shown in B and C of FIG. 8, respectively. Shown.

Comparing the feedback rates of A1 to C1 in FIG. 8, the feedback rate increases in conjunction with the increase in volume. Further, when the auto-equalizer values of A2 to C2 in FIG. 8 are compared, the feedback rate F (ω) is corrected in the auto-equalizer unit 101 in conjunction with the increase in volume, and the corrected feedback rate F (ω) is corrected. The suppression gain G (ω) calculated according to the threshold value determination of) is applied to the voice signal.

<2. Second Embodiment>

Next, with reference to FIGS. 9 and 10, not only the volume but also the interlocking with the processing that affects the sound wraparound such as beamforming and the equalizer will be described.

(Other first configuration of the voice processing device)
FIG. 9 shows an example of another configuration of an embodiment of a voice processing device to which the present technology is applied. In FIG. 9, the same parts as those in FIG. 2 are designated by the same or corresponding reference numerals, and the description thereof will be omitted as appropriate.

The microphone 10 is composed of a microphone unit 11-1 and a microphone unit 11-2. The A / D conversion units 12-1 and 12-2 convert the audio signals from the microphone units 11-1 and 11-2 from analog signals to digital signals, respectively, and supply them to the signal processing unit 13A.

The signal processing unit 13A is further provided with a beamforming unit 111 and an equalizer unit 112 in addition to the auto equalizer unit 101 to the feedback rate calculation unit 106.

The beamforming unit 111 performs beamforming processing based on the audio signals from the A / D conversion units 12-1 and 12-2.

In this beamforming process, it is possible to reduce the sensitivity other than the target sound direction while ensuring the sensitivity in the target sound direction. That is, in the beamforming process, the directivity of the microphone 10 is formed so as not to take sound from the direction in which the speaker 20 is installed (as little as possible), and is passed to the subsequent process as a monaural signal.

Here, using a method such as an adaptive beam former, as the directivity of the microphone 10 (microphone units 11-1, 11-2), a directivity that reduces the sensitivity in the direction in which the speaker 20 is installed is formed, and a monaural signal is formed. Is generated.

In order to suppress the sound from the direction of the speaker 20 (to prevent howling) by using a method such as an adaptive beam former, the beam former is used in the section where the sound is output only from the speaker 20 at the time of calibration or the like. The internal parameters of are learned, and the directivity is calculated so that a blind spot (NULL directivity) is formed in the direction in which the speaker 20 is installed.

The beamforming unit 111 can obtain the suppression amount B (ω) of the loudspeaker component by beamforming based on the directivity coefficient and the information on the loudspeaker sound calculated at the time of calibration.

Note that the beamforming method is not limited to the adaptive beamformer, and other methods such as the delay sum method and the three-microphone integration method are also known, and any method may be used.

The beamforming processed audio signal is supplied to the equalizer unit 112.

The equalizer unit 112 changes the frequency characteristics of the audio signal supplied from the beamforming unit 111 according to the user's operation, and supplies the sound signal to the auto equalizer unit 101. The equalizer unit 112 holds the gain EQ (ω) for each frequency band as an internal parameter.

In the signal processing unit 13A of FIG. 9, the auto equalizer unit 101 uses the internal parameters of the beamforming unit 111 and the equalizer unit 112 to calculate the feedback rate F (ω) of the above equation (4) by the following equation. It can be modified as in (6).

The internal parameters are input to the auto equalizer unit 101 during the beamforming process or the frequency characteristic change process.

F (ω) = P x B (ω) x EQ (ω) x F0 (ω) ... (6)

That is, the suppression gain G (ω) is calculated by the relationship shown in the following equation (7) according to the result of the threshold value determination of the corrected feedback rate F (ω).

F (ω) ≤ T (ω): G (ω) = 1.0
F (ω)> T (ω): G (ω) = T / F (ω) ・・・ (7)

In the auto equalizer unit 101, the suppression gain G (ω) for each frequency band calculated according to the threshold value determination of the corrected feedback rate F (ω) according to the equation (7) is applied to the audio signal.

In this way, by updating (correcting) the feedback rate F (ω) according to the internal parameters of the volume, equalizer, and beamforming, howling is suppressed and the suppression gain G (ω) that maintains high sound quality is obtained. be able to.

In particular, by changing the suppression gain G (ω) of howling countermeasures in conjunction with signal processing such as beamforming processing and frequency characteristic change processing, howling can be suppressed even if each internal parameter is changed. Can be done. In addition, by linking with each internal parameter, it is possible to save the trouble of additional adjustment and calibration.

In the signal processing unit 13A of FIG. 9, audio signals are input to the input sound power value calculation unit 105 from both the A / D conversion unit 12-1 and the A / D conversion unit 12-2. Only the audio signal from either one may be input.

(Other second configuration of voice processing device)
FIG. 10 shows an example of still another configuration of one embodiment of the voice processing device to which the present technology is applied. In FIG. 10, the same parts as those in FIG. 2 or 9 are designated by the same or corresponding reference numerals, and the description thereof will be omitted as appropriate.

In addition to the auto equalizer unit 101 to the feedback rate calculation unit 106, the signal processing unit 13B is further provided with a beamforming unit 111, an equalizer unit 112, a low frequency cut filter unit 121, and an auto gain control unit 122.

The low frequency cut filter unit 121 cuts a specific low frequency component among the frequency components of the audio signal supplied from the beamforming unit 111, and supplies the equalizer unit 112.

The auto gain control unit 122 automatically corrects the gain according to the input level of the audio signal supplied from the auto equalizer unit 101, keeps the output level of the signal having a level difference constant, and supplies it to the volume unit 102. To do.

In the signal processing unit 13B of FIG. 10, the calculation formula of the feedback rate F (ω) in the above equation (6) can be further modified by using the internal parameters of the low frequency cut filter unit 121 and the auto gain control unit 122. it can. Further, the suppression gain G (ω) is obtained according to the result of the threshold value determination of the corrected feedback rate F (ω), as in the above-described equation (7).

The equalizer unit 112, the low frequency cut filter unit 121, and the auto gain control unit 122 are examples of effector units that perform signal processing related to sound effects (signal processing that affects the volume), and when other signal processing is used. Similarly, the formula for calculating the feedback rate F (ω) in the above formula (6) may be further modified according to the internal parameters.

Further, in the signal processing unit 13B of FIG. 10, the arrangement position of the low frequency cut filter unit 121 and the auto gain control unit 122 is an example, and may be arranged at other positions.

<3. Third Embodiment>

By the way, by using the suppression gain G (ω) described above, it is possible to estimate the loudspeaking sound quality at the time of off-mic loudspeaking. For example, when the average value of the suppression gain G (ω) for each frequency band is the sound quality score S (t), the average suppression gain for the time t can be calculated by the following equation (8).

... (8)

At this time, when the sound quality score S (t) is 1.0, the suppression gain G (ω) is 1.0 in all frequency bands, which means that the suppression filter is not applied, so that the loudspeaker sound quality is good. ing.

When the sound quality score S (t) is equal to or higher than the threshold value T, the suppression gain G (ω) is applied in a part of the frequency band, but the amount of suppression by the suppression filter is limited, and the loudspeaker sound quality is relatively good. It means that.

When the sound quality score S (t) is less than the threshold value T, the suppression gain G (ω) is applied to many frequency bands, and the suppression filter is applied, so that the loudspeaker sound quality is relatively deteriorated. It means that you are.

By presenting the numerical value of the sound quality score S (t) and the result of the threshold value determination to the user, the installer, etc. using the GUI (Graphical User Interface) display, the LED (Light Emitting Diode), etc. as evaluation information, Loudspeaker sound quality can be conveyed in an easy-to-understand manner. By presenting the loudspeaker sound quality to the user, the installer, etc., it is possible to support volume adjustment and setting.

Specifically, by confirming this evaluation information, the user, the installer, etc., in order to improve the sound quality, for example, reduce the volume or make the installation positions of the microphone 10 and the speaker 20 suitable for loudspeaking. You can change it.

(Configuration of information processing device)
FIG. 11 is a block diagram showing an example of a configuration of an embodiment of an information processing device to which the present technology is applied.

The information processing device 100 is a device for calculating and presenting a sound quality score as an index for evaluating whether or not the loudspeaker sound quality is appropriate.

The information processing device 100 calculates the sound quality score based on the data for calculating the sound quality score (hereinafter referred to as score calculation data). Further, the information processing device 100 generates evaluation information based on data for generating evaluation information (hereinafter referred to as evaluation information generation data) and presents it to the display device 30.

The score calculation data includes the suppression gain G (ω) for each frequency band input from the voice processing device 1 (signal processing unit 13). Further, the evaluation information generation data can include, for example, in addition to the calculated sound quality score, information obtained when performing off-mic loudspeaker.

The display device 30 is a device having a display such as an LCD (Liquid Crystal Display) or an OLED (Organic Light Emitting Diode), for example. The display device 30 presents the evaluation information output from the information processing device 100.

The information processing device 100 is, of course, configured as a single electronic device such as an audio device constituting a loudspeaker system, a dedicated measuring device, or a personal computer, as well as the voice processing device 1 and the microphone 10 described above. It may be configured as a part of a function of an electronic device such as a speaker 20. Further, the information processing device 100 and the display device 30 may be integrated into one electronic device.

In FIG. 11, the information processing device 100 includes a sound quality score calculation unit 151, an evaluation information generation unit 152, and a presentation control unit 153.

The sound quality score calculation unit 151 calculates the sound quality score S (t) by applying the above equation (8) using the suppression gain G (ω) for each frequency band input therein as score calculation data. It is supplied to the evaluation information generation unit 152.

The evaluation information generation unit 152 generates evaluation information based on the sound quality score S (t) input as evaluation information generation data, and supplies the evaluation information to the presentation control unit 153. The evaluation information includes information on the sound quality at the time of off-mic loudspeaker, such as the numerical value of the sound quality score S (t) and the result of the threshold value determination.

The presentation control unit 153 controls to present the evaluation information supplied from the evaluation information generation unit 152 on the screen of the display device 30.

In the information processing apparatus 100 configured as described above, the sound quality score is calculated using the suppression gain for each frequency band, and evaluation information including the calculated numerical value of the sound quality score and the result of the threshold value determination is generated. The presentation of the generated evaluation information is controlled.

FIG. 12 shows an example of evaluation information presented as a GUI display.

In FIG. 12, the evaluation information screen 201 is presented with a user interface for adjusting the volume and three rectangular areas showing the result of the threshold value determination of the sound quality score.

For example, when the sound quality score S (t) is represented by three stages of 1.0, a threshold value T or more, or a threshold value T or less, the first from the left according to these three stages. The lighting display of the three rectangular areas is controlled, such as the lighting display of the rectangular areas of the above, the lighting display of the first and second rectangular areas from the left, and the lighting display of all the rectangular areas.

As a result of the threshold determination of the sound quality score, the suppression filter is not applied (loud sound quality is good), the suppression amount by the suppression filter is limited (loud sound quality is relatively good), or many frequency bands. The state in which the suppression filter is applied (loud sound quality is relatively deteriorated) is indicated by the lighting display.

Further, by checking the evaluation information screen 201, the user, the installer, etc. can intuitively recognize how loud the sound quality is due to the influence of howling suppression when the volume is adjusted. it can. For example, when the user, the installer, or the like recognizes from the evaluation information screen 201 that the amount of sound wraparound is large, the user or the installer can take measures such as not raising the volume any more.

In the evaluation information screen 201, the result of the threshold value determination of the sound quality score is presented by displaying the rectangular area in a different color such as green, yellow, or red, not limited to the number of lighting displays in the rectangular area. May be good. Further, when displaying the rectangular areas in different colors, the number of the rectangular areas may be one and the color of the rectangular areas may be changed.

Moreover, the evaluation information is not limited to the GUI display, and may be presented by lighting the LED. FIG. 13 shows an example of evaluation information presented by lighting the LED.

In A of FIG. 13, the information processing device 100 is provided with an LED 202 indicating the result of threshold value determination of the sound quality score.

For example, depending on the three stages of the sound quality score S (t) being 1.0, the threshold value T or more, or the threshold value T or less, the LED 202 capable of emitting three colors may be green, yellow, or red. It is controlled to light in any of the three colors. As a result, as a result of the threshold value determination of the sound quality score, a state in which the suppression filter is not applied (loud sound quality is good) is presented by the LED 202 lit in green or the like.

In B of FIG. 13, the information processing device 100 is provided with LEDs 202-1 to 202-3 indicating the result of the threshold value determination of the sound quality score.

For example, LED202-1 lights up in green and LED202-2 lights up in yellow according to the three stages of the sound quality score S (t) being 1.0, above the threshold value T, or below the threshold value T. , LED202-3 is controlled to light up in red. As a result, as a result of the threshold value determination of the sound quality score, a state in which the suppression filter is not applied (loud sound quality is good) is presented by the LED202-1 lit in green or the like.

By checking the LED 202, users, installers, etc. can intuitively recognize how loud the sound quality is due to the effect of howling suppression.

The method of presenting the evaluation information shown in FIGS. 12 and 13 is an example, and the evaluation information may be presented by another presentation method. For example, the score value of the sound quality score itself may be presented, not limited to the GUI display and the identification by the number or color by lighting the LED, or the score value of the sound quality score is read aloud and output by voice (sound). May be good.

<4. Modification example>

In the above description, the case where the feedback rate is obtained for each frequency band has been shown, but the feedback rate is a value corresponding to the ratio of the audio signal output to the speaker 20 and the audio signal input from the microphone 10, and is a frequency. It is not limited to each area, and may be obtained, for example, in the time domain.

For example, if the output sound in the time domain is Y, the input sound is M, and the feedback rate in the time domain is F0, F0 is expressed by the following equation (9).

F0 = | M | / | Y | ... (9)

The auto equalizer unit 101 performs a threshold value determination using this feedback rate F0, and calculates a suppression gain G (ω) according to the result of the threshold value determination.

When calculating the feedback rate, noise included in the input sound may be added.

In this specification, "voice signal" may be read as "sound signal", and "sound processing device" may be read as "sound processing device" or "signal processing device".

<5. Computer configuration>

The series of processes of the voice processing device 1 described above can be executed by hardware or software. When a series of processes are executed by software, the programs constituting the software are installed on the computer of each device.

FIG. 14 is a block diagram showing an example of the hardware configuration of a computer that executes the above-mentioned series of processes programmatically.

In a computer, a CPU (Central Processing Unit) 1001, a ROM (Read Only Memory) 1002, and a RAM (Random Access Memory) 1003 are connected to each other by a bus 1004. An input / output interface 1005 is further connected to the bus 1004. An input unit 1006, an output unit 1007, a storage unit 1008, a communication unit 1009, and a drive 1010 are connected to the input / output interface 1005.

The input unit 1006 includes a microphone, a keyboard, a mouse, and the like. The output unit 1007 includes a speaker, a display, and the like. The storage unit 1008 includes a hard disk, a non-volatile memory, and the like. The communication unit 1009 includes a network interface and the like. The drive 1010 drives a removable recording medium 1011 such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory.

In the computer configured as described above, the CPU 1001 loads the program recorded in the ROM 1002 and the storage unit 1008 into the RAM 1003 via the input / output interface 1005 and the bus 1004 and executes the above-mentioned series. Is processed.

The program executed by the computer (CPU1001) can be recorded and provided on the removable recording medium 1011 as a package medium or the like, for example. Programs can also be provided via wired or wireless transmission media such as local area networks, the Internet, and digital satellite broadcasting.

In the computer, the program can be installed in the storage unit 1008 via the input / output interface 1005 by mounting the removable recording medium 1011 in the drive 1010. Further, the program can be received by the communication unit 1009 and installed in the storage unit 1008 via a wired or wireless transmission medium. In addition, the program can be pre-installed in the ROM 1002 or the storage unit 1008.

Here, in the present specification, the processing performed by the computer according to the program does not necessarily have to be performed in chronological order in the order described as the flowchart. That is, the processing performed by the computer according to the program also includes processing executed in parallel or individually (for example, parallel processing or processing by an object). Further, the program may be processed by one computer (processor) or may be distributed by a plurality of computers.

The embodiment of the present technology is not limited to the above-described embodiment, and various changes can be made without departing from the gist of the present technology.

In addition, each step of the above-mentioned processing can be executed by one device or shared by a plurality of devices. Further, when a plurality of processes are included in one step, the plurality of processes included in the one step can be executed by one device or shared by a plurality of devices.

Note that this technology can have the following configuration.

(1)
A signal processing unit that processes the audio signal picked up by the microphone, compares the index related to the amount of wraparound of the sound output from the speaker to the microphone with a predetermined threshold, and determines the threshold. A voice processing device to be equipped.
(2)
The voice processing device according to (1), wherein the index is a value corresponding to a ratio of a voice signal output to the speaker and a voice signal input from the microphone.
(3)
The voice processing device according to (2), wherein the signal processing unit applies a gain according to the result of the threshold value determination to the voice signal.
(4)
The voice processing device according to (3) above, wherein the index is calculated for each frequency band or in the time domain.
(5)
When the index value calculated for each frequency band exceeds the threshold value, the signal processing unit applies the index value and the gain calculated from the threshold value to the audio signal for each frequency band (4). The audio processing device described in.
(6)
The voice processing device according to (5), wherein the signal processing unit calculates the gain so that the value of the index becomes equal to or less than the threshold value.
(7)
The signal processing unit includes an auto equalizer unit.
The voice processing device according to any one of (3) to (6) above, wherein the auto equalizer unit applies the gain to the voice signal.
(8)
The signal processing unit
Generate a calibration signal and
Described in any one of (2) to (7) above, which calculates a value according to the ratio of the calibration signal output to the speaker and the audio signal input from the microphone as the index during the calibration period. Audio processing device.
(9)
The signal processing unit
With more volume
When the volume is changed in the volume unit, a value corresponding to the ratio of the audio signal output to the speaker and the audio signal input from the microphone is calculated.
The voice processing device according to (8) above, wherein the calculated value is used to correct the value of the index.
(10)
The signal processing unit
Including a beamforming section that performs beamforming processing,
The voice processing device according to (9) above, wherein the value of the index is corrected by using the internal parameters of the beamforming unit.
(11)
The signal processing unit
It further includes an effector unit that performs processing on the voice signal.
The voice processing device according to (9) or (10), wherein the value of the index is corrected by using the internal parameters of the effector unit.
(12)
The audio processing device according to (11) above, wherein the effector unit includes at least one of an equalizer unit, an auto gain control unit, and a filter unit.
(13)
The voice processing device according to any one of (1) to (12) above, wherein the threshold value is set according to a criterion for determining sound quality with a reverberant feeling.
(14)
The voice processing device according to any one of (3) to (7) above, further comprising a calculation unit for calculating a score according to the gain.
(15)
A generator that generates evaluation information based on the score,
The voice processing device according to (14), further including a presentation control unit that controls the presentation of evaluation information.
(16)
The voice processing device according to (15) above, wherein the evaluation information includes information on sound quality at the time of loudspeaking.
(17)
The voice processing device according to any one of (1) to (16) above, wherein the microphone is installed at a position away from the mouth of the speaker.
(18)
The voice processing device according to (17), wherein the microphone and the speaker are fixedly installed at predetermined positions in the same space.
(19)
The voice processing device according to any one of (1) to (18), which is provided in the housing of the microphone or in an external housing.
(20)
The voice processing device
A voice processing method in which a voice signal picked up by a microphone is processed, an index relating to the amount of wraparound of sound corresponding to the voice signal output from a speaker to the microphone is compared with a predetermined threshold value, and a threshold value is determined.

1 sound processing device, 10 microphone, 11-1, 11-2, 11 microphone unit, 12 A / D conversion unit, 13, 13A, 13B signal processing unit, 14 signal output unit, 20 speakers, 30 display devices, 100 information Processing device, 101 auto equalizer unit, 102 volume unit, 103 calibration signal generation unit, 104 output sound power value calculation unit, 105 input sound power value calculation unit, 106 feed rate calculation unit, 111 beam forming unit, 112 equalizer unit, 121 low frequency cut filter unit, 122 auto gain control unit, 151 sound quality score calculation unit, 152 evaluation information generation unit, 153 presentation control unit, 1001 CPU

Claims

A signal processing unit that processes the audio signal picked up by the microphone, compares the index related to the amount of wraparound of the sound output from the speaker to the microphone with a predetermined threshold, and determines the threshold. A voice processing device to be equipped.
The voice processing device according to claim 1, wherein the index is a value corresponding to a ratio of a voice signal output to the speaker and a voice signal input from the microphone.
The voice processing device according to claim 2, wherein the signal processing unit applies a gain according to the result of the threshold value determination to the voice signal.
The voice processing device according to claim 3, wherein the index is calculated for each frequency band or in the time domain.
According to claim 4, when the index value calculated for each frequency band exceeds the threshold value, the signal processing unit applies the index value and the gain calculated from the threshold value to the audio signal for each frequency band. The described audio processing device.
The voice processing device according to claim 5, wherein the signal processing unit calculates the gain so that the value of the index becomes equal to or less than the threshold value.
The signal processing unit includes an auto equalizer unit.
The voice processing device according to claim 5, wherein the auto equalizer unit applies the gain to the voice signal.
The signal processing unit
Generate a calibration signal and
The audio processing device according to claim 2, wherein a value is calculated as an index according to the ratio of the calibration signal output to the speaker and the audio signal input from the microphone during the calibration period.
The signal processing unit
With more volume
When the volume is changed in the volume unit, a value corresponding to the ratio of the audio signal output to the speaker and the audio signal input from the microphone is calculated.
The voice processing device according to claim 8, wherein the calculated value is used to correct the value of the index.
The signal processing unit
Including a beamforming section that performs beamforming processing,
The voice processing apparatus according to claim 9, wherein the value of the index is corrected by using the internal parameters of the beamforming unit.
The signal processing unit
It further includes an effector unit that performs processing on the voice signal.
The voice processing device according to claim 9, wherein the value of the index is corrected by using the internal parameter of the effector unit.
The audio processing device according to claim 11, wherein the effector unit includes at least one of an equalizer unit, an auto gain control unit, and a filter unit.
The voice processing device according to claim 1, wherein the threshold value is set according to a criterion for determining sound quality with a reverberant feeling.
The voice processing apparatus according to claim 3, further comprising a calculation unit for calculating a score according to the gain.
A generator that generates evaluation information based on the score,
The voice processing device according to claim 14, further comprising a presentation control unit that controls the presentation of the evaluation information.
The voice processing device according to claim 15, wherein the evaluation information includes information on sound quality at the time of loudspeaking.
The voice processing device according to claim 1, wherein the microphone is installed at a position away from the mouth of the speaker.
The voice processing device according to claim 17, wherein the microphone and the speaker are fixedly installed at predetermined positions in the same space.
The voice processing device according to claim 1, which is provided in the housing of the microphone or in an external housing.
The voice processing device
A voice processing method in which a voice signal picked up by a microphone is processed, an index relating to the amount of wraparound of the sound corresponding to the voice signal output from the speaker to the microphone is compared with a predetermined threshold value, and a threshold value is determined.