JP6096437B2 - Audio processing device - Google Patents

Description

  The present invention relates to an audio processing device, and more particularly to an audio processing device that adjusts gains of a plurality of audio signals acquired in parallel.

  An example of this type of device is disclosed in Patent Document 1. According to this background art, the audio signal captured by one microphone is input to the first detector via the amplifier and the first LPF. The audio signal captured by the other microphone is input to the second detector via the variable gain amplifier and the second LPF. The output of the first detector and the output of the second detector are compared with each other by the comparator, and the amplification factor of the variable gain amplifier is adjusted based on the comparison result. Thereby, variation in sensitivity of the microphone can be suppressed.

JP 2005-136628 A

  However, in the background art, the amplification factor of the variable gain amplifier is not adjusted by the incident angle of the audio signal to the microphone, and the phase variation between the audio signals captured by each microphone is not suppressed. For this reason, in the background art, the quality of the audio signal after adjustment is limited.

  Therefore, a main object of the present invention is to provide an audio processing apparatus capable of improving the quality of an audio signal.

  The speech processing apparatus according to the present invention (10: reference numerals corresponding to the embodiments; the same applies hereinafter) receives N (N: 2 integers) of M speech signals (M: an integer of 2 or more) acquired in parallel. Classification means (54, S1 to S5, S13 to S15) for classifying into N signal components respectively corresponding to the frequencies of the above integers), and M corresponding to each of the N frequencies with reference to the output of the classification means Detecting means (S7 to S11) for detecting a phase difference between the signal components, and a first specifying for specifying a phase difference lower than the first threshold (TH1) among the N phase differences detected by the detecting means. Means (S19), and first adjustment means for adjusting the amplitudes of the M audio signals so that the level difference between the M signal components defining the phase difference specified by the first specifying means is suppressed ( 50, S23 to S25, S33 to S37).

  Preferably, the first threshold value is a value based on a distance between M microphones (34L, 34R) that respectively acquire M sound signals and an upper limit of an allowable incident angle of the M sound signals.

  Preferably, the second specifying means (S21, S27) for specifying a phase difference indicating a value equal to or greater than the second threshold (TH2) from the N phase differences detected by the detecting means, and the second specifying means Second adjustment means (52, S39 to S41) for adjusting the delay amount of the M audio signals so as to suppress the phase difference is further provided.

  More preferably, the second threshold value indicates a value based on a distance between M microphones (34L, 34R) that respectively acquire M audio signals.

  Preferably, the classification means includes conversion means (54) for Fourier transforming each of M (M: an integer of 2 or more) audio signals.

  An audio processing apparatus (10) according to the present invention has N (M: integer greater than or equal to 2) audio signals acquired in parallel, each of which corresponds to N (N: integer greater than or equal to 2) frequencies. Classification means (54, S1 to S5, S13 to S15) for classifying into signal components, and detecting the phase difference between M signal components corresponding to each of N frequencies with reference to the output of the classification means Detecting means (S7 to S11), specifying means (S21, S27) for specifying a phase difference indicating a value equal to or greater than a threshold value (TH2) among N phase differences detected by the detecting means, and specifying by the specifying means Adjusting means (52, S39 to S41) for adjusting the delay amount of the M audio signals so as to suppress the phase difference.

  The sound processing device (10) according to the present invention is a detection means (S1 to S15) for detecting relative phase difference information of a plurality of sound signals acquired in parallel, the amplitude between the plurality of sound signals caused by component variations, Discriminating means (S17 to S21, S29 to S31) for discriminating the phase shift based on the relative phase difference information detected by the detecting means, correcting means (50, 52) for correcting the amplitude and phase of a plurality of audio signals, and Adjustment means (S23 to S27, S33 to S41) for adjusting the correction amount of the correction means based on the determination result of the determination means is provided.

  The audio processing program according to the present invention is configured such that the processor (56) of the audio processing device (10) receives N (N: 2 or more) each of M (M: integer of 2 or more) audio signals acquired in parallel. Classification step (S1 to S5, S13 to S15) for classifying into N signal components respectively corresponding to frequencies of M), and M signals corresponding to each of the N frequencies with reference to the output of the classification step A detection step (S7 to S11) for detecting a phase difference between components, a specification step (S19) for specifying a phase difference below a threshold (TH1) among N phase differences detected by the detection step, and a specification An adjustment step (50, S23 to S25, S33 to S37) for adjusting the amplitude of the M audio signals so that the level difference between the M signal components defining the phase difference identified by the step is suppressed. This is a voice processing program for execution.

  The audio processing method according to the present invention is an audio processing method executed by the processor (56) of the audio processing device (10), and is an M (M: integer of 2 or more) audio signals acquired in parallel. Classification steps (S1 to S5, S13 to S15) for classifying each of the signals into N signal components corresponding to N (N: an integer of 2 or more) frequencies, and N frequencies with reference to the output of the classification step Detection step (S7 to S11) for detecting the phase difference between M signal components corresponding to each of the above, and the phase difference below the threshold (TH1) is identified from the N phase differences detected by the detection step A specific step (S19), and an adjusting step (50,) for adjusting the amplitude of the M audio signals so that a level difference between the M signal components defining the phase difference specified by the specific step is suppressed. S23 to S25, S33 to S37).

  The audio processing program according to the present invention is configured such that the processor (56) of the audio processing device (10) receives N (N: 2 or more) each of M (M: integer of 2 or more) audio signals acquired in parallel. Classification step (54, S1 to S5, S13 to S15) for classifying into N signal components respectively corresponding to frequencies of M), and M corresponding to each of the N frequencies with reference to the output of the classification step Detection step (S7 to S11) for detecting a phase difference between the signal components of the signal, and a specifying step for specifying a phase difference indicating a value equal to or greater than a threshold value (TH2) among the N phase differences detected by the detection step ( S21, S27), and an audio processing program for executing an adjustment step (52, S39 to S41) for adjusting the delay amount of the M audio signals so that the phase difference specified by the specific step is suppressed is there.

  The audio processing method according to the present invention is an audio processing method executed by the processor (56) of the audio processing device (10), and is an M (M: integer of 2 or more) audio signals acquired in parallel. Classification step (54, S1 to S5, S13 to S15) for classifying each signal into N signal components corresponding to N (N: integer greater than or equal to 2) frequencies, N with reference to the output of the classification step A detection step (S7 to S11) for detecting a phase difference between M signal components corresponding to each of the frequencies, and a value equal to or greater than a threshold (TH2) among the N phase differences detected by the detection step. A specific step (S21, S27) for identifying the phase difference shown, and an adjustment step (52, S39 to S41) for adjusting the delay amount of the M audio signals so that the phase difference identified by the specific step is suppressed. Prepare.

  The amplitudes of the M audio signals are adjusted so that the level difference between the M signal components defining the phase difference below the first threshold is suppressed. That is, the level difference between the sound components incident at an angle lower than the angle corresponding to the first threshold is suppressed. This improves the quality of the audio signal.

  The phases of the M audio signals are adjusted so that a phase difference equal to or greater than a threshold value is suppressed. That is, by setting the threshold value as the theoretical maximum threshold value determined from the microphone interval, a phase difference exceeding the maximum phase difference caused by quality variation is suppressed. By repeating this suppression processing, the phase difference falls within the maximum threshold value for sound coming from any direction. As a result, the delay due to the quality variation is corrected, and the quality of the audio signal is improved.

  The above object, other objects, features and advantages of the present invention will become more apparent from the following detailed description of embodiments with reference to the drawings.

(A) is a block diagram showing a basic configuration of one embodiment of the present invention, and (B) is a block diagram showing a basic configuration of another embodiment of the present invention. It is a block diagram which shows the structure of one Example of this invention. FIG. 3 is a block diagram illustrating an example of a configuration of a sound processing circuit applied to the embodiment in FIG. 2; FIG. 4 is a flowchart showing a part of the operation of a control circuit provided in the sound processing circuit shown in FIG. 3. FIG. 4 is a flowchart showing another part of the operation of the control circuit provided in the sound processing circuit shown in FIG. 3. FIG. 4 is a flowchart showing another part of the operation of the control circuit provided in the sound processing circuit shown in FIG. 3. It is an illustration figure which shows an example of the audio | voice signal which injects into a microphone. (A) is an illustrative view showing an example of a waveform of an L channel frequency component, and (B) is an illustrative view showing an example of a waveform of an R channel frequency component. It is an illustration figure which shows another example of the audio | voice signal which injects into a microphone. (A) is an illustrative view showing another example of the waveform of the L channel frequency component, and (B) is an illustrative view showing another example of the waveform of the R channel frequency component. It is an illustration figure which shows another example of the audio | voice signal which injects into a microphone. (A) is an illustrative view showing another example of the waveform of the L channel frequency component, and (B) is an illustrative view showing another example of the waveform of the R channel frequency component. It is a block diagram which shows another example of a structure of the audio | voice processing circuit applied to the FIG. 2 Example. FIG. 7 is a block diagram illustrating another example of the configuration of the sound processing circuit applied to the embodiment in FIG. 2;

Embodiments of the present invention will be described below with reference to the drawings.
[Basic configuration 1]

  Referring to FIG. 1A, the sound processing apparatus of this embodiment is basically configured as follows. The classifying means 1a classifies each of M (M: integer greater than or equal to 2) audio signals acquired in parallel into N signal components respectively corresponding to N (N: integer greater than or equal to 2) frequencies. To do. The detection means 2a refers to the output of the classification means 1a and detects the phase difference between the M signal components corresponding to each of the N frequencies. The first specifying unit 3a specifies a phase difference that is lower than the first threshold value from among the N phase differences detected by the detecting unit 2a. The first adjusting unit 4a adjusts the amplitudes of the M audio signals so that the level difference between the M signal components defining the phase difference specified by the first specifying unit 3a is suppressed.

The amplitudes of the M audio signals are adjusted so that the level difference between the M signal components defining the phase difference below the first threshold is suppressed. That is, the amplitude of the entire area of the M audio signals is adjusted so that the level difference between the audio components incident at an angle lower than the angle corresponding to the first threshold is suppressed. As a result, variations in microphone sensitivity are corrected, and the quality of the audio signal is improved.
[Basic configuration 2]

  With reference to FIG. 1 (B), the speech processing apparatus of another Example is fundamentally comprised as follows. The classifying means 1b classifies each of M (M: integer greater than or equal to 2) audio signals acquired in parallel into N signal components respectively corresponding to N (N: integer greater than or equal to 2) frequencies. To do. The detection means 2b refers to the output of the classification means 1b and detects the phase difference between the M signal components corresponding to each of the N frequencies. The specifying unit 3b specifies a phase difference indicating a value greater than or equal to a threshold value from among the N phase differences detected by the detecting unit 2b. The adjusting unit 4b adjusts the delay amount of the M audio signals so that the phase difference specified by the specifying unit 3b is suppressed.

The phases of the M audio signals are adjusted so that a phase difference equal to or greater than a threshold value is suppressed. That is, by setting the threshold value as the theoretical maximum threshold value determined from the microphone interval, a phase difference exceeding the maximum phase difference caused by quality variation is suppressed. By repeating this suppression processing, the phase difference falls within the maximum threshold value for sound coming from any direction. As a result, the delay due to the quality variation is corrected, and the quality of the audio signal is improved.
[Example]

  Referring to FIG. 2, the digital camera 10 of this embodiment includes a focus lens 12 and an aperture unit 14 driven by drivers 18a and 18b, respectively. The optical image that has passed through these members is irradiated onto the imaging surface of the imager 16 and subjected to photoelectric conversion.

  When the power is turned on, the CPU 30 instructs the driver 18c to repeat the exposure operation and the charge readout operation in order to execute the moving image capturing process. The driver 18c exposes the imaging surface of the imager 16 in response to a periodically generated vertical synchronization signal Vsync, and reads out the charges generated on the imaging surface in a raster scanning manner. From the imager 16, raw image data based on the read charges is periodically output.

  The camera processing circuit 20 performs processing such as white balance adjustment, color separation, and YUV conversion on the raw image data output from the imager 16. The YUV format image data generated thereby is written into the YUV image area 24 a of the SDRAM 24 through the memory control circuit 22. The LCD driver 26 repeatedly reads out the image data stored in the YUV image area 24a through the memory control circuit 22, and drives the LCD monitor 28 based on the read image data. As a result, a real-time moving image (through image) representing the scene captured on the imaging surface is displayed on the monitor screen.

  The camera processing circuit 20 also provides the CPU 30 with Y data generated by YUV conversion. The CPU 30 performs AE processing on the given Y data to calculate an appropriate EV value, and sets the aperture amount and the exposure time that define the calculated appropriate EV value in the drivers 18b and 18c, respectively. This ensures the brightness of the through image. The CPU 30 also continuously executes the AF process with reference to the high frequency component of the Y data given from the preprocessing circuit 20. Accordingly, the focus lens 12 is continuously disposed in the vicinity of the in-focus point, and the sharpness of the through image is ensured.

  When the movie button 32mv provided on the key input device 32 is operated, the CPU 30 activates the voice processing circuit 36 and the memory I / F 38. The audio processing circuit 36 performs audio processing described later on the L channel audio data and the R channel audio data output from the microphones 34L and 34R, respectively. The processed L-channel audio data and R-channel audio data are written into the audio area 24 b of the SRAM 24 via the memory control circuit 22.

  The memory I / F 38 creates a new image file on the removable recording medium 38 (the created image file is opened), and the image data stored in the YUV image area 24a and the 2 stored in the audio area 24b. The audio data of the channel is repeatedly read through the memory control circuit 22, and the read image data and audio data are stored in an open image file.

  When the movie button 34mv is operated again, the CPU 30 stops the sound processing circuit 36 and the memory I / F 38. The memory I / F 38 finishes reading data from the YUV image area 24a and the audio area 24b, and closes the open image file. As a result, the moving image that continuously represents the imaging scene and the sound around the imaging scene are recorded in the recording medium 40 in a file format.

  The audio processing circuit 36 is configured as shown in FIG. The L-channel audio data and the R-channel audio data are input to amplitude correction circuits 50L and 50R forming the amplitude correction system 50, respectively. Each of the amplitude correction circuits 50L and 50R corrects the amplitude of the input audio data in accordance with the setting of the control circuit 56, and supplies the corrected audio data to the delay correction system 52. The L channel audio data is input to the delay correction circuit 52L, and the R channel audio data is input to the delay correction circuit 52R. Each of the delay correction circuits 52L and 52R delays the input audio data according to the setting of the control circuit 56, and outputs the delayed audio data to the memory control circuit 22.

  The delay-corrected L-channel sound data and R-channel sound data are also input to FFT analysis circuits 54L and 54R that form an FFT (Fast Fourier Transform) analysis system 52, respectively. Each of the FFT analysis circuits 54L and 54R performs Fourier transform on the input audio data, and gives the analysis result, that is, Nmax (Nmax: integer of 2 or more) frequency components obtained thereby, to the control circuit 56.

For the frequency where the phase difference between the frequency component of the L channel and the frequency component of the R channel is shifted by ½ period (= π) or more, the phase difference between the channels cannot be accurately determined. For this reason, each frequency of the Nmax frequency components needs to satisfy Equation 1.
[Equation 1]
D / V * 2πf <π
D: Distance between microphones 34L and 34R V: Sound velocity f: Frequency

  If the interval D is 20 millimeters and the sound speed is 340 m / sec, all Nmax frequency components correspond to data components having a frequency lower than 8.5 kHz.

  The control circuit 56 controls the settings of the amplitude correction system 50 and the delay correction system 52 based on the frequency component thus given. The control circuit 56 is specifically a DSP (Digital Signal Processor), and executes processing according to the flowcharts shown in FIGS. 4 to 6 for every 1024 samples. The settings of the amplitude correction system 50 and the delay correction system 52 are initialized when the power is turned on. Both the L channel audio data and the R channel audio data correspond to data sampled at a clock frequency of 48 kHz.

  Referring to FIG. 4, in step S1, the FFT analysis result of the L channel audio data is acquired from the FFT analysis circuit 54L, and in step S3, the FFT analysis result of the R channel audio data is acquired from the FFT analysis circuit 54R. When the acquisition is completed, the variable N is set to “1” in step S5.

In step S7, the phase of the Nth frequency component belonging to the L channel is calculated as “Ph_L (N)”, and in step S9, the phase of the Nth frequency component belonging to the R channel is calculated as “Ph_R (N)”. The phase Ph_L (N) is calculated according to Equation 2, and the phase Ph_R (N) is calculated according to Equation 3.
[Equation 2]
Ph_L (N) = atan (real (f_N_L) / image (f_N_L))
atan: arctangent real (f_N_L): real part of the Nth frequency component belonging to the L channel imag (f_N_L): lie part of the Nth frequency component belonging to the L channel [Equation 3]
Ph_L (R) = atan (real (f_N_R) / image (f_N_R))
real (f_N_R): real part of the Nth frequency component belonging to the R channel imag (f_N_R): lie part of the Nth frequency component belonging to the R channel

  In step S11, the difference absolute value of the phases Ph_L (N) and Ph_R (N) calculated in this way is calculated as “ΔPh (N)”. In step S13, it is determined whether or not the variable N has reached the maximum value Nmax. If the determination result is NO, the variable N is incremented in step S15 and then the process returns to step S7. If the determination result is YES, the process proceeds to step S17.

In step S17, the variable N is set to “1” again. In step S19, it is determined whether or not the difference absolute value ΔPh (N) is less than the threshold value TH1, and in step S21, it is determined whether or not the difference absolute value ΔPh (N) is greater than or equal to the threshold value TH2. Here, the threshold value TH1 is calculated according to Equation 4, and the threshold value TH2 is calculated according to Equation 5. In addition, “85 °” in Equation 4 corresponds to a limit of an angle that can be regarded as an audio signal from the front direction that can be detected with the same amplitude. Equation 5 represents the phase difference when coming from the direction of the extended line of the straight line connecting the microphones, and shows the theoretical maximum phase difference.
[Equation 4]
TH1 = D * cos85 ° / V * 2πf
[Equation 5]
TH2 = D * cos0 ° / V * 2πf

  If the decision result in the step S19 is YES, the level of the Nth frequency component belonging to the L channel is saved in a step S23, and the level of the Nth frequency component belonging to the R channel is saved in a step S25. If the determination result in the step S21 is YES, the difference absolute value ΔPh (N) is stored in a step S27.

  If the processing of step S25 or S27 is completed, or if the determination results of steps S19 and S21 are both NO, it is determined in step S29 whether the variable N has reached the maximum value Nmax. If the determination result is NO, the variable N is incremented in step S31 and then the process returns to step S19. If the determination result is YES, the process proceeds to step S33.

  In step S33, the average value of the levels saved by the process of step S23 is calculated as “LVav_L”. In step S35, the average value of the levels saved by the process of step S25 is calculated as “LVav_R”. In step S37, the settings of the amplitude correction circuits 50L and 50R are adjusted so that the absolute difference value between the calculated average values LVav_L and LVav_R is suppressed.

  In step S39, the average value of the absolute differences stored by the process of step S27 is calculated as “ΔPhav”. In step S41, the settings of the delay correction circuits 52L and 52R are adjusted so that the calculated average value ΔPhav is suppressed. When the adjustment is completed, the processing for the target 1024 samples is terminated.

  When an audio signal is incident from the front as shown in FIG. 7, the L-channel data component and the R-channel data component belonging to a certain frequency are the waveform shown in FIG. 8A and the waveform shown in FIG. 8B, respectively. Draw. Also, as shown in FIG. 9, when the audio signal is incident obliquely from the right front, the L-channel data component and the R-channel data component belonging to a certain frequency have the waveforms shown in FIG. Draw the waveform shown in Further, when the audio signal is incident from the right side as shown in FIG. 11, the L-channel data component and the R-channel data component belonging to a certain frequency are respectively shown in the waveform shown in FIG. 12A and FIG. 12B. Draw the waveform shown.

  Here, in the waveform shown by the solid line in FIG. 8B, FIG. 10B, or FIG. 12B, the characteristic of the amplitude correction circuit 50R matches the characteristic of the amplitude correction circuit 50L, and the waveform of the delay correction circuit 52R. This represents a change in the data component of the R channel when the characteristic matches the characteristic of the delay correction circuit 52L.

  8B, 10B, or 12B, the waveform of the amplitude correction circuit 50R is different from that of the amplitude correction circuit 50L, and the waveform of the delay correction circuit 52R is different. This represents a change in the data component of the R channel when the characteristic matches the characteristic of the delay correction circuit 52L.

  8B, 10B, or 12B, the waveform of the amplitude correction circuit 50R matches the characteristic of the amplitude correction circuit 50L, and the characteristic of the delay correction circuit 52R. Represents a change in the data component of the R channel when the characteristic differs from the characteristic of the delay correction circuit 52L.

  Differences in characteristics between the amplitude correction circuit 50L and the amplitude correction circuit 50R occur due to variations in the performance of components. Differences in characteristics between the delay correction circuit 52L and the delay correction circuit 52R also occur due to variations in component performance.

  Further, since the incident angle of the audio signal is different between FIGS. 7, 9 and 11, the phase of the waveform in FIG. 10B is ahead of the phase of the waveform shown in FIG. The phase of the waveform in (B) is ahead of the phase of the waveform shown in FIG.

  Based on this, the determination result in step S19 shown in FIG. 5 indicates YES for the sound signal incident as shown in FIG. 7 or FIG. 9, while NO for the sound signal incident as shown in FIG. Indicates. On the other hand, the determination result of step S21 shown in FIG. 5 shows NO for the audio signal incident as shown in FIG. 7 or FIG. 9, while YES for the audio signal entered as shown in FIG. Indicates.

  Therefore, the setting of the amplitude correction system 50 is adjusted so that the difference between the waveform level shown in FIG. 8A and the waveform level shown in FIG. 8B is suppressed, or in FIG. Adjustment is made so that the difference between the waveform level shown and the waveform level shown in FIG. On the other hand, the setting of the delay correction system 52 is adjusted so that the difference between the phase of the waveform shown in FIG. 12A and the phase of the waveform shown in FIG.

  As can be seen from the above description, the control circuit 56 classifies each of the two channels of audio data acquired in parallel into Nmax frequency components respectively corresponding to Nmax (Nmax: an integer of 2 or more) frequencies. (S1 to S5, S13 to S15), the phase difference between the two frequency components corresponding to each of the Nmax frequencies is detected as difference absolute values ΔPh (1) to ΔPh (Nmax) (S7 to S11). The control circuit 56 also identifies a difference absolute value below the threshold TH1 from the detected difference absolute values ΔPh (1) to ΔPh (Nmax) (S19), and defines two frequencies defining the identified difference absolute value The setting of the amplitude correction system 50 is adjusted so that the level difference between the components is suppressed (S23 to S25, S33 to S37). Here, the threshold value TH1 indicates a value based on the distance between the microphones 34L and 34R and the upper limit of the allowable incident angle of sound.

  The control circuit 56 also specifies a difference absolute value indicating a value equal to or greater than the threshold value TH2 from the Nmax difference absolute values ΔPh (1) to ΔPh (Nmax) (S21, S27), and sets the specified difference absolute value. The setting of the delay correction system 52 is adjusted so that the corresponding phase difference is suppressed (S39 to S41). Here, the threshold value TH2 also indicates a value based on the distance between the microphones 34L and 34R.

  As described above, the amplitude of the audio data is adjusted so that the level difference between the two frequency components defining the absolute difference value below the threshold value TH1 is suppressed. In other words, the amplitudes of the entire area of the M audio signals are adjusted so that the level difference between the audio components incident at an angle lower than the angle corresponding to the threshold value TH1 is suppressed. Further, the delay amount of the audio data is adjusted so that the phase difference corresponding to the absolute difference value equal to or greater than the threshold value TH2 is suppressed. In other words, by setting the threshold value TH2 as the theoretical maximum threshold value determined from the microphone interval, a phase difference exceeding the maximum phase difference caused by the quality variation is suppressed. By repeating this suppression processing, the phase difference falls within the maximum threshold value for sound coming from any direction. As a result, the delay due to the quality variation is corrected, and the quality of the audio signal is improved.

  Although the audio processing circuit 36 of this embodiment is configured as shown in FIG. 3, the audio processing circuit 36 may be configured as shown in FIG. 13 or FIG.

  According to FIG. 13, the FFT analysis system 54 is provided before the amplitude correction system 50, and the inverse FFT system 58 is provided after the delay correction system 52. The L channel audio data is provided to the amplitude correction circuit 50L via the FFT analysis circuit 54L, and the R channel audio data is provided to the amplitude correction circuit 50R via the FFT analysis circuit 54R. Further, the control circuit 56 executes the processes shown in FIGS. 4 to 6 based on the output of the delay correction system 52. Further, the output of the delay correction circuit 52L is output to the memory control circuit 22 after being returned to the audio data by the inverse FFT circuit 58L, and the output of the delay correction circuit 52R is returned to the audio data by the inverse FFT circuit 58R. It is output toward the memory control circuit 22.

  According to FIG. 14, a phase / amplitude correction filter 60L is provided instead of the amplitude correction circuit 50L and the delay correction circuit 52L, and a phase / amplitude correction filter 60R is provided instead of the amplitude correction circuit 50R and the delay correction circuit 52R. Each of the phase / amplitude correction filters 60L and 60R corresponds to a weighting filter for controlling directivity or enhancing stereo feeling. At this time, in steps S37 and S41 shown in FIG. 6, the settings of the weighting filters 60L and 60R are adjusted.

  In this embodiment, a DSP is adopted as the control circuit 56 shown in FIG. 3, but a CPU may be adopted instead of the DSP. In this case, a control program corresponding to the processing shown in FIGS. 4 to 6 is stored in a flash memory (not shown).

10 ... Digital camera 16 ... Imager 24 ... SDRAM
30 ... CPU
36 ... Audio processing circuit 50 ... Amplitude correction system 52 ... Delay correction system 54 ... FFT analysis system 56 ... Control circuit

Claims (3)

Classifying means for classifying each of M (M: integer greater than or equal to 2) audio signals acquired in parallel into N signal components respectively corresponding to N (N: integer greater than or equal to 2) frequencies;
Detecting means for detecting a phase difference between M signal components corresponding to each of the N frequencies with reference to an output of the classifying means;
A specifying means for specifying a phase difference indicating a value equal to or greater than a threshold value from among the N phase differences detected by the detecting means; and
Adjusting means for adjusting a delay amount of the M audio signals so that the phase difference specified by the specifying means is suppressed;
The threshold value indicates a value based on a distance between M microphones that respectively acquire the M audio signals.
Audio processing device. The specifying unit specifies L (L: integer) phase differences indicating a value equal to or greater than the threshold value from among the N phase differences detected by the detection unit,
  Calculating means for calculating an average value of the L phase differences specified by the specifying means;
  The adjusting unit adjusts the delay amount of the M audio signals so that the average value of the phase differences calculated by the calculating unit is suppressed.
The speech processing apparatus according to claim 1.   The threshold value represents a phase difference in a case where the M audio signals arrive from a direction on an extension line of a straight line connecting the M microphones.
The speech processing apparatus according to claim 1.

Images