JP5246134B2 - Signal processing apparatus and imaging apparatus - Google Patents

Description

  The present invention relates to a signal processing device and an imaging device that reduce a noise signal included in an audio signal.

  As a method of reducing the component (noise component) based on the noise sound from the sound signal in which the target sound and the noise sound are mixed, the noise sound is estimated from the acquired sound signal, and the estimated noise sound signal (hereinafter referred to as the noise signal) In general, the noise signal is subtracted from the audio signal (see, for example, Patent Document 1).

JP 2005-195955 A

  In such a method, when the noise sound is acquired in advance or when the noise sound is a periodic sound, the size of the noise sound and the timing including the noise sound can be easily estimated. The noise signal can be appropriately reduced. However, when the sound generated when various mechanisms in the device are driven (hereinafter referred to as operation sound) is a noise sound, the above-mentioned operation sound is generated at irregular timings, so that the noise sound included in the audio signal is not generated. It is difficult to estimate. Furthermore, in the case of an audio signal whose target sound volume changes, it is more difficult to estimate this noise signal. Therefore, the audio signal after subtracting the noise signal contains a noise component called musical noise. End up.

  The present invention provides a signal processing device and an imaging device that can effectively reduce the influence of an operation sound included in an audio signal even when the size of the audio signal itself changes. Objective.

In order to solve the above-described problem, the signal processing apparatus of the present invention provides a plurality of first audio signals obtained by dividing an audio signal indicated by a time function at predetermined time intervals, and second audio signals indicated by a frequency function. a signal conversion means for converting each of the plurality of said second audio signal, and a second audio signal dynamic Sakuon are mixed, and a voice signal indicative of the operation sound, the effect of the operation sound a third signal calculating means for obtaining an audio signal with reduced out of the plurality of said second audio signal, operation signal for executing the operation to emit pre Symbol operation sound obtained when the non-output, the Ratio calculating means for obtaining a ratio between the magnitude of the reference signal and the magnitude of the second audio signal mixed with the operation sound, using the second audio signal not mixed with the operation sound as a reference signal; The audio signal multiplied by the ratio The frequency characteristic of the third audio signal is compared with the frequency characteristic of the reference signal multiplied by the ratio when the comparison result is not within a predetermined range. Correction means for correcting; and signal inverse conversion means for inversely converting the third audio signal corrected by the correction means from the audio signal indicated by the frequency function to the audio signal indicated by the time function. It is characterized by having.

In addition, the correction means may include the at least one frequency band when a value of a frequency spectrum of at least one frequency band is not in a predetermined range between the third audio signal and the reference signal multiplied by the ratio. The third audio signal is corrected so that the value of the frequency spectrum becomes closer to the reference signal multiplied by the ratio.
Further, the correcting means multiplies the third audio signal and the ratio by a ratio between a frequency spectrum value of a specific frequency band and a value obtained from the frequency spectrum of at least one frequency band among all frequency bands. Obtained from the reference signal, and the ratio obtained from the third audio signal is approximately equal to the ratio obtained from the reference signal multiplied by the ratio. The value is corrected.

Further, the correction means is a ratio between the value of the frequency spectrum of the specific frequency band and either the sum of the values of the frequency spectrum of the entire frequency band or the average value of the frequency spectrum of the entire frequency band. Are obtained from the third audio signal and the reference signal multiplied by the ratio, respectively.
  In addition, the correction means includes a sum of a frequency spectrum value of the specific frequency band and a frequency spectrum value of the specific frequency band and a frequency band in the vicinity of the specific frequency band, or the specific frequency band and A ratio with any one of average values of frequency spectra in a frequency band in the vicinity of the specific frequency band is obtained from the third audio signal and the reference signal multiplied by the ratio.

  In addition, an imaging apparatus of the present invention includes an imaging unit that acquires an image signal, a sound collection unit that acquires an audio signal in synchronization with the acquisition of the image signal by the imaging unit, and any one of the signal processing devices described above. And a storage means for storing the image signal acquired by the imaging means and the audio signal subjected to the signal processing by the signal processing device.

  According to the present invention, it is possible to effectively reduce the influence of the operation sound superimposed on the audio signal even when the magnitude of the audio signal itself changes.

It is a functional block diagram which shows the structure of a digital camera. It is a functional block diagram which shows the structure of a signal processing apparatus. It is a figure which shows the relationship between the audio | voice signal acquired, a window function, a frame division | segmentation, and AF operation signal. It is a figure which shows the flow of the signal processing in a signal processing apparatus.

  FIG. 1 is a functional block diagram showing the configuration of the digital camera 10. As is well known, the digital camera 10 photoelectrically converts the subject light captured by the imaging optical system 15 by the imaging device 16 and acquires image data from the electrical signal (image signal) after the photoelectric conversion.

  The imaging optical system 15 includes a plurality of lenses. The lenses constituting the imaging optical system 15 move along the optical axis L by driving the lens driving unit 18 when changing the zoom magnification or adjusting the focus. The imaging optical system 15 is provided with a diaphragm 19. This diaphragm 19 changes the amount of subject light incident on the image sensor 16 by changing the size of the aperture. The aperture of the aperture 19 is changed by the aperture driver 20 so that the aperture value is set in advance.

  The shutter 21 is disposed between the imaging optical system 15 and the imaging element 16. The shutter 21 can be switched between an open state in which the imaging element 16 is irradiated with subject light captured via the imaging optical system 15 and a light shielding state in which the subject light is shielded. At the time of photographing, the shutter 21 is once held in a light shielding state and then switched to an open state. Then, when a preset time elapses after the shutter 21 is switched to the open state, the shutter 21 is switched to the light shielding state again. Note that switching between the light shielding state and the open state of the shutter 21 is performed by the shutter driving unit 22.

  The image pickup device 16 is configured by, for example, a charge coupled device (CCD), a complementary metal-oxide semiconductor (CMOS), or the like. The image sensor 16 receives subject light captured by the imaging optical system 15, converts the received light amount into signal charges (photoelectric conversion), and accumulates the converted signal charges. Thereafter, the signal charge accumulated in the image sensor 16 is output to an AFE (Analog Front End) circuit 25.

  The AFE circuit 25 includes an AGC circuit, a CDS circuit, and an A / D conversion circuit (not shown). The AFE circuit 25 performs processing such as gain control and noise removal on the input image signal. After these processes, the AFE circuit 25 converts the analog image signal into a digital image signal. The digital image signals are collected for each frame and recorded in the image memory 30. The drive timing of the image sensor 16 and the AFE circuit 25 is controlled by a timing generator (not shown).

  The image processing circuit 35 performs image processing such as white balance processing, color interpolation processing, contour compensation processing, and gamma processing on the image signal stored in the image memory 30. After these image processes, the image processing circuit 35 performs a format process for compression using a storage system such as the JPEG system. The image processing circuit 35 performs encoding processing and decoding processing on the image data. Reference numeral 37 denotes a recording I / F.

  The LCD 38 displays an image acquired by the digital camera 10, an image stored in the storage medium 36, a through image captured in a shooting standby state, and a setting image for setting or changing a shooting condition. Etc. are displayed. Examples of the image acquired by the digital camera 10 and the image stored in the storage medium 36 include moving images in addition to still images. The LCD 38 is controlled by a display control circuit 39. The speaker 40 outputs sound or the like associated with the moving image when the moving image is displayed on the LCD 38. The sound output control in the speaker 40 is executed by the sound control circuit 41.

  The sound collection unit 43 is constituted by, for example, a microphone, and acquires, for example, sound during moving image shooting or recording. The audio signal acquired by the sound collection unit 43 is recorded in the audio memory 44.

  The signal processing device 45 performs processing for reducing the noise sound included in the audio signal with respect to the audio signal acquired by the sound collection unit 43. Note that the noise sound includes an operation sound that is generated when each mechanism provided in the digital camera 10 is driven at the time of moving image recording or recording, and an operation sound that is operated when various mechanisms are driven. Can be mentioned. In addition, the signal processing device 45 compresses and encodes the acquired audio signal (including the audio signal subjected to the noise noise reduction process described above) and the audio signal subjected to the compression encoding process. The process which decodes is performed.

  The CPU 50 comprehensively controls each unit of the digital camera 10 by executing a control program (not shown) stored in the built-in memory 51. Examples of the control in the CPU 50 include control based on the operation of the release button 52 and control based on the operation of the setting operation unit 53. Examples of the control based on the operation of the release button 52 include known AE processing, AF processing, imaging processing, and the like. Control based on the operation of the setting operation unit 53 includes processing such as initial setting and setting of shooting conditions.

  In addition, the CPU 50 performs a process of writing image data acquired at the time of shooting to the storage medium 36. For example, when still image shooting is performed, the CPU 50 converts the image data that has been encoded by the image processing circuit 35 into one file (model information of the digital camera 10, shooting information at the time of shooting, and the like). (Still image files) are collectively written in the storage medium 36. Similarly, when moving image shooting is performed, each frame image data obtained by moving image shooting is subjected to encoding processing by the image processing circuit 35. Therefore, each frame image subjected to these encoding processing is processed. The data and the audio data subjected to the signal processing by the signal processing circuit 45 are collectively written in the storage medium 36 as one file (moving image file) together with the model information of the digital camera 10 and shooting information at the time of shooting. .

  Next, the configuration of the signal processing device 45 described above will be described with reference to the functional block diagram of FIG. As shown in FIG. 2, the signal processing device 45 includes a frequency conversion unit 61, a signal calculation unit 62, a storage unit 63, a ratio calculation unit 64, a signal correction unit 65, and a frequency inverse conversion unit 66.

  The frequency conversion unit 61 converts the audio signal acquired by the sound collection unit 43 from a signal (time domain signal) indicated by a time function to a signal (frequency domain signal) indicated by a frequency function. First, the frequency converter 61 determines a window width in a window function described later. After determining the window width in this window function, the frequency converting unit 61 converts the input audio signal so that the number of samples per frame is, for example, 1024 when the determined window width is one frame. To divide.

  Next, the frequency conversion unit 61 multiplies a window function such as a Hanning window while shifting by 0.5 frames, and then performs a Fourier transform process on the audio signal to which the window function is applied. As is well known, since a window function called a Hanning window is a function having both end values of 0 and a median value of 1, a signal obtained by multiplying the window functions is a signal in which the central portion is emphasized. For this reason, when a signal such as vibration that changes with time is shifted and analyzed for each frame, it is difficult to capture a characteristic part. For this reason, a characteristic portion of the signal is detected by performing an analysis that is overlapped by shifting by 0.5 frame. By performing these processes, the acquired audio signal is converted from a time domain signal to a frequency domain signal for each frame while shifting by 0.5 frame. The sound signal subjected to these processes is output to the signal calculation unit 62 and the signal correction unit 65.

  The signal calculation unit 62 performs a process of reducing noise noise included in the acquired audio signal. As described above, the audio signal acquired by the sound collection unit 43 includes a signal in which a target sound (target sound) and a noise sound (operation sound) are mixed. The signal calculation unit 62 subtracts a signal (noise signal) based on the noise sound from the sound signal output from the frequency conversion unit 61 to reduce the noise sound included in the sound signal. This noise signal consists of a frequency domain signal corresponding to the same frame width as the audio signal output from the frequency converter 61. The noise signal is stored in the storage unit 63 in advance.

  In the present embodiment, the noise signal is stored in the storage unit 63 in advance, and the noise signal stored in the storage unit 63 is subtracted from the audio signal for each frame output from the frequency conversion unit 61. The present invention is not limited to this, and the noise signal multiplied by the coefficient can be subtracted from the audio signal of each frame output from the frequency converter 61.

  In addition, the noise signal is not stored in the storage unit 63 in advance, but the noise signal is acquired using a conventional noise estimation method, or the acquired noise signal or a signal obtained by multiplying the noise signal by a coefficient is acquired. May be subtracted from the audio signal for each frame output from the frequency converter 61.

When the audio signal of only the target sound is used as the reference signal, the ratio calculation unit 64 obtains the ratio of the input audio signal size for each frame with respect to the reference signal size. As described above, the audio signal for each frame is converted by the frequency converter 61 into a frequency domain signal, in other words, a signal indicated by the relationship between the frequency band and the value of the frequency spectrum (the intensity of the sound in that frequency band). . The ratio calculator 64 calculates the sum A0 of the frequency spectrum values in each frequency band in the reference signal and the sum A1 of the frequency spectrum values in each frequency band in the audio signal. After calculating these values, determining the ratio factor B by B = A1 / A0. That is, by determining the ratio B, it can be determined whether or not the size of the acquired audio signal has changed. The ratio calculation unit 64 outputs the obtained ratio B and the reference signal to the signal correction unit 65.

  The signal correction unit 65 performs correction on the audio signal whose noise component has been reduced by the signal calculation unit 62 (hereinafter, the audio signal that has been subjected to reduction processing). First, the signal correction unit 65 multiplies the reference signal by the ratio B obtained by the ratio calculation unit 64. Thereafter, the signal correction unit 65 corrects the audio signal that has been subjected to the reduction process, based on the reference signal multiplied by the ratio B. The corrected audio signal subjected to this correction is output to the frequency inverse transform unit 66. The frequency inverse transform unit 66 inversely transforms the corrected audio signal from the frequency domain signal to the time domain signal. The audio signal inversely converted to the time domain signal is written into the audio memory 44.

  Next, the flow of signal processing in the signal processing device 45 will be described with reference to FIGS. 3 and 4. As shown in FIG. 3, the sound acquired by the sound collection unit 43 becomes a periodic signal in a short time of about several tens of ms. As described above, when an audio signal is input, the frequency converting unit 61 sets a window width in the window function and then performs frame division.

  As described above, the frequency conversion unit 61 multiplies a window function such as a Hanning window while shifting by 0.5 frames, and then performs a Fourier transform process on the audio signal to which the window function is applied. For this reason, when the above-described processing is performed on the audio signal input to the frequency converting unit 61, the audio signal having the area indicated by reference numeral 71 and the area indicated by reference numeral 72 as one frame, the area indicated by reference numeral 72, and An audio signal for each frame is generated in the order of an audio signal having the region indicated by reference numeral 73 as one frame, and so on, and is output to the signal calculation unit 62. When an audio signal for each frame is input, the signal calculation unit 62 reads a noise signal composed of an operation sound recorded in the recording unit 63 and subtracts the noise signal from the audio signal for each frame.

  For example, when AF (autofocus) processing is executed while sound is acquired by the sound collection unit 43, an AF operation signal is output. In response to the output of the AF operation signal, the lens driving unit 18 is driven, and the lens constituting the imaging optical system 15 moves in the optical axis L direction. When the lens driving unit 18 is driven and the lens constituting the imaging optical system 15 is moved, the operation sound is generated. For this reason, the sound signal acquired by the sound collection unit 43 is a sound signal in which the target sound and the noise sound are mixed. For example, when the timing at which the AF drive signal is output is within the region indicated by reference numeral 76, it can be estimated that the noise component is superimposed on the audio signal in the subsequent region.

  For example, an audio signal on which a noise component is superimposed can be expressed by the following equation (1).

x (t) = s (t) + n (t) (1)
Here, x (t) is an audio signal acquired by the sound collection unit 43, s (t) is an audio signal of a target sound, and n (t) is a noise signal such as an operation sound. These signals are shown as a time function.

  By performing the Fourier transform described above, the audio signal in which the target sound and the noise sound are mixed is converted from the signal x (t) indicated by the time function into the signal X (f) indicated by the frequency function. Note that f indicates a frequency.

  Here, if the sound signal of the target sound is Se (f), Se (f) of the sound signal of the target sound is expressed by the following equation (2).

| Se (f) | = | X (f) | -α | Ne (f) | (2)
Note that Ne (f) is a noise signal, and α is a subtraction coefficient. The value of α depends on the magnitude of the noise signal to be subtracted when obtaining the audio signal of only the target sound using the above-described equation (2). In some cases, musical noise or the like may be artificially superimposed on the audio signal after subtraction. For this reason, as a value of (alpha), it is desirable to use the value of 0.5-4.

  FIG. 4 shows the frequency spectrum of each signal (81, 82, 83, 83 ′, 84, 84 ′). In the graph of each frequency spectrum, the horizontal axis is the frequency band, and the vertical axis is the sound intensity (hereinafter referred to as the sound intensity). , Also referred to as “frequency spectrum value”).

  Hereinafter, as illustrated in FIG. 4, the acquired audio signal is denoted by reference numeral 81 and the noise signal is denoted by reference numeral 82. The signal calculation unit 62 reads out the noise signal 82 stored in the storage unit 63 and then converts the frequency spectrums 82a to 82h in the noise signal 82 from the frequency spectra 81a to 81h in the Fourier-transformed audio signal 81, respectively. Subtract for each band. By this subtraction process, a reduced audio signal 83 with a reduced noise component is generated.

Then, ratio calculation unit 64, the sum of the magnitude of the frequency spectrum 84a~84h of each frequency band in the reference signal 84 A0, the frequency spectrum 81 a to 81 h of each frequency band in及beauty speech signal 81 magnitude Is obtained. After obtaining these sums A0 and A1, the signal correction unit 65 calculates the ratio B = A1 / A0.

  The signal correction unit 65 generates a signal (symbol 84 ′) obtained by multiplying the ratio B calculated by the ratio calculation unit 64 by the spectrum value of each frequency band of the reference signal 84. Then, the signal correction unit 65 compares the reference signal 84 ′ multiplied by the ratio B with the reduced audio signal 83. In this comparison, the values of the frequency spectra 83a to 83h of the audio signal 83 subjected to the reduction process are the same as those of the frequency spectra 84′a to 84′h of the reference signal 84 ′ multiplied by the ratio B in the same frequency band. It is compared whether or not the value is within a predetermined error range. That is, when the frequency spectrum value of the reduced audio signal 83 is S1, and the frequency spectrum value S0 of the reference signal 84 multiplied by the ratio B is satisfied, S0−K ≦ S1 ≦ S0 + K (K is an error). It is determined whether or not. Note that the value of the error K may be set as appropriate. FIG. 4 shows a case where the values of the frequency spectrum 83f and the frequency spectrum 83g among the frequency spectra 83a to 83h of the audio signal 83 subjected to the reduction process do not satisfy the above-described equation.

In this case, the signal correction unit 65 obtains a ratio between the sum of the values of the frequency spectra 83a to 83h in the entire frequency band of the audio signal 83 subjected to the reduction process and the values of the target frequency spectrum 83f and the frequency spectrum 83g. That is, the ratio (first intensity ratio) of the sum of the values of the frequency spectra 83a to 83h and the value of the frequency spectrum 83f (the first intensity ratio), and the ratio of the sum of the values of the frequency spectra 83a to 83h and the value of the frequency spectrum 83g (second intensity). Ratio). The signal correction unit 65 also adds the sum of the values of the frequency spectra 84′a to 84′h in the entire frequency band of the reference signal 84 ′ multiplied by the ratio B, the target frequency spectrum 84′f, and the frequency spectrum 84. Find the ratio of 'g values. That is, the ratio (third intensity ratio) between the sum of the values of the frequency spectra 84′a to 84′h and the value of the frequency spectrum 84′f, and the sum of the values of the frequency spectra 84′a to 84′h and the frequency spectrum. A ratio with respect to a value of 84′g (fourth intensity ratio) is obtained. Then, the signal correction unit 65 corrects the value of the target frequency spectrum 83f so that the first intensity ratio becomes equal to the third intensity ratio. Further, the value of the target frequency spectrum 83g is corrected so that the second intensity ratio becomes equal to the fourth intensity ratio. In this way, a corrected audio signal 83 ′ is generated. In FIG. 4, correction is performed to increase the value of the frequency spectrum 83f of the audio signal 83 that has been reduced and to decrease the value of the frequency spectrum 83g, thereby generating an audio signal 83 ′.

  The corrected audio signal 83 ′ is converted from a signal represented by a frequency function into a signal represented by a time function by inverse Fourier transform or the like by the frequency inverse transform unit 65. Since the signal correction described above is executed for each frame shifted by 0.5 frames, the audio signal indicated by the time function subjected to the inverse Fourier transform process by the frequency inverse transform unit 65 is 0.5 After being connected while shifting the frame, it is written in the audio memory 44.

  In this way, by subtracting the component of the operation sound that occurs when the internal mechanism of the camera is driven from the sound signal, after generating the sound signal that reduces the effect of the operation sound, the effect of this operation sound is reduced. The sound signal is corrected based on the sound signal of only the target sound. At this time, it is possible to determine whether or not the size of the audio signal has changed by obtaining the ratio between the size of the reference signal and the size of the audio signal, and multiply the audio signal by the ratio. Thus, it is possible to cope with the case where the size of the audio signal changes. Further, since the audio signal with the reduced noise signal can be corrected based on the reference signal multiplied by the ratio, it is possible to prevent the occurrence of musical noise that occurs in the audio signal with the reduced noise signal. As described above, in the present embodiment, it is possible to appropriately reduce the noise sound included in the audio signal even if the audio signal has a variable signal magnitude.

  In the present embodiment, the timing for outputting the AF drive signal is given as a requirement for the generation of the operation sound as the noise sound, but in addition to this, it is output when an operation unit such as a zoom button is operated. The timing at which the operation signal of the operation unit provided in the digital camera, such as the operation signal, is output, and the timing at which the start signal of the camera shake correction process is output in the case of a digital camera having the camera shake correction function. It is done.

  In the present embodiment, the intensity ratio is calculated from the ratio between the sum of the frequency spectrum values of the entire frequency band of the audio signal and the value of the frequency spectrum to be corrected, but the present invention is not limited to this. For example, the intensity ratio may be obtained from the ratio between the average value of the frequency spectrum values of the entire frequency band of the audio signal and the value of the frequency spectrum to be corrected. Also, it is not the sum of the frequency spectrum values of all frequency bands of the audio signal, but the sum of the frequency spectrum values to be corrected and the frequency spectrum values of the frequency bands near the frequency band of the frequency spectrum, or the average By using the value, the above-described intensity ratio can also be obtained. Furthermore, instead of obtaining the above-described sum and average value, a ratio between the frequency spectrum value of the frequency band to be corrected and the frequency spectrum value of the frequency band adjacent to the frequency spectrum may be used.

  In the present embodiment, the value of the frequency spectrum to be corrected is corrected so that the intensity ratio obtained from the reduced audio signal is equal to the intensity ratio obtained from the reference signal 84 multiplied by the ratio B. However, the present invention is not limited to this, and the intensity ratio obtained from the reduced audio signal is within the error range of the intensity ratio obtained from the reference signal 84 multiplied by the ratio B. The frequency spectrum to be corrected may be corrected. In this case, the error is calculated based on the value K used when, for example, the frequency spectrum value of the reduced audio signal 83 is compared with the frequency spectrum value of the reference signal 84 multiplied by the ratio B. A value can be used.

  In addition, it is not necessary to calculate the intensity ratio based on the reduced audio signal 83 or the intensity ratio based on the reference signal 84 multiplied by the ratio B. For example, the value of the frequency spectrum of the reduced audio signal is the ratio. You may correct | amend so that it may be contained in the error range of the value of the frequency spectrum of the reference signal 84 by which B was multiplied.

  In the present embodiment, the audio signal acquired at the time of moving image shooting is described as an example. However, the present invention is not limited to this, and can be applied to the case of acquiring only an audio signal, for example. That is, any electronic device having a recording function can be applied. In addition, although a digital camera has been described as an example of a device that performs moving image shooting, a mobile terminal such as a mobile phone or a PDA may be used. Furthermore, the program may be a program that allows a computer to execute the functions of the signal processing device shown in FIG. In this case, the program is preferably stored in a computer-readable storage medium such as a memory card, an optical disk, or a magnetic disk.

  DESCRIPTION OF SYMBOLS 10 ... Digital camera, 15 ... Imaging optical system, 16 ... Imaging device, 18 ... Lens drive part, 19 ... Aperture, 20 ... Aperture drive part, 21 ... Shutter, 22 ... Shutter drive part, 36 ... Storage medium, 43 ... Collection Sound part, 44... Voice memory, 45... Signal processing device, 61... Frequency conversion part, 62... Signal calculation part, 63 ... recording part, 64 ... ratio calculation part, 65 ... signal correction part, 66.

Claims (6)

Signal conversion means for converting a plurality of first audio signals obtained by dividing an audio signal represented by a time function at predetermined time intervals into second audio signals represented by a frequency function;
A signal for obtaining a third audio signal in which the influence of the operation sound is reduced from the second audio signal in which the operation sound is mixed and the audio signal indicating the operation sound among the plurality of second audio signals. A calculation means;
Among the plurality of second audio signals, a second audio signal that is obtained when the operation signal for executing the operation for generating the operation sound is not output and that is not mixed with the operation sound is used as a reference signal. A ratio calculating means for obtaining a ratio between the magnitude of the reference signal and the magnitude of the second audio signal mixed with the operation sound;
The frequency characteristic of the reference signal multiplied by the ratio is compared when the frequency characteristic of the third audio signal and the reference signal multiplied by the ratio are compared, and the comparison result is not in a predetermined range. Correction means for correcting the frequency characteristics of the third audio signal so as to be close to
Signal inverse conversion means for inversely converting the third sound signal corrected by the correction means from the sound signal indicated by the frequency function to the sound signal indicated by the time function;
A signal processing apparatus comprising: The signal processing device according to claim 1,
When the frequency spectrum value of at least one frequency band is not within a predetermined range between the third audio signal and the reference signal multiplied by the ratio, the correction unit is configured to output the frequency of the at least one frequency band. The signal processing apparatus, wherein the third audio signal is corrected so that a spectrum value approaches the reference signal multiplied by the ratio. The signal processing device according to claim 1 or 2,
The correction means is configured to multiply a ratio of a frequency spectrum value of a specific frequency band and a value obtained from a frequency spectrum of at least one frequency band of all frequency bands by the third audio signal and the ratio. The value of the frequency spectrum of the specific frequency band is determined such that the ratio obtained from the reference signal and obtained from the third audio signal is approximately equal to the ratio obtained from the reference signal multiplied by the ratio. A signal processing device characterized by correcting. The signal processing device according to claim 3.
The correction means is a ratio of the value of the frequency spectrum of the specific frequency band and the sum of the values of the frequency spectrum of the entire frequency band or the average value of the frequency spectrum of the entire frequency band, A signal processing apparatus, wherein the signal processing apparatus obtains each of the third audio signal and the reference signal multiplied by the ratio. The signal processing device according to claim 3.
The correction means includes a sum of a frequency spectrum value of the specific frequency band and a frequency spectrum value of the specific frequency band and a frequency band near the specific frequency band, or the specific frequency band and the specific frequency band. A signal processing characterized in that a ratio with any one of average values of frequency spectrums in a frequency band in the vicinity of the frequency band is obtained from the third audio signal and the reference signal multiplied by the ratio. apparatus. Imaging means for obtaining an image signal;
  Sound collection means for acquiring an audio signal in synchronization with acquisition of the image signal by the imaging means;
  A signal processing device according to any one of claims 1 to 5,
  Storage means for storing the image signal acquired by the imaging means and the audio signal subjected to signal processing by the signal processing device;
  An imaging apparatus comprising:

Images