JP2011253126A - Voice signal processor and control method thereof - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce the possibility to subtract spectrum excessively when reducing noises contained in a voice in spectrum subtraction.SOLUTION: The voice signal processor having plural drive units comprises: a generating means that generates voice signals from voices around the voice signal processor; and subtraction means that applies spectrum subtraction to frequency spectrum of the voice signals to reduce drive sounds accompanying the operation of the plural drive units and contained in the voice signals. When the plural drive units are in operation simultaneously, the subtraction means processes the frequency components of the voice signals to subtracts a maximum value of each frequency components of plural drive sounds accompanying the operation of the plural drive units.

Description

  The present invention relates to an audio signal processing device and a control method thereof.

  2. Description of the Related Art In recent years, a photographing apparatus capable of photographing a moving image such as a camera is known as an audio signal processing apparatus. When shooting a moving image with sound using such a photographing device, it is desired to suppress the influence of driving sound (noise) due to driving of a driving unit inside the photographing device.

  As a technique for suppressing the influence of driving sound, Patent Document 1 is known. Patent Document 1 discloses a method of removing drive noise generated by a photographing operation by spectral subtraction. According to the method of Patent Document 1, when a plurality of driving units are driven at the same time, each spectrum obtained in advance representing the driving sound of each driving unit is subtracted from the spectrum of the sound input from the microphone.

JP 2006-279185 A

  However, the phases of the drive sounds of the drive units do not necessarily match, and the drive sounds may interfere with each other and cancel each other out. Therefore, when each spectrum representing the drive sound of each drive unit is simply subtracted, more spectrum than necessary may be subtracted from the spectrum of the sound input from the microphone, which may cause distortion in the sound.

  The present invention has been made in view of such a situation, and an object of the present invention is to reduce the possibility of subtracting more spectrum than necessary when reducing noise contained in speech by spectral subtraction. To do.

  In order to solve the above-described problem, the first aspect of the present invention is an audio signal processing device including a plurality of driving units, the generating means for generating an audio signal from audio around the audio signal processing device, and the audio Subtracting means for applying spectral subtraction to the frequency spectrum of the audio signal in order to reduce driving sound accompanying the operation of the plurality of driving units included in the signal, the subtracting means comprising the plurality of subtracting means When the drive unit is operating simultaneously, the maximum value of the frequency component of each of the plurality of drive sounds accompanying the operation of each of the plurality of drive units is subtracted from the frequency component of the audio signal for each frequency. An audio signal processing apparatus is provided.

  Other features of the present invention will become more apparent from the accompanying drawings and the following description of the preferred embodiments.

  With the above configuration, according to the present invention, it is possible to reduce the possibility of subtracting more spectrum than necessary when reducing noise contained in speech by spectral subtraction.

(A) is a perspective view of the imaging device 1 equipped with the lens 2, which is an example of the audio signal processing device of the present invention, and (b) is a cross-sectional view of the imaging device 1 and the lens 2. 2 is a block diagram illustrating an electrical configuration of the photographing apparatus 1 and the lens 2. FIG. 3 is a block diagram showing a detailed configuration of an audio processing circuit 26. FIG. 3 is a block diagram illustrating a detailed configuration of a noise processing unit 44. FIG. FIG. 3 is an operation sequence diagram of the photographing apparatus 1. It is a figure which shows typically the waveform which extracted a certain frequency component of drive sound, (a) shows the case where the phase of two drive sounds is equal, (b) shows the phase shift of two drive sounds. It is a figure which shows the case. FIG. 3 is a schematic diagram of a spectrum of an audio signal generated by a microphone 7 of the photographing apparatus 1. It is a figure which shows the relationship between the gain ratio between several drive sounds, and the gain of a synthetic wave. It is a schematic diagram of the synthetic spectrum of two driving sounds, (a) is a case where a synthetic spectrum is generated by simple addition, (b) is a case where a synthetic spectrum is obtained according to Equation 5, and (c) is two driving It is a figure which shows the case where the larger one of the frequency components of a sound is made into a synthetic spectrum.

  Embodiments of the present invention will be described below with reference to the accompanying drawings. The technical scope of the present invention is determined by the claims, and is not limited by the following individual embodiments. In addition, not all combinations of features described in the embodiments are essential to the present invention.

[First Embodiment]
FIG. 1A is a perspective view of a photographing apparatus 1 equipped with a lens 2, which is an example of an audio signal processing apparatus of the present invention. The photographing apparatus 1 includes a microphone 7 (see FIG. 1B), and can acquire and record sound simultaneously with image acquisition. In FIG. 1A, 4 indicates the optical axis of the lens 2, 30 indicates a release button, 31 indicates an operation button, and 32 indicates an opening for the microphone 7.

  FIG. 1B is a cross-sectional view of the photographing device 1 and the lens 2. 1B, the same or similar components as those in FIG. 1A are denoted by the same reference numerals, and description thereof is omitted. In FIG. 1B, 3 is an imaging optical system, 5 is a lens barrel, 6 is an imaging element as imaging means for photoelectrically converting incident light from the imaging optical system 3 to generate image data, and 7. Indicates a microphone, and 8 indicates a display device provided on the back of the photographing apparatus 1. Reference numeral 9 denotes a drive unit for adjusting the photographing optical system 3, reference numeral 10 denotes a contact for connecting the photographing apparatus 1 and the lens 2, reference numeral 11 denotes a so-called quick return mirror mechanism, and reference numeral 12 denotes a focus detection unit including an AF sensor. , 14 indicates a blur sensor. The shake sensor 14 includes, for example, an acceleration sensor. In this embodiment, the shake sensor 14 is referred to as a shake sensor 14 in order to detect that the camera vibrates due to the shake of the user's hand.

  As shown in FIG. 1A, the photographing apparatus 1 is provided with a plurality of microphone openings 32 at locations that are not projected on the cross section of FIG. However, in order to clarify the presence of the microphone 7 and the opening 32, these are schematically shown in FIG.

  Here, a still image shooting operation will be described. The photographing apparatus 1 performs focus detection and exposure detection using the lens 2, the focus detection unit 12, and an exposure detection unit (not shown). The photographing apparatus 1 also drives a part of the photographing optical system 3 to form an image near the image sensor 6 by adjusting the photographing optical system 3. Further, the photographing apparatus 1 operates the diaphragm so as to achieve appropriate exposure. Detailed operation will be described later with reference to the block diagram of FIG. The photographing apparatus 1 further sets various shooting conditions according to the operation of the operation button 31 by the user, obtains information on the subject from the image sensor 6 in synchronization with the operation of the release button 30, and the memory 24 (see FIG. 2). To record.

  Next, a moving image shooting operation will be described. Prior to shooting a moving image, an image of the image sensor 6 is displayed on the display device 8 by pressing a live view button (not shown). In addition, displaying an image acquired by the image sensor 6 on the display device 8 in real time is referred to as “live view”. The photographing apparatus 1 synchronizes with the operation of a moving image photographing button (not shown), obtains subject information from the image sensor 6 at a set frame rate, obtains an audio signal from the microphone 7, and synchronizes them to store the memory. Record to 24. When adjustment of the photographing optical system 3 is necessary during moving image photographing, the photographing apparatus 1 appropriately adjusts the photographing optical system 3 using the drive unit 9. The photographing apparatus 1 ends the photographing in synchronization with the operation of the moving image photographing button. Also, the photographing apparatus 1 can take a still image at any time by operating the release button 30 even during moving image shooting.

  FIG. 2 is a block diagram showing an electrical configuration of the photographing apparatus 1 and the lens 2. The photographing apparatus 1 equipped with the lens 2 has an imaging system, an image processing system, an audio processing system, a recording / reproducing system, and a control system. The imaging system includes the photographing optical system 3 and the imaging element 6, the image processing system includes the A / D converter 20 and the image processing circuit 21, and the sound processing system includes the microphone 7 and the sound processing circuit 26, and recording. The reproduction system includes a recording processing circuit 23 and a memory 24. The control system includes a camera system control circuit 25, a focus detection unit 12 including an AF sensor, an exposure detection unit 13 including an AE sensor, a blur sensor 14, an operation detection circuit 27, a lens system control circuit 28, a release button 30, and a drive unit 9. . The drive unit 9 includes a focus lens drive unit 9a, a shake correction drive unit 9b, and an aperture drive unit 9c, and drives the focus lens, the shake correction lens, and the stop included in the lens 2, respectively. The blur correction driving unit 9b performs driving to move the lens so that the optical image is not shaken by the vibration of the imaging device 1 detected by the blur sensor 14.

  As an example of the drive unit, the focus lens drive unit 9a, the shake correction drive unit 9b, and the aperture drive unit 9c are described, but the present embodiment is not limited to this. The photographing apparatus 1 equipped with the lens 2 as the audio signal processing apparatus of the present invention may be provided with a plurality of driving sections, regardless of the driving section.

  The imaging system is an optical processing system that forms an image of light from an object on the imaging surface of the imaging device 6 via the imaging optical system 3. During a preliminary shooting operation such as aiming, a part of the light beam is also guided to the focus detection unit 12 through a mirror provided in the quick return mirror mechanism 11. Further, as will be described later, the photographing optical system 3 is appropriately adjusted by the control system, so that an appropriate amount of object light is exposed to the image sensor 6 and a subject image is formed in the vicinity of the image sensor 6.

  The image processing circuit 21 is a signal processing circuit that processes image data received from the image sensor 6 via the A / D converter 20, and includes a white balance circuit, a gamma correction circuit, and an interpolation calculation for increasing the resolution by interpolation calculation. Circuit and the like.

  The audio processing system generates an audio signal for recording by performing an appropriate process on the audio signal generated by the microphone 7 by the audio processing circuit 26. The audio signal for recording is compressed and linked with an image by a recording processing unit to be described later at the time of moving image shooting.

  The recording processing circuit 23 outputs image data to the memory 24 and generates and stores an image to be output to the display unit 22. In addition, the recording processing circuit 23 compresses images, moving images, sounds, and the like using a predetermined method.

  The camera system control circuit 25 generates and outputs a timing signal at the time of imaging. The focus detection unit 12 detects the focus state of the photographing apparatus 1. The exposure detection unit 13 detects the luminance of the subject by processing the signal from the image sensor 6. The lens system control circuit 28 adjusts the photographing optical system 3 by appropriately driving the lens 2 in accordance with the signal from the camera system control circuit 25.

  The control system controls the imaging system, the image processing system, and the recording / reproducing system in response to an external operation. For example, in still image shooting, the operation detection circuit 27 detects pressing of the release button 30 to control driving of the image sensor 6, operation of the image processing circuit 21, compression processing of the recording processing circuit 23, and the like. The display unit 22 controls the state of each segment of the information display device that displays information on an optical finder, a liquid crystal monitor, or the like.

  The adjustment operation of the photographing optical system 3 by the control system will be described. A focus detection unit 12 and an exposure detection unit 13 are connected to the camera system control circuit 25. In still image shooting, an appropriate focus position and aperture position are obtained based on these signals. The camera system control circuit 25 issues a command to the lens system control circuit 28 via the electrical contact 10, and the lens system control circuit 28 appropriately controls the focus lens driving unit 9a and the aperture driving unit 9c. On the other hand, in moving image shooting, the camera system control circuit 25 causes the focus lens drive unit 9a to finely move the focus lens and analyze the signal from the image sensor 6 to obtain the focus position from the signal contrast. Further, the camera system control circuit 25 obtains the aperture position from the signal level of the image sensor 6.

  Further, the blur sensor 14 is connected to the lens system control circuit 28, and in the mode in which camera shake correction is performed in still image shooting, the blur correction drive unit 9b is appropriately controlled based on the signal of the blur sensor 14. On the other hand, in a mode in which camera shake correction is performed in moving image shooting, it is possible to drive the blur correction drive unit 9b similarly to still image shooting, and so-called changing the readout position of the image sensor 6 based on the signal from the blur sensor 14. Electronic vibration isolation can also be performed.

  Here, consider shooting with sound recording such as moving image shooting. In photographing with sound recording, a sound (hereinafter referred to as “mechanical driving sound”) accompanying driving of the actuators of the photographing device 1 and the lens 2 is an unnecessary sound and becomes noise. In this specification, “noise” refers to this “mechanical drive sound”, not background noise such as white noise.

  FIG. 3 is a block diagram showing a detailed configuration of the audio processing circuit 26. 3, the same or similar components as those in FIG. 2 are denoted by the same reference numerals, and description thereof is omitted. In FIG. 3, reference numeral 41 denotes a gain adjustment unit, 42 denotes a filter, 43 denotes an A / D converter, 44 denotes a noise processing unit, and 45 denotes a filter.

  The microphone 7 generates a sound signal from the sound around the photographing apparatus 1, and the generated sound signal is supplied to the gain adjustment unit 41. The gain adjusting unit 41 adjusts the signal level of the microphone 7 so that the dynamic range of the A / D converter 43 can be fully utilized. That is, when the signal level of the microphone 7 is low, the gain is increased to amplify the signal, and when the signal level of the microphone 7 is high, the gain is decreased to prevent saturation. The filter 42 is configured by a low-pass filter having an appropriate cutoff frequency considering the sampling frequency of the A / D converter 43. When the microphone 7 is in the vicinity of an element that emits a specific frequency, the filter 42 may include an appropriate notch filter in addition to the low-pass filter. The A / D converter 43 converts the analog signal processed by the gain adjusting unit 41 and the filter 42 into a digital signal.

  The noise processing unit 44 functions as a subtracting unit, and applies a spectral subtraction process (SS process) by a spectral subtraction method (SS method) to the audio signal. SS is an abbreviation for Spectral Subtraction.

  Here, the SS process will be described. A noise spectrum (in this specification, a spectrum obtained by Fourier transforming noise is referred to as a noise spectrum) is prepared in advance, and the noise spectrum is subtracted from the spectrum of the acquired speech signal. In the present embodiment, it is assumed that the noise spectrum is identified in advance and stored in a storage unit such as the memory 24 in the photographing apparatus 1. In this specification, “spectrum” means “frequency spectrum” unless otherwise specified.

  The SS method assumes that noise is additively added to the subject sound.

It becomes. Here, x (t) represents the acquired sound, s (t) represents the subject sound, n (t) represents noise, and t represents time. When the number 1 is Fourier transformed,

It becomes. However, X (ω), S (ω), and N (ω) are the Fourier transforms of x (t), s (t), and n (t), respectively, and ω is the frequency. In the photographing apparatus 1, an audio signal is divided into frames by applying an appropriate window function and sequentially processed. Here, for simplification of description, the description will be given focusing on a specific frame. As is apparent from Equation 2, N (ω) may be subtracted from X (ω) to obtain S (ω). Therefore, the following equation is given.

Where N ′ (ω) is an estimated value of N (ω), S ′ (ω) is an estimated value of S (ω) obtained using N ′ (ω), and β is a flooring coefficient. ∠ indicates an operation for obtaining the argument of a complex number. As can be seen from Equation 3, the spectrum is subtracted using a noise spectrum given in advance, but the value of X (ω) is used as it is for the phase. Further, the flooring coefficient β is a coefficient introduced in order to suppress voice distortion by the SS method (β = 0 in the original SS method). As shown in Equation 3, in the SS method, it is assumed that noise acts additively. However, in practice, there are cases where a plurality of noises weaken each other in the audio signal as a result of phase inversion and addition. For this reason, the value obtained by subtracting N ′ (ω) from X (ω) may be negative. Therefore, in the SS method, when it is smaller than β, processing is performed so as to be β.

  Finally, S ′ (ω) is subjected to inverse Fourier transform to obtain s ′ (t), which is used as the speech after SS processing.

  In order to properly perform the SS method, it is important that the noise spectrum estimated in advance is close to the actually acquired noise. When the noise spectrum is not appropriate, the result is undersubtraction or oversubtraction. In the case of under-subtraction, noise removal is insufficient. On the other hand, in the case of excessive subtraction, many solitary peaks are observed in the spectral region, and annoying noise called so-called musical noise is generated.

  The SS process described above is schematically shown in FIG. In FIG. 4, the same or similar components as those in FIG. In FIG. 4, FFT 44 a indicates fast Fourier transform processing including window function processing, IFFT 44 c indicates fast inverse Fourier transform processing, and S ′ (ω) estimation 44 b indicates processing of Equation 3. As can be seen from FIG. 4, the SS method is applicable to a single channel signal (monaural sound). On the other hand, it is necessary to give N ′ (ω) by some method in advance. In the example of FIG. 4, the noise spectrum generation unit 44d generates an estimated noise spectrum N ′ (ω).

  The camera system control circuit 25 acquires a pre-estimated noise spectrum from the memory 24 and stores it in the temporary storage unit 44f in the noise spectrum generation unit 44d. The noise spectrum generator 44d appropriately selects and synthesizes a noise spectrum (using the spectrum synthesizer 44e) according to a control signal from the camera system control circuit 25, and generates an estimated noise spectrum N ′ (ω). Hereinafter, the generation process of generating the estimated noise spectrum N ′ (ω) will be described in detail together with the operation of the imaging apparatus 1.

  The camera system control circuit 25 acquires a noise spectrum corresponding to the drive unit 9 in the lens 2 from the memory 24 when the lens 2 is attached to the photographing apparatus 1 or when the power is turned on, and a noise spectrum generation unit 44d. In the example shown in FIG. 2, the noise spectra of the focus lens driving unit 9a, the blur correction driving unit 9b, and the aperture driving unit 9c are supplied to the noise spectrum generating unit 44d. The noise spectrum generation unit 44d stores the supplied noise spectrum in the temporary storage unit 44f.

  Next, user operations before and during moving image shooting and processing inside the shooting device 1 will be described with reference to FIG. When performing moving image shooting, the user operates the moving image shooting button after pressing a live view button (not shown) as described above. At this time, the camera system control circuit 25 instructs start of blur correction and performs a focus adjustment operation. In addition, even during moving image shooting, the focus adjustment operation can be performed again by an appropriate operation by the user. For example, the focus adjustment operation is performed by half-pressing the release button 30 (the release button 30 is a two-stage push switch, and pushing to the first stage is called “half-press”). Each sequence of user operation, blur correction driving, and focus driving at this time is shown in FIG. In FIG. 5, the horizontal axis represents time, and the vertical axis represents the state of each sequence. In FIG. 5, t1 indicates a time during which focus drive is performed for the focus adjustment operation.

  During moving image shooting, the noise processing unit 44 receives the driving state of each driving unit from the camera system control circuit 25, and performs noise processing as shown in FIG. That is, when only blur correction driving is performed, the noise processing unit 44 processes only noise (driving sound) accompanying the operation of the blur correction driving unit 9b. On the other hand, when the blur correction drive and the focus drive are performed at the same time (that is, when a plurality of drive units are operating simultaneously), the noise processing unit 44 performs noise and focus associated with the operation of the blur correction drive unit 9b. Noise associated with the operation of the lens driving unit 9a is also processed. As described above, in the photographing apparatus 1 equipped with the lens 2, a single drive unit (noise source) operates, or a plurality of drive units (noise source) operate simultaneously.

  Consider the waveform of the drive sound when multiple drive units are operating simultaneously. FIG. 6 is a diagram schematically showing a waveform obtained by extracting a certain frequency component of the drive sound. In FIG. 6, the horizontal axis represents time, and the vertical axis represents sound pressure (sound volume). FIG. 6A is a diagram illustrating a case where the phases are aligned, and FIG. 6B is a diagram illustrating a case where the phase is shifted by an appropriate amount. In FIG. 6, G1 represents the one-sided amplitude of the sound pressure generated by the driving unit 1, G2 represents the one-sided amplitude of the sound pressure generated by the driving unit 2, and Gs represents the one-sided amplitude of the combined waveform of the driving unit 1 and the driving unit 2. , Respectively. As can be seen from FIG. 6A, when the phases of the driving unit 1 and the driving unit 2 coincide with each other, the waves are strengthened and Gs = G1 + G2 is established. At this time, the driving sound can be removed by the conventional spectral subtraction. However, in general, in each frequency component, it is a very rare case that the phases of the drive unit 1 and the drive unit 2 coincide with each other. In many cases, when attention is paid to a certain frequency component, there is a shift in the phases of the drive unit 1 and the drive unit 2 as shown in FIG. That is, Gs <G1 + G2 is a general state.

  FIG. 7 is a schematic diagram of a spectrum of an audio signal generated by the microphone 7 of the photographing apparatus 1. FIG. 7 is described with the driving unit 1 as the focal lens driving unit 9a and the driving unit 2 as the blur correction driving unit 9b. When attention is paid to the specific frequency f1, when the phase of the driving sound of the focus lens driving unit 9a and the driving sound of the blur correction driving unit 9b match, by subtracting (G1 + G2) from the frequency component Sa at the frequency f1. A component S1 excluding noise is obtained. However, when the driving sound of the focus lens driving unit 9a and the driving sound of the shake correction driving unit 9b are out of phase, the actual component Gs ′ of the driving sound included in Sa is Gs ′ <G1 + G2. Accordingly, assuming that the actual speech component excluding noise is S1 ', when (G1 + G2) is subtracted from Sa, excessive subtraction of (S1'-S1) occurs. In order to prevent such excessive subtraction, it is necessary to calculate true Gs (indicated as Gs ′ in FIG. 7) according to the following equation.

Here, θ is the phase difference between the drive unit 1 (focus lens drive unit 9a) and the drive unit 2 (blur correction drive unit 9b). As is clear from Equation 4, when θ = 0, Gs = G1 + G2 holds as described above. When θ = π, Gs = | G1-G2 |. Generally between these. As described above, since the information on the drive sound generated by the drive unit 9 of the photographing apparatus 1 with the lens 2 attached is known in advance, G1 and G2 in Equation 4 are known. Therefore, in order to obtain Gs, it is sufficient to know θ.

  However, since it is difficult to obtain θ, the expected value of Equation 4 is used. For example, the following formula may be calculated.

Thereby, the one-side amplitude Gs of the synthesized signal in the time domain of the plurality of driving sounds is obtained as an expected value regarding the phase difference between the plurality of driving sounds.

  The noise spectrum generation unit 44d of the imaging device 1 acquires the expected value Gs for solving Equation 5 for each frequency (may be calculated in advance by another computer or the like and stored in the memory 24 or the like). . In S ′ (ω) estimation 44b, the expected value Gs is subtracted from the frequency component of the audio signal generated by the microphone 7 for each frequency. As a result, it is possible to reduce the possibility of subtracting more spectra than necessary while reducing noise.

  In addition, a lookup table can be used to obtain the expected value Gs. A specific method will be described with reference to FIG. FIG. 8 is a diagram showing how Gs changes depending on the ratio of G1 and G2 when the larger gain of G1 and G2 is 1. As shown in FIG. The horizontal axis in FIG. 8 represents the smaller gain when the larger one of G1 and G2 is set to 1. The vertical axis in FIG. 8 is the ratio of Gs to G1 at that time (= Gs / G1).

  That is, in the look-up table of FIG. 8, the ratio between frequency components of a plurality of driving sounds is input (horizontal axis), and the magnification of the expected value Gs for the maximum frequency component is output (vertical axis). This lookup table is generated by calculating in advance the magnification that is the output.

  In order to obtain the expected value Gs, for example, G2 / G1 is obtained when G1> G2. The ratio of Gs to G1 at this time is read from the graph of FIG. Since Gs = G1 * (Gs / G1), the expected value Gs can be obtained by multiplying G1 by a predetermined magnification (the magnification read from FIG. 8). In this way, the expected value Gs can be obtained while reducing the amount of calculation.

  As yet another method, Gs may subtract only the larger one of G1 and G2 (that is, the maximum value of the frequency components of each of the plurality of driving sounds) from the frequency component of the audio signal. Expressed as an expression

It is. However, in Expression 6, max () is an operator for selecting the maximum value. Equation 6 is an approximation in the case of Equation 4 where G1 >> G2 or G1 << G2. The drive sound generated from the drive unit 1 and the drive unit 2 generally has a spectrum that has a peak at a specific frequency, not white noise. Moreover, since this peak portion occupies a large proportion as noise, it is desired to approximate this portion precisely. G1 >> G2 or G1 << G2 holds at the peak unless the peaks of the drive unit 1 and the drive unit 2 coincide with the gain. For this reason, Equation 6 is a good approximation of the combined waveform of the drive sounds of the drive unit 1 and the drive unit 2. The calculation of Equation 6 can be regarded as a modification of the method using the above-described lookup table. That is, it is equal to the case where the ratio of Gs to the larger of G1 and G2 is always 1. In order to always set to 1, no special storage unit is required. When the maximum value is used, the memory usage of the photographing apparatus 1 can be reduced compared to the case where the lookup table is used, and the consumption of calculation resources is reduced compared to the case where the calculation of Formula 5 is performed. be able to.

  The effect of the present invention will be described with reference to FIG. In FIG. 9, the horizontal axis is frequency (unit is hertz and logarithmic axis), and the vertical axis is gain (unit is decibel). 9A shows an example in which noise is synthesized using simple addition in the spectral domain, FIG. 9B shows an example in which noise is synthesized by Equation 5, and FIG. 9C shows noise by Equation 6. An example of synthesizing is shown. The case where a detailed lookup table is used is as shown in FIG. In addition, when the storage capacity of the lookup table is reduced and roughened, a result about between FIG. 9 (b) and FIG. 9 (c) is obtained (the processing is performed using only the maximum value as shown in Equation 6). Is equivalent to the case where all the output of the look-up table is 1.)

  As described above, when the phase difference between the drive unit 1 and the drive unit 2 is unknown, it seems appropriate to generate the synthesized noise by the equation (5). Therefore, it can be said that the synthesized noise spectrum shown in FIG. As a result, appropriate SS processing is performed and high-quality sound can be obtained.

  In the example in which the noise spectrum is estimated using the addition in the spectral region shown in FIG. 9A, the synthesized noise spectrum is excessively estimated. As a result, there is a possibility that musical noise or the like occurs due to excessive subtraction from the subject sound in the SS processing.

  On the other hand, in the example in which the noise spectrum is estimated using the maximum value shown in FIG. 9C, a result close to FIG. 9B is obtained, and it is expected that the noise is appropriately processed. . In particular, at the peak portion where the noise has a large sound pressure, the same gain as in FIG. 9B is obtained. As a result, appropriate SS processing is performed and high-quality sound can be obtained.

  As described above, according to the present embodiment, the photographing apparatus 1 estimates the composite spectrum of a plurality of driving sounds by calculating the expected value regarding the phase difference and adopting the maximum value, and the audio signal generated by the microphone 7. The estimated synthesized spectrum is subtracted from the spectrum of. As a result, it is possible to reduce the possibility of subtracting more spectrum than necessary when reducing noise contained in speech by spectral subtraction.

[Other Embodiments]
The present invention can also be realized by executing the following processing. That is, software (program) that realizes the functions of the above-described embodiments is supplied to a system or apparatus via a network or various storage media, and a computer (or CPU, MPU, or the like) of the system or apparatus reads the program. It is a process to be executed.

Claims (8)

An audio signal processing device including a plurality of driving units,
Generating means for generating an audio signal from audio around the audio signal processing device;
Subtracting means for applying spectral subtraction to the frequency spectrum of the audio signal in order to reduce driving sound accompanying the operation of the plurality of driving units included in the audio signal;
With
When the plurality of driving units are operating at the same time, the subtracting means, for each frequency, from the frequency component of the audio signal, the maximum frequency component of each of the plurality of driving sounds accompanying the operation of the plurality of driving units An audio signal processing device characterized by subtracting a value. An audio signal processing device including a plurality of driving units,
Generating means for generating an audio signal from audio around the audio signal processing device;
Subtracting means for applying spectral subtraction to the frequency spectrum of the audio signal in order to reduce driving sound accompanying the operation of the plurality of driving units included in the audio signal;
With
The subtracting means corresponds to the frequency components of the plurality of driving sounds associated with the operations of the plurality of driving units from the frequency components of the audio signal for each frequency when the plurality of driving units are operating simultaneously. An audio signal processing apparatus, wherein an expected value related to a phase difference between the plurality of driving sounds is subtracted from one-side amplitude of a synthesized signal obtained by synthesizing signals in a time domain. The subtracting unit is configured to calculate the expected value by multiplying a maximum value of a frequency component of each of the plurality of driving sounds by a predetermined magnification for each frequency,
The subtracting unit is configured to acquire the predetermined magnification from a look-up table that inputs a ratio of frequency components of each of the plurality of driving sounds and outputs a magnification for each frequency. Item 3. The audio signal processing device according to Item 2. Storage means for storing a frequency spectrum of each of the plurality of driving sounds;
The audio signal processing apparatus according to any one of claims 1 to 3, wherein the subtracting unit acquires the frequency component of each of the plurality of driving sounds from the storage unit. Photographic optics,
Imaging means for photoelectrically converting incident light from the imaging optical system to generate image data;
Further comprising
The audio signal processing device according to any one of claims 1 to 4, wherein the plurality of driving units include a driving unit of the photographing optical system. A photographic optical system including a focus lens, a diaphragm, and an image stabilization lens;
Imaging means for photoelectrically converting incident light from the imaging optical system to generate image data;
Further comprising
5. The drive unit of the focus lens, the drive unit of the diaphragm, and the drive unit of the blur correction lens, wherein the plurality of drive units includes the drive unit of the blur correction lens. Audio signal processing device. A method of controlling an audio signal processing device including a plurality of driving units,
A generating step for generating a sound signal from the sound around the sound signal processing device;
A subtracting step in which a subtracting unit applies a spectral subtraction to a frequency spectrum of the audio signal in order to reduce a driving sound accompanying an operation of the plurality of driving units included in the audio signal;
With
In the subtracting step, when the plurality of driving units are operating at the same time, the subtracting unit, for each frequency, from the frequency component of the audio signal, a plurality of driving sounds accompanying each operation of the plurality of driving units A control method characterized by subtracting the maximum value of frequency components. A method of controlling an audio signal processing device including a plurality of driving units,
A generating step for generating a sound signal from the sound around the sound signal processing device;
A subtracting step in which a subtracting unit applies a spectral subtraction to a frequency spectrum of the audio signal in order to reduce a driving sound accompanying an operation of the plurality of driving units included in the audio signal;
With
In the subtracting step, when the plurality of driving units are operating at the same time, the subtracting unit, for each frequency, from the frequency component of the audio signal, a plurality of driving sounds accompanying each operation of the plurality of driving units A control method comprising: subtracting an expected value related to a phase difference between the plurality of driving sounds from one-side amplitude of a synthesized signal obtained by synthesizing signals in a time domain corresponding to a frequency component.