JP2013183355A - Imaging device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an imaging device capable of performing suitable noise reduction processing.SOLUTION: An imaging device 100 of the present invention comprises a sound collection unit 131 for collecting sounds and a drive unit 112 for driving a mechanism unit 111. The imaging device 100 also comprises an extraction unit 133A for extracting information of a specific frequency range from sound information collected by the sound collection unit 131, and a noise point determination unit 133B for determining a noise generation point in a sound signal by comparing the information extracted by the extraction unit 133A with a threshold.

Description

  The present invention relates to an imaging apparatus.

In recent years, an increasing number of imaging devices such as digital cameras are capable of capturing moving images.
Sound is also recorded at the same time as shooting a movie. At this time, drive sound noise of autofocus drive or zoom drive is recorded, and sound quality may be impaired. These drive sound noises may be due to auto focus or zoom mechanism parts, but are often generated by a motor or the like that drives them. These driving noises have a wide frequency spectrum and are harsh.
For this reason, a method has been proposed in which a sound whose generation time is equal to or shorter than a predetermined time and whose power is within a predetermined range is detected and noise reduction processing is performed on the sound (see, for example, Patent Document 1).

JP 2008-77707 A

  However, in the case of the noise reduction processing according to Patent Document 1, it is determined that the driving sound noise is only from the sound information collected by the sound collecting device. For this reason, for example, a sound similar to an impact sound is generated around the shooting location, and even when this sound is collected, it is erroneously detected as drive noise and noise reduction processing is performed. As a result, there is a problem that the sound quality of the target sound is degraded.

  An object of the present invention is to provide an imaging device capable of performing a suitable noise reduction process.

  The present invention solves the above problems by the following means. In addition, in order to make an understanding easy, although the code | symbol corresponding to embodiment of this invention is attached | subjected and demonstrated, it is not limited to this.

The invention according to claim 1 is an imaging apparatus including a sound collecting unit (131) that collects sound and a drive unit (112) that drives the mechanism unit (111), wherein the sound collecting unit ( 131) an extraction unit (133A) for extracting information in a specific frequency range from the audio information collected by the sound, and a noise occurrence location in the audio signal by comparing the information extracted by the extraction unit (133A) with a threshold value. It is an imaging device (100) characterized by including the noise location determination part (133B) which determines.
Invention of Claim 2 is an imaging device of Claim 1, Comprising: The said extraction part (133A) of several frequency range different from the audio | voice information collected by the said sound collection part (131) The information is extracted, and the noise determination unit (133B) compares the plurality of pieces of information extracted by the extraction unit (133A) with threshold values, and determines the noise occurrence location in the audio signal by summing the results. The imaging device (100) characterized by the above.
Invention of Claim 3 is an imaging device of Claim 2, Comprising: The said several extraction part (133A) is a several different frequency range which takes out from the audio | voice information collected by the said sound collection part (131) Is an imaging device (100) characterized in that none of the frequencies is set to be an integral multiple of other frequencies.
Invention of Claim 4 is an imaging device of Claims 1-3, Comprising: The said extraction part (133A) is specified by the spectrum analysis from the audio | voice information collected by the said sound collection part (131) It is an imaging device (100) characterized by taking out the information of the frequency range.
A fifth aspect of the present invention is the imaging apparatus according to the first to third aspects, wherein the extraction unit (133A) is collected by the sound collection unit (131) by a band-pass filter or a high-pass filter. The imaging apparatus (100) is characterized by extracting information in a specific frequency range from the audio information.
Invention of Claim 6 is an imaging device of any one of Claims 1-5, Comprising: The said extraction part (133A) is the rotation speed of the said drive part (112), and the said drive part ( 112) Extracting information in a specific frequency range from the sound information collected by the sound collection unit (133A) using a weighting window having a period corresponding to the number of noise occurrences per rotation in step 112). Device (100).
A seventh aspect of the present invention is the imaging apparatus according to the sixth aspect, wherein the extracting unit (133A) moves the weighting window relative to the audio signal on the time axis to move the collection unit. The imaging device (100) is characterized in that information in a specific frequency range is extracted from audio information collected by the sound unit (131).

  According to the present invention, it is possible to provide an imaging device capable of performing a suitable noise reduction process.

It is a block diagram which shows the structure of the camera which is one Embodiment of this invention. It is a figure explaining the time change and the detection of timing of audio | voice information at the time of zoom operation | movement. It is a figure which shows typically the difference in the spectrum by the presence or absence of a drive sound noise. It is a figure explaining the time change and timing detection of audio | voice information at the time of zoom operation of 2nd Embodiment. It is a figure explaining the time change and timing detection of audio | voice information at the time of zoom operation of 3rd Embodiment.

(First embodiment)
Embodiments of the present invention will be described below with reference to the drawings.
FIG. 1 is a block diagram illustrating a configuration of a camera 100 according to an embodiment of the present invention.
As shown in FIG. 1, the camera 100 includes a lens barrel 110 and a camera body 160.

The lens barrel 110 includes an imaging optical system including a zoom lens 111, a zoom drive motor 112 that drives the zoom lens 111, and a zoom encoder 113.
The zoom lens 111 is driven to move in the optical axis direction by the zoom drive motor 112 to change the focal length of the lens barrel 110 (imaging optical system).

The zoom drive motor 112 is controlled by a drive control signal input from the control unit 150 and drives the zoom lens 111 to move. Note that the zoom drive motor 112 in this embodiment is a two-pole, three-slot DC motor.
The zoom encoder 113 detects information related to the rotation (rotation angle) of the zoom drive motor 112 and outputs it to the control unit 150 as control information. The zoom encoder 113 outputs a pulse signal for each predetermined rotation angle of the zoom drive motor 112, for example.

  The camera main body 160 has an image processing unit (not shown) that captures the subject light that has passed through the lens barrel 110 and performs A / D conversion and image processing to generate image data. A sound information processing unit 130 for A / D converting sound information and performing noise reduction processing, a recording unit 140 for recording image data obtained by the image processing unit and an audio signal obtained by the sound information processing unit 130, and a CPU And a control unit 150 that controls each functional unit of the camera 100 and a noise information memory 151 that stores information used by the sound information processing unit 130.

  When the camera 100 is instructed to perform shooting by pressing a shutter button (not shown) by the user, the lens barrel 110 is set according to the imaging conditions (for example, aperture value, exposure value, etc.) set by the control unit 150. The camera unit 160 controls each mechanism unit to perform the photographing operation.

  In other words, the subject image light is converted into an electrical signal by the image sensor, and the image data processed by the image processing unit is recorded (photographed) in the recording unit 140. During moving image shooting, the sound data collected by the sound information processing unit 130 is subjected to noise reduction processing and recorded (recorded) in the recording unit 140 together with the image data.

In addition, the control unit 150 generates a drive control signal for driving the zoom drive motor 112 in response to an operation of an operation member (not shown), and outputs the drive control signal to the zoom drive motor 112 to control the drive. The drive control of the zoom drive motor 112 by the control unit 150 is performed based on the rotation detection signal of the zoom drive motor 112 input from the zoom encoder 113.
For example, the control unit 150 controls the position of the zoom lens 111 by positioning the zoom lens 111 at a mechanical origin and counting the pulse signal of the zoom encoder 113 therefrom.

Next, the sound information processing unit 130 will be described in detail.
The sound information processing unit 130 includes a microphone 131 that is a sound collection device, a sound signal processing unit 132 that processes sound information that has been collected and A / D converted, a noise timing detection unit 133, and a noise reduction processing unit 134. And comprising.
The sound signal processing unit 132 weights the A / D converted microphone sound signal with a window function for each predetermined section, and converts the microphone sound signal for each section into a spectrum expressed in the frequency domain. Then, the spectrum represented in this frequency domain is output to the noise timing detection unit 133 and the noise reduction processing unit 134.

  Here, the sound signal processing unit 132 converts the microphone sound signal to the frequency domain by performing, for example, Fourier transform or fast Fourier transform (FFT) on the microphone sound signal, and each section of the window function. A frequency spectrum corresponding to is calculated. Each section of the window function is a unit (frame) of signal processing and is a section repeated at a constant interval. As the window function, for example, a Hanning window (Hanning window) function can be used.

The noise timing detection unit 133 is caused by the rotation of the zoom drive motor 112 during the operation of the zoom drive motor 112 based on the sound information input from the sound signal processing unit 132 and the operation information of the zoom drive motor 112 input from the zoom encoder 113. Then, the timing of the periodic drive sound noise generated is detected.
Then, the drive timing noise generation timing information is stored in the noise information memory 151 via the control unit 150 described later. Here, the drive sound noise generation timing information is rotation angle information of the zoom drive motor 112 that generates the drive sound noise, and directly relates to the pulse signal output from the zoom encoder 113 and the drive sound noise generation. It is. The noise timing detection unit 133 will be described in detail later.

The noise reduction processing unit 134 performs noise reduction processing on the sound information input from the sound signal processing unit 132.
The noise reduction processing unit 134 in the present embodiment includes the operation information of the zoom drive motor 112 detected by the zoom encoder 113, the generation timing information of the driving sound noise detected by the noise timing detection unit 133 and held in the noise information memory 151. Based on the above, the occurrence location of the drive sound noise on the time axis is specified, and the drive sound noise reduction processing is performed on the occurrence location of the drive sound noise. As a result, the noise reduction processing unit 134 performs a reduction process on periodic drive sound noise caused by the rotation of the zoom drive motor 112 that occurs during zooming (when the zoom drive motor 112 is driven).

The noise reduction processing by the noise reduction processing unit 134 uses a frequency spectrum subtraction method in which a frequency component of a frequency spectrum representing noise determined in advance according to a drive pattern is subtracted from the frequency component of the target spectrum.
Note that the frequency spectrum of noise determined in advance according to the drive pattern is preset in the noise information memory 151 as a set value. Also, the noise frequency spectrum is obtained by the noise reduction processing unit 134 subtracting the frequency spectrum of the section where the driving sound is not generated from the frequency spectrum of the section where the driving sound is generated based on the past microphone sound signal. Thus, the frequency spectrum of the estimated noise of the driving sound may be calculated for each driving pattern.

  In addition to the above, the noise reduction processing by the noise reduction processing unit 134 includes a frequency spectrum in a place where driving sound noise is generated and a frequency spectrum in a place where driving sound noise adjacent thereto is not generated in a frequency band in a high frequency range. In comparison with each other, it may be performed by replacing a frequency component having a high intensity at a place where the driving sound noise is generated with a low intensity of a comparison target.

  Further, the noise reduction processing unit 134 performs, for example, inverse Fourier transform or inverse fast Fourier transform (IFFT) on the frequency spectrum subjected to the noise reduction process, thereby converting the frequency spectrum into the time domain, The recording unit 140 records the sound signal converted into the time domain.

Next, with reference to FIG. 2 and FIG. 3 in addition to FIG. 1 described above, the timing detection of the drive sound noise during the operation of the zoom drive motor 112 by the noise timing detector 133 will be described.
FIGS. 2A and 2B are diagrams for explaining the temporal change and timing detection of audio information during the zoom operation, where FIG. 2A is an audio waveform, FIG. 2B is a diagram obtained by converting it into a spectrum, and FIG. It is explanatory drawing of a detection. 2 (a) and 2 (c), the vertical axis represents intensity (sound pressure), the vertical axis of (b) represents frequency, and the horizontal axis of all drawings represents time. FIG. 3 is a diagram schematically showing the difference in spectrum depending on the presence or absence of driving sound noise.

The noise timing detection unit 133 includes an extraction unit 133A and a noise location determination unit 133B.
The extraction unit 133A extracts specific frequency band information from the audio information.
The noise location determination unit 133B determines a noise occurrence location by comparing the magnitude of the signal extracted by the extraction unit 133A with a threshold value.

In the spectrum shown in FIG. 2 (b), a characteristic pattern with periodicity is seen as indicated by an upward arrow in the figure. This indicates the drive sound noise of the zoom drive motor 112.
That is, the zoom drive motor 112 according to the present embodiment has a configuration of two poles and three slots as described above, and due to its mechanical factors, the drive noise noise of six times which is the least common multiple of 2 and 3 per rotation. Has occurred.
On the other hand, the sound waveform shown in FIG. 2A is a mixture of target sound and drive sound noise (the target sound is superimposed with drive sound noise), and it is difficult to separate and recognize the drive sound noise from this. It is.

Here, FIG. 3 schematically shows the difference in spectrum depending on the presence or absence of driving sound noise. FIG. 3A shows a spectrum where there is driving sound noise, and FIG. 3B shows a spectrum where there is no driving sound noise.
That is, FIG. 3A shows a voice waveform cut out by a Hamming window and converted into a spectrum, and the drive noise is almost in the center. FIG. 3B also shows a voice waveform cut out by a Hamming window and converted into a spectrum, which does not include drive noise.

In the spectrum of the portion without drive noise shown in FIG. 3B, it can be seen that the signal intensity up to several KHz is high, and for example, high components of 10 KHz or higher are hardly detected.
On the other hand, in the spectrum of the portion with driving sound noise shown in FIG. 3A, a high component of 10 KHz or more is detected, and it can be seen that particularly a component near 13 KHz is included.

The extraction unit 133A extracts the high sound portion by paying attention to the high sound portion, and the noise location determination unit 133B measures the intensity of the extracted high temperature portion to identify the generation location of the driving sound noise.
Specifically, in FIG. 2B in which the extraction unit 133A has converted the speech waveform of FIG. 2A into a spectrum, a frequency region (for example, about 13 KHz) indicated by A in the drawing is extracted, and a noise location determination unit 133B is extracted. Identifies the location of the drive noise occurrence based on the spectral intensity.

That is, as shown in FIG. 2 (c), the extraction unit 133A extracts a signal in the 13 KHz frequency band from the spectrum, and the noise location determination unit 133B compares the magnitude with a threshold value set to a predetermined size. The location exceeding the threshold is determined as the drive noise location.
In this way, by converting the audio signal into a spectrum and detecting the intensity of the frequency component of the treble part where the intensity of the normal audio signal is low and the intensity of the drive sound noise is high, the location where the drive noise is generated can be identified. It becomes possible.

Note that the frequency band extracted from the spectrum in order to specify the location where the driving sound noise is generated is not limited to 13 KHz and can be changed as appropriate.
Further, as shown by A and B in FIG. 2B, a plurality of (in the figure, two) frequency band signals are extracted and individually compared with thresholds, and the plurality of judgment results are added together. You may identify the drive noise noise generation | occurrence | production location (by AND of the result of a several frequency band).
In this case, the plurality of frequency bands used for extraction are set so that none of the frequencies has a relationship (overtone) that is an integral multiple of other frequencies. As a result, it is possible to specify the drive noise noise occurrence location with high accuracy.

Also, as shown in FIG. 2, when drive noise is generated periodically, the drive signal (in this example, the control unit 15 sets the rotation speed of the zoom drive motor 112, so the noise cycle can be predicted. If the width (size) of the above-mentioned window (Humming window) is set so that a plurality of noise portions are not included in the window (below the noise period), accurate detection can be performed.
The drive sound noise generation timing information detected by the noise timing detection unit 133 when driving the zoom drive motor 112 (during zooming) is stored in the noise information memory 151 as described above.

  The detection (acquisition) of the generation timing of the driving sound noise by the noise timing detection unit 133 and the storage in the noise information memory 151 are performed by driving the zoom driving motor 112 at the time when the assembly of the camera 100 is completed. This is done by moving 111 to the entire movement range. In addition, the detection (acquisition) of the generation timing of the driving sound noise may be performed as necessary, and the generation timing information of the driving sound noise stored in the noise information memory 151 may be updated.

As described above, the sound information processing unit 130 in this configuration is based on the spectrum intensity of the high-pitched portion of the sound collected by the noise timing detection unit 133 (about 13 KHz obtained by converting the sound waveform into a spectrum). Location of occurrence of driving sound noise during driving (during zooming) 112 (occurrence timing is specified, and the noise reduction processing unit 134 reduces driving sound noise generated during zooming based on the driving noise noise generation timing information. .
As a result, it is possible to accurately perform noise reduction processing on the location where the drive sound noise is generated, the processing speed can be increased by simplifying the noise reduction processing, and deterioration of voice information due to the noise reduction processing can be suppressed.

(Second Embodiment)
Next, a second embodiment of detection of driving noise noise timing during the operation of the zoom drive motor 112 by the noise timing detector 133 shown in FIG. 4 will be described.
4A and 4B are diagrams for explaining time change and timing detection of audio information during the zoom operation of the second embodiment, in which FIG. 4A is an audio waveform, FIG. 4B is a waveform obtained by high-pass filtering the audio waveform, (C) is an explanatory diagram of timing detection. In FIG. 4, the vertical axis represents intensity (sound pressure) and the horizontal axis represents time.

The second embodiment is different from the first embodiment described above in the operation of the noise timing detection unit 133.
In the noise timing detection unit 133 according to the second embodiment, the extraction unit 133A extracts a high-frequency signal having a predetermined frequency or higher by performing a band-pass filter or a high-pass filter process on the audio signal, and the noise location determination unit 133B determines the size. Compared with a predetermined threshold value, a portion exceeding the threshold value is determined as a drive sound noise portion.

That is, the extraction unit 133A passes through the 12 kHz high-pass filter with respect to the collected audio signal shown in FIG. 4A to obtain the waveform shown in FIG. In FIG.4 (b), the low frequency part is cut and the signal of 13 KHz is remarkable. Then, by comparing the signal intensity shown in FIG. 4B with a predetermined intensity threshold by the noise location determination unit 133B, it is possible to specify the location where the driving noise noise is generated, as shown in FIG. 4C. Become.
As described above, in the second embodiment, the location of the drive sound noise can be specified by detecting the strength of the frequency component of the high sound portion where the intensity of the normal sound signal is small and the drive sound noise is large. It becomes possible.

(Third embodiment)
Next, a description will be given of a third embodiment of detection of driving sound noise timing during operation of the zoom drive motor 112 by the noise timing detection unit 133 shown in FIG.
FIGS. 5A and 5B are diagrams for explaining the temporal change and timing detection of audio information during the zoom operation of the third embodiment. FIG. 5A is a waveform obtained by performing high-pass filter processing on the audio waveform, and FIGS. These are explanatory diagrams of timing detection. In FIG. 5A, the vertical axis represents intensity (sound pressure) and the horizontal axis represents time.

The third embodiment is different from the first embodiment described above in the operation of the noise timing detection unit 133.
In the noise timing detection unit 133 according to the third embodiment, the extraction unit 133A passes the 12 kHz high-pass filter for the collected audio signal to obtain the waveform shown in FIG. In FIG. 5A, the low frequency portion is cut, and a signal of 13 KHz is remarkable.

  Then, for the waveform of FIG. 5A, the noise location determination unit 133B uses a weighting window in which the weighting period as shown in FIG. 5B is substantially the same as the noise period (drive of the zoom drive motor 112). The period can be determined by the signal), and the relative position of the weighting window with respect to the waveform is changed, and the output is detected.

In this example, a weighting window having four weighting peaks (each corresponding to a Hamming window) as shown in FIG. 5B is used.
FIG. 5C shows the change in the output after the weighting window due to the change in the relative position of the waveform and the weighting window.

  In the configuration as in this example, when the weighting center of the weighting window overlaps with the noise position, a large output can be obtained (window position A). Further, when the weighting window position is changed and there is a window at the position B, the output decreases. Further, when the position is changed and the positional relationship is D, a large output can be obtained again (position D).

  Thus, the waveform shown in FIG. 5C is obtained by the relative positional relationship between the noise period and the weighting window period in the audio signal. Using a plurality of noises enables accurate position detection.

  The filtered waveform shown in FIG. 5C is cut out in a rectangular area indicated by a broken line in the figure, and FFT (Fast Fourier Transform: Fast Fourier Transform) is performed to detect the phase (in this case, since three wavelengths are included) (Detects the phase at a frequency component that is three times the fundamental frequency). The noise position is specified based on the phase information and the filter position information (A to D) (when the Cos wave is used as a reference, in this example, the phase is 0 °, and the cutout start position A of the rectangular area indicated by the broken line in the figure is shown. And the noise position match.)

  According to the third embodiment, even if a sudden noise that is not related to the rotation of the zoom drive motor 112 is included in the audio signal, the periodic drive sound noise caused by the rotation of the zoom drive motor 112 is included. Can be detected. As a result, it is possible to accurately identify the location where the drive noise is generated. As a result, good noise reduction is possible.

As described above, this embodiment has the following effects.
In the sound information processing unit 130 of the camera 100 according to the present embodiment, the extraction unit 133A in the noise timing detection unit 133 extracts a high-pitched portion including a lot of noise components in the collected audio signal, and the noise location determination unit 133B By making the determination based on the intensity of the high-pitched sound part, it is possible to identify the location (occurrence timing) of the drive sound noise when the zoom drive motor 112 is driven.
As a result, the noise reduction processing unit 134 can accurately perform noise reduction processing on the location where the driving sound noise occurs during zooming, and the noise reduction processing can be speeded up, and voice information is deteriorated due to the noise reduction processing. Can be suppressed.

(Deformation)
The present invention is not limited to the above-described embodiment, and various modifications and changes as described below are possible, and these are also within the scope of the present invention.
In the above embodiment, the present invention is applied to driving sound noise generated by the zoom driving motor 112 that moves and drives the zoom lens 111. However, the present invention is not limited to this, and can also be applied to other drive sources that generate drive sound noise, such as an AF motor that drives an AF lens.

  In addition, although embodiment and a deformation | transformation form can also be used in combination as appropriate, detailed description is abbreviate | omitted. Further, the present invention is not limited to the embodiment described above.

  DESCRIPTION OF SYMBOLS 100: Camera, 111: Zoom lens, 112: Zoom drive motor, 130: Sound information processing part, 131: Microphone, 132: Sound signal processing part, 133: Noise timing detection part, 133A: Extraction part, 133B: Noise location determination Part 134: Noise reduction processing part

Claims (7)

An imaging apparatus comprising a sound collection unit that collects sound and a drive unit that drives a mechanism unit,
An extraction unit for extracting information in a specific frequency range from the audio information collected by the sound collection unit;
A noise location determination unit that determines a noise occurrence location by the drive unit in the audio signal by comparing the information extracted by the extraction unit with a threshold;
An imaging apparatus comprising: The imaging apparatus according to claim 1,
The extraction unit extracts information of a plurality of different frequency ranges from the audio information collected by the sound collection unit,
The noise part determination unit compares a plurality of pieces of information extracted by the extraction unit with threshold values, and determines a noise occurrence location in the audio signal by summing the results,
An imaging apparatus characterized by the above. The imaging apparatus according to claim 2,
The plurality of different frequency ranges that the extraction unit extracts from the voice information collected by the sound collection unit is set so that any frequency is not an integral multiple of other frequencies,
An imaging apparatus characterized by the above. The imaging device according to claim 1, wherein
The extraction unit extracts information in a specific frequency range by spectrum analysis from the voice information collected by the sound collection unit;
An imaging apparatus characterized by the above. The imaging device according to claim 1, wherein
The extraction unit extracts information of a specific frequency range from the audio information collected by the sound collection unit by a band-pass filter or a high-pass filter;
An imaging apparatus characterized by the above. The imaging apparatus according to any one of claims 1 to 5,
The extraction unit uses a weighting window having a period corresponding to the number of rotations of the driving unit and the number of noise generations per rotation of the driving unit, from the sound information collected by the sound collecting unit, in a specific frequency range. Retrieving information,
An imaging apparatus characterized by the above. The imaging apparatus according to claim 6,
The extraction unit extracts information in a specific frequency range from the audio information collected by the sound collection unit by moving the weighting window relative to the audio signal on a time axis;
An imaging apparatus characterized by the above.