JPH0482381A - Noise reducer - Google Patents

Abstract

PURPOSE:To avoid deterioration of a picture effectively and to improve noise reduction effect by setting a feedback characteristic and a threshold level of a multiplied value to an optimum value based on noise quantity. CONSTITUTION:A digital video signal SDV stored in a frame memory circuit 4 is subject to movement correction and the result is subtracted from a digital video signal SDV of a succeeding frame at a subtractor circuit 8, a difference data DS is generated thereby, a difference data correction circuit 10 corrects a signal level and the result is subtracted from a digital video signal SDV of a succeeding frame at a subtractor circuit 2 thereby suppressing noise component included in the digital video signal SDV. In the threshold level setting mode, a level of a difference correction circuit 10 is set to a value close sequentially to a 0 level from a preset value and when noise is rapidly increased to be a level depending on a noise setting value DELTAX1 or over, the noise is set to just preceding values ta, -ta.

Description

[Detailed description of the invention] The present invention will be explained in the following order.

A: Industrial field of application B: Overview of the invention C: Conventional technology Means for solving the 8 problems to be solved by the invention (Figs. 1 and 7) F Effects (Figs. 1 and 7) G Example (G1) First Example (Figs. 1 to 6) (G2) Second Example
Embodiments (FIGS. 7 to 10) (G3) Other Embodiments H Effects of the Invention A Industrial Application Field The present invention relates to a noise reducer and can be applied to an audio recorder.

For example, the present invention can effectively avoid deterioration of image quality and reduce noise by switching the threshold value in a noise reducer according to the amount of noise.

C. Prior Art Conventionally, noise reducers have been designed to reduce noise by utilizing frame correlation of video signals.

That is, in a still image, noise components can be extracted by detecting difference data for each pixel between frames.

On the other hand, in an image with movement, noise components can be similarly extracted by performing motion correction and detecting difference data.

Therefore, by subtracting the sequentially extracted difference data from the original image data, noise can be reduced.

On the other hand, in general video signals, not only noise components but also information representing changes in the image are extracted as difference data.

Therefore, in the noise reducer, the signal level of the difference data is corrected using a predetermined feedback characteristic and then subtracted from the original image data, thereby reducing noise based on the above-mentioned noise reduction principle.

Problems to be Solved by the Invention In this type of noise reducer, it would be convenient if it were possible to effectively avoid deterioration of image quality and improve the noise reduction effect.

The present invention has been made in consideration of the above points, and aims to propose a noise reducer that can improve the noise reduction effect.

E! II! Means for Solving the Problem In order to solve this problem, in the present invention, difference data D is sequentially generated between frames of the video signal SDV.

and convert the difference data D into a video signal SD with a predetermined feedback characteristic.
By subtracting from v or based on the difference data D3, predetermined multiplication values (1-K), (K) are generated, and the video signal sev is framed using the multiplication values (1-K), (K). In the noise reducer 1140, which reduces the noise of the video signal 5Elv by performing summer processing in between, the noise amount detection means 12 detects the noise amount of the video signal SDV.
Based on the detection results of the noise amount detection means 12, the feedback characteristics or the multiplication values (1-K), thresholds t1 and t of (K) are determined.
Threshold control means 20.22.24.2 for switching l
6.30.31.

Based on the detection result of the F-effect noise amount detection means 12, the feedback characteristic or the multiplication value (1-K), the threshold value ta of (K),
By switching ta, the threshold value can be set to an optimal value and the noise reduction effect can be improved.

G example An embodiment of the present invention will be described in detail below with reference to the drawings.

(G1) First Embodiment In FIG. 1, 1 generally indicates a noise reducer of a video tape recorder, which converts the reproduced video signal into a digital signal (hereinafter referred to as digital video signal) SDV and generates a subtraction circuit 2. give to

The frame memory circuit 4 sequentially stores the digital video signal SDV inputted via the subtraction circuit 2.

The motion correction circuit 32 switches and outputs read address data according to the horizontal and vertical motion vectors D and Dl detected by a motion vector detection circuit (not shown), and thereby outputs the read address data. The digital video signal S□ stored in the digital video signal S□ is motion-corrected and output.

The fiX circuit 8 receives a motion-compensated digital video signal 5.
IIIF is subtracted from the digital video signal So of the following frame, and difference data D3 is output.

As a result, the digital video signal S
The noise component of □ can be extracted.

The differential data correction circuit 10 is composed of a read-only memory circuit, and stores threshold control data D7 and slope control data D.
The values of K and difference data D are used as address data, and preset data is output.

As a result, the difference data correction circuit 10 adjusts the difference data D according to the threshold control data DT and the slope control data D.
, is corrected and output to the subtraction circuit 2, and subtracted from the original digital video signal S0.

That is, the differential data correction circuit 10 has a feedback characteristic as shown in FIG. ), and when the range of thresholds t and -t is exceeded, the signal level of the output signal is suddenly lowered and output.

That is, as shown in FIG. 3, when the distribution of noise is expressed by taking the signal level as the horizontal axis, the noise is normally distributed around the O level.

Therefore, if only the difference data Ds near the 0 level is fed back to the subtraction circuit 2, only the noise component can be selectively subtracted, and deterioration in image quality can be effectively avoided and noise can be reduced.

That is, as shown in FIG. 4, the digital video signal S output from the subtraction circuit 2. At V, as the thresholds t, -t, -ts of the differential data correction circuit 10 are sequentially increased, the amount of noise gradually decreases, and the thresholds t, -t,
When becomes equal to or higher than a predetermined value, the reduction in the amount of noise becomes saturated.

On the other hand, with respect to the image quality of the digital video signal S□, as the threshold values ts and -t are sequentially increased, the image information is correspondingly impaired and the image quality is degraded.

Therefore, by setting the threshold values t, . . . -ta based on the signal level at which the frequency of noise occurrence sharply decreases, deterioration of image quality can be effectively avoided and noise can be efficiently reduced.

In other words, if the amount of noise is detected and the threshold value t, , -ta is set at a signal level at which the frequency of noise occurrence sharply decreases based on the detection result, the threshold value t, , -ta
ta can be set to an optimal value, deterioration of image quality can be effectively avoided and noise can be efficiently reduced.

Furthermore, at this time, as shown in FIG. 5, at the slope K -
If the slope is gradually increased within the range of 1 or less, the amount of noise in the digital video signal 5IIV output from the subtraction circuit 2 will decrease as the slope K increases, and the image quality will follow this. .

Therefore, the slope K can be set to the optimum value by setting it to the value that reduces the amount of noise the most within a practically sufficient range based on the noise amount detection results, thereby effectively avoiding deterioration in image quality. Noise can be reduced efficiently.

Based on this setting principle, the noise reducer 1 detects the amount of noise with the noise amount detection circuit 12, and sets the threshold value 1 to 1 based on the detection result. , -1. and set the slope K to the optimum value.

That is, the noise amount detection circuit 12 detects the digital video signal S! output from the subtraction circuit 2. lv to the comparator circuit 14, which obtains a comparison result at the clock cycle of the digital video signal SOW with reference to the reference level held at the signal level during the vertical blanking period, thereby determining the signal level of the digital video signal SDV. Detect rising and falling edges.

The absolute value conversion circuit 16 converts the output signal of the comparison circuit 14 into an absolute value and outputs the absolute value.

The adder circuit 1 detects the amount of variation in the signal level of the digital video signal SOV during the vertical retrace period by cumulatively adding the output signal of the absolute value converting circuit 16 using a predetermined gate signal 5IIL as a reference. .

In other words, the video signal is maintained at a predetermined signal level during the retrace period, so that during the retrace period,
By detecting fluctuations in the digital video signal S6v, the amount of noise can be detected.

As a result, in the noise amount detection circuit 12 , the addition circuit 18 obtains a noise amount detection result and outputs the detection result to the register circuit 20 .

The register circuit 20 takes in the noise amount detection result based on the count up pulse Up, and outputs the noise amount detection result with a delay of one frame period.

The subtraction circuit 22 obtains difference data between the noise amount detection results output from the noise amount detection circuit 12 and the register circuit 20, and outputs the difference data to the comparison circuit 24.

The comparison circuit 24 switches its operation based on the control signal output from the control circuit 28, and in the threshold setting mode, the noise setting value ΔX input via the selection circuit 26
A count-up pulse U or a count-down pulse Dp is output based on I.

In other words, the comparison circuit 24 determines that the difference data is the noise setting value ΔX
When the difference data is smaller than 1, a count up pulse U is output, whereas when the difference data is greater than or equal to the noise setting value ΔX1, a count down pulse DP is output.

Furthermore, in the amplitude setting mode, the comparison circuit 24 receives the noise setting value x8 output from the selection circuit 26, and when the noise amount detection result is larger than the noise setting value X2,
Outputs count-up pulse U2.

In the threshold setting mode, the counter circuit 30 first calculates the count value! After initialization to O, the count up pulse U is counted up, whereas when the count down pulse D is output, the count value is down counted and then the counting operation is stopped.

The threshold control circuit 31 updates the threshold control data D7 according to the count value of the counter circuit 30 in the threshold setting mode.

That is, the threshold control circuit 31 switches the threshold values ta and tm of the differential data correction circuit 10 so that they approach the 0 level sequentially from a preset value as the count value increases, and generates a countdown pulse. When DP is output, it returns to the previous value and stops the update operation.

As a result, the threshold control circuit 31 sequentially updates the threshold in the direction in which the amount of noise increases, as shown by the broken line in FIG. 4, and the noise suddenly increases beyond the value determined by the noise setting value ΔXI. Then, the threshold value is set to the previous value t, , -ta.

Thereby, the threshold value can be set to an optimal value, and deterioration in image quality can be effectively avoided and the noise reduction effect can be improved.

In fact, as shown in FIG. 6, according to the experimental results, it can be confirmed that changing the threshold value can follow the change in the amount of noise and improve the S/N ratio. By setting the threshold value to the optimum value as shown in FIG. 1, it was possible to effectively avoid deterioration of image quality and efficiently improve the S/N ratio.

In the threshold setting mode, the counter circuit 32 outputs a predetermined count value, whereas in the amplitude setting mode, the counter circuit 32 first initializes the count value to a value of 0, and then counts up a count-up pulse U. .

The slope control circuit 33 outputs slope control data D according to the count value of the counter circuit 32, thereby setting the slope K of the differential data correction circuit 30 to a value proportional to the count value.

As a result, the slope control circuit 33 maintains the slope K of the differential data correction circuit lO at a predetermined value in the threshold setting mode, whereas in the amplitude setting mode, the slope is sequentially adjusted as shown by the broken line in FIG. Update K to a larger value.

At this time, in the comparator circuit 24, when the noise amount detection result is larger than the noise setting value Xt, the count up pulse UP is outputted, so that the noise amount detection result is larger than the noise setting value Xt.
When the amount of noise determined by 2 is reached, the updating operation of the slope K is stopped.

As a result, in the noise reducer 1, the slope K can be set to an optimal value, effectively avoiding deterioration of image quality and improving efficiency (
Noise can be reduced.

Therefore, in addition to the threshold values tm and ta, the slope K is set to an optimal value, and the noise reduction effect can be improved compared to the conventional method.

The control circuit 28 controls the overall operation of the noise reducer 1 and sequentially switches the operation of the noise reducer 1 to a threshold setting mode and an amplitude setting mode at the time of rising.

That is, the control circuit 28 switches and controls the contacts of the selection circuit 35, and during the threshold setting mode and the amplitude setting mode, the digital video signal Sl output from the subtraction circuit 2,
Instead of v, the digital video signal SaW input to the noise reducer 1 is directly output.

As a result, in the control circuit 28, the threshold values ta and Lm are set during the threshold setting mode and the amplitude setting mode.
The output of the digital video signal Slv, which changes unnaturally with the switching operation of the slope K, is controlled to stop.

Thus, in this embodiment, the noise amount detection circuit 12 constitutes noise amount detection means for detecting the noise amount of the video signal S00, whereas the register circuit 20 and the addition circuit 22
, comparison circuit 24, selection circuit 26, counter circuit 30.3
2. The threshold interspecies m circuit 30 determines the threshold value 1.2 of the feedback characteristic based on the detection result of the noise amount detection means. , -1. constitutes a threshold control means for switching.

In the above configuration, the reproduced video signal is converted into a digital video signal S0 and then sequentially stored in the frame memory circuit 4 via the subtraction circuit 2.

The digital video signal S0 stored in the frame memory circuit 4 is motion-compensated and then subtracted from the digital video signal S□ of the following frame in the subtraction circuit 8, thereby generating difference data D3.

After the signal level of the difference data D is corrected in the difference data correction circuit 10, the digital video signal S! of the following frame is converted into the digital video signal S! in the subtraction circuit 2. It is subtracted from lv, thereby suppressing the noise component contained in the digital video signal SHv.

Furthermore, the digital video signal S output from the subtraction circuit 2
0, the noise amount detection circuit 12 detects fluctuations in the signal level based on the reference level held at the signal level during the vertical blanking period, and obtains a cumulative addition result of the detection results during the vertical blanking period. It will be done.

As a result, in the noise amount detection circuit 12, the subtraction circuit 2
It is possible to detect the amount of noise in the digital video signal S0 output from the digital video signal S0.

On the other hand, when the noise reducer 1 starts up, the operation mode is sequentially switched to the threshold setting mode and the amplitude setting mode based on the control signal output from the control circuit 28.

In the threshold setting mode, the differential data correction circuit 1
When the 0 threshold value is sequentially switched from a preset value to a value approaching the O level, and the noise suddenly increases beyond the value determined by the noise setting value ΔXl, the value immediately before that, t,...
-ta.

As a result, the threshold value can be set to an optimal value just before the noise suddenly increases, effectively avoiding deterioration of image quality and improving the noise reduction effect.

When switching to the amplitude setting mode in this state, the slope K of the differential data correction circuit IO is updated to a larger value in sequence, and when the noise amount detection result becomes smaller than the noise setting value X (when it becomes smaller, the updating operation is stopped).

As a result, at the slope K, the noise setting value X8
It is set to the optimum value that reduces the amount of noise the most within a practically sufficient range determined by , and thus it is possible to effectively avoid deterioration of image quality and improve the noise reduction effect.

According to the above configuration, the threshold and slope of the differential data correction circuit 0 are set to optimal values based on the noise amount detection results, thereby effectively avoiding deterioration of image quality and efficiently reducing noise. Therefore, the noise reduction effect can be improved compared to the conventional method.

(G2) Second Embodiment In FIG. 7, in which parts corresponding to those in FIG. By performing arithmetic processing between the digital video signals S9. reduce noise.

That is, as shown in FIG. 8, the constant setting circuit 48 is constituted by a read-only memory circuit, and outputs preset data using the difference data D3 as address data.

As a result, the constant setting circuit 48 outputs a constant H when the signal level of the differential data D falls within the range from the threshold t to -t, while If the value of the constant H exceeds the range of , the value of the constant H is suddenly lowered and output.

At this time, as shown in FIGS. 9 and 10, the constant setting circuit 48 uses the threshold control data Dr and the slope control data D as address data in addition to the difference data D, and outputs a constant H. As a result, the threshold value t is set according to the threshold control data DT and the slope control data DK.
, and -t1 and the maximum value, ,h.

The multiplication circuit 42 sequentially multiplies each image data of the digital video signal SOV by a value (1-H) and outputs the result.

On the other hand, the multiplication circuit 44 receives the motion-compensated digital video signals S and v output from the frame memory circuit 4.
Each image data is sequentially multiplied by a value H and output.

Addition circuit 46 adds the output signals of multiplication circuits 42 and 44.

As a result, the adder circuit 46 and the multiplier circuits 42 and 44 perform a predetermined multiplication value (1-H) that changes based on the difference data D.
), H, by performing arithmetic processing on the digital video signal S between frames, the digital video signal S
Designed to reduce 6V noise.

Further, in this embodiment, at the time of rising, the mode is switched to the threshold setting mode and the amplitude setting mode, and the threshold value 1 of the constant setting circuit 48 is switched. , -1m and the amplitude value are updated to optimum values to improve the noise reduction efficiency.

In other words, the threshold values t- and Lm of the constant H are set to optimal values in the threshold setting mode, similar to the threshold values of the differential data correction circuit 10 (FIG. 1), thereby effectively preventing image quality deterioration. It is possible to avoid this and reduce noise efficiently.

On the other hand, the amplitude value is set to the optimum value in the amplitude setting mode, similar to the slope of the differential data correction circuit 10 (Fig. 1), thereby effectively avoiding image quality deterioration within a practically sufficient range. It is possible to efficiently reduce noise.

According to the configuration shown in FIG. 7, even when the digital video signal S0 is processed between frames based on the difference data D to reduce noise, the threshold value ta is set based on the amount of noise.
, -ta to optimal values, the noise reduction effect can be improved.

(G3) Other Embodiments In the above-described embodiments, a case has been described in which the amount of noise is detected by detecting fluctuations in the signal level during the vertical blanking period, but the present invention is not limited to this, and the present invention is not limited to this. The quantity detection means can be widely applied.

Further, in the above-described embodiments, a case has been described in which the present invention is applied to a video tape recorder, but the present invention is not limited to this, but can be widely applied to reduce noise in various video equipment.

Effects of the Invention As described above, according to the present invention, by setting the feedback characteristic or the threshold value of the multiplication value to the optimum value based on the amount of noise, image deterioration is effectively avoided and the noise reduction effect is improved. You can get a noise reducer that can.

[Brief explanation of the drawing]

Fig. 1 is a block diagram showing a noise reducer according to an embodiment of the present invention, Figs. 2 to 6 are characteristic curve diagrams for explaining its operation, and Fig. 7 shows a noise reducer according to a second embodiment. The block diagram and FIGS. 8 to 10 are characteristic curves for explaining the operation. 1.40...Noise reducer, 2.8.22
・・・・・・Reduction circuit, 4：Frame memory circuit, 10：Differential data correction circuit, 12：
... Noise amount detection circuit, 24 ... Comparison circuit, 31 ... Threshold control circuit, 33 ...
・Slope control circuit, 42.44... Multiplication circuit, 4
8...Constant setting circuit. Noise difference Dependency on the positive circuit 1. Figure 2. Distribution of noise. Figure 3. Support for noise level. Change in noise level. Figure 6. Law of law data. Figure 6. Illustrated Poka' is the 1st year of the first year of life.

Claims (1)

[Claims] By obtaining differential data between frames of a sequential video signal and subtracting the differential data from the video signal with a predetermined feedback characteristic, or by generating a predetermined multiplication value based on the differential data. , a noise reducer that reduces noise in the video signal by performing arithmetic processing on the video signal between frames using the multiplication value, comprising: noise amount detection means for detecting the amount of noise in the video signal; A noise reducer comprising threshold control means for switching the threshold value of the feedback characteristic or the multiplication value based on the detection result of the means.