JP2002525927A - Video signal processing method and apparatus - Google Patents

Abstract

(57) Abstract: A video processing method and apparatus, comprising detecting a possible movement between consecutive fields of a video divided into even and odd numbers, and determining a mode, that is, a video mode or a film mode. In the method, a motion or stationary sequence between consecutive fields is detected and this sequence over several fields is stored in a memory, after which the sequence is compared with a pattern specific to the mode, and If a video mode is detected, median filtering is performed, while if a film mode (2: 2 pulldown or 3: 2 pulldown) is detected, the median filtering is switched off and synchronized with the film phase, Even and odd numbers belonging to one group, derived from the same film image Fields, methods again merged to the original film image is obtained, the the video, subsequent original film image is characterized in that it consists, repeating until again constituted by the merger.

Description

DETAILED DESCRIPTION OF THE INVENTION

The present invention relates to a video processing method for detecting a possible movement between continuous fields of a video divided into even and odd numbers and determining a mode, that is, a video mode or a film mode.

More particularly, the present invention relates to the processing of video signals from video tapes, DVDs, laser discs, tuners, etc., and thus also to the processing of standard PAL, SECAM and NTSC video signals. .

[0003] In particular, film modes such as so-called "2: 2 pulldown" or so-called "3: 2 pulldown" involve dividing the film image into even and odd fields, and these signals are standardized. When processed in this way, the quality of the video on the monitor can be annoying. In this standard manner, the detection of the motion between successive fields determines which mode will convert the current video signal and then perform further specific processing of the film mode signal.

In the known method, the detection of this movement is performed by detecting possible movements at the edges of the field, and based on the result, if there are many details in the field, a quick intervention is possible. Done. This requires a number of auxiliary devices, for example a complete correlator.

[0005] The use of these known methods therefore requires relatively expensive and large equipment.

[0006] It is an object of the present invention to provide a video processing device that can display the high quality video on a monitor even in the film mode by using a device that can be designed to be relatively inexpensive and compact. Is to provide a way.

It is an object of the present invention to detect a motion or still sequence between consecutive fields, store this sequence over several fields in a memory, and then store the sequence in a mode specific to the mode. If a normal video mode is detected in comparison with the pattern, median filtering is performed, while if a film mode (2: 2 pulldown or 3: 2 pulldown) is detected, the median filtering is switched off. Synchronizing with the film phase, merging matching even and odd fields derived from the same film image again until the original film image is obtained, which is merged with the subsequent original film image. Iterating until it can be configured again by Wherein the is realized by.

[0008] The detection of the motion, ie the detection of the sequence of motion and stillness, can be performed by a three-point median filtering operation, after which the result of the median filtering operation and the input information of the following fields are reduced to two low Filter through a low pass filter, calculate the absolute difference between the results of these two low pass filters, add these differences, possibly divide the sum by a number, and then compare this to a threshold. The result of the comparison constitutes the sequence, stored in memory, of motion and still between successive fields.

[0009] The video signal to be processed preferably has a doubled, and also preferably a 4
A doubling, and preferably also a doubling of the field speed, is added.

[0010] The invention also relates to an apparatus which is particularly suitable for using the method according to any of the preceding embodiments.

[0011] Accordingly, the present invention provides a motion detector, a film mode /
The present invention relates to a video mode detector, a synchronizing device for synchronizing processing with a film phase, and an apparatus having a film processor body.

Preferably, the motion detector is connected as input to a median filter having the current field and the next field of the video image and has two low-pass filters, one of these filters being a median filter. The other has as input the information of the next input field, and the motion detector comprises:
A differentiator connected to the two low-pass filters to calculate the difference between these outputs, a summer connected to the differentiator, a counter connected to the adder, and connected to the counter And a comparator for comparing the output of the counter with a threshold value.

Preferably, the film mode / video mode detector has a shift register in which the result of the comparator over several fields is stored,
Thereby, comparison with a pattern specific to a specific mode is enabled.

The device may have means for using doubling and / or means for performing quadrupling, and a 50 or 60 Hz to 100 or 120 Hz field rate converter.

In the following, in order to more fully describe the features of the present invention, a preferred embodiment of a method and apparatus for video processing according to the present invention will be described by way of non-limiting example with reference to the accompanying drawings. explain.

Starting with an interlaced video having 525 lines at 60 Hz or 625 lines at 50 Hz, interlacing is started first, so that the video is first converted to non-interlaced video.

For this purpose, the known or progressive scanning method is used, in which a three-point median is implemented by the median filter 1 for motion compensation.

For interlaced video, each frame consists of two fields,
Successive even and odd fields are offset from each other by half a line. The three-point median is, as its inputs, the pixel vertically above one of the two consecutive lines from the current field entering 1A, and the pixel vertically below the middle line of the next field entering 1A. With pixels.

The result of this median processing is merged with the original line of the current image to provide a non-interlaced 50/60 Hz 625 or 525 line motion compensated image.

The result of the median filter 1 is stored in a memory bank (display bank) 2. Bank 2 consists of two video FIFOs 3 and 4, which are written in interlaced form. Even pixels are in FIFO 3 and odd pixels are in FIFO 4.

“Double” refers to constant-speed reading of the display bank 2 at a normal speed.

When the interlacing is removed, a stable image without line flicker is obtained. In order to obtain the optimal screen coverage on a monitor system using a CRT, four more
A doubling method is used.

Simultaneously with the calculation of the median in the median filter 1, the intermediate pixels in the vertical direction are calculated by vertical linear interpolation in the interpolation circuit 5. The interpolation circuit 5
A connects to the output of the median filter 1. Again, 50
A / 60 Hz non-interlaced 625/525 line is obtained.

The result of this interpolation is stored in a memory bank (display bank) 6. Bank 6 consists of two video FIFOs 7 and 8, as before, which are written in an interleaved form. Even pixels go into FIFO 7 and odd pixels go into FIFO 8.

In order to obtain a quadrupled image, the memory controller 9 reads the lines of the display bank 2 and the lines of the display bank 6 alternately. The result is sent to the digital-analog converter 10.

Each line consists of 1024 pixels, 512 even pixels from FIFO 3 or 7, and 512 odd pixels from FIFO 4 or 8. These are read out in the correct pixel phase in an interleaved manner. Eventually, a complete image of 1250 or 1050 lines, respectively, which is not interlaced,
Obtained at a line frequency of 64 kHz and a frame rate of 50 or 60 Hz.

In order to eliminate the inherent screen flicker at low frame rates of 50 or 60 Hz, further up-conversion of the field rate, ie an increase in frequency to 100 or 120 Hz, must be performed.

In the case of the first embodiment, the lines of the display bank 2 are controlled by the memory controller 9 twice in succession, twice the normal speed, ie at a frequency of 100 or 120 Hz instead of 50 or 60 Hz. And 625 or 525 lines of non-interlaced video, respectively, are formed and displayed at twice the frame rate. In this case, the line frequency is 64 kHz.

To maintain the advantage of better screen coverage, the line frequency should be 128 kHz
, But this is not possible with most display devices. In order to solve this problem, in the second embodiment, the memory controller 9 first reads out the entire contents of the display bank 2 at twice the normal speed, and then reads out the entire contents of the display bank 6. The content is read at twice the normal speed. Since the two display banks 2 and 6 are alternately read at twice the speed, a field rate of 100 or 120 Hz is again obtained.

Since the contents of the display bank 2 are spatially different from the contents of the display bank 6, it is necessary to prevent data of the display banks 2 and 6 that coincide with each other from being written to the same position on the monitor. There is. 100 or 120
In order to achieve correct interlacing in Hz, the vertical frame outgoing pulse (100/120 Hz) is shifted up by half a 64 kHz line every other frame.

As a result of the interlacing, the information from one display bank is placed between the information from the other display banks 6 and therefore in the same time as in the normal quadrupling (20 ms or 16 ms). .6 ms), the optimum screen coverage is obtained.

Since this type of video is interlaced using non-equidistant frame pulses instead of equidistant frame pulses, it can only be used in devices that support this type of interlacing, such as HDTV compatible devices. .

It is necessary to calculate the frame pulse for the second 100 or 120 Hz field. The length of the frame is determined based on the number of 64 kHz lines in the original field. The second frame pulse is located between the two original frame pulses. Without interlacing, equally spaced 100 or 120 Hz frame pulses are obtained.

Thus, field rate upconversion can be applied to both doubling and quadrupling methods, with doubling and quadrupling separately with or without field rate upconversion. You can choose to use it. 2
In the case of doubling, the display bank 2 is read at a constant speed at a normal speed, and in the case of quadrupling, the lines of the display banks 2 and 6 are alternately read at a normal speed, and
In the case of doubling and doubling the field speed, the display bank 2 is read at a constant speed at twice the speed, and in the case of doubling and doubling the field speed, all the fields are alternately read from the display banks 2 and 6. Reading is performed at 2 × speed, and then interlacing for frame pulses is performed.

If not in video mode, ie, not a normal video recording where motion can occur between fields (20 or 16.6 ms intervals), the device can handle transmissions in other modes. For example, film mode, or a derivative mode thereof (eg, video editing mode transmission where normal video images are mixed with film mode video, and film editing modes where the input consists only of film mode video but the correct order is lost) Can be processed.

The film mode has a so-called “2: 2 pull-down” form or a so-called “3: 2 pull-down”.
2 "pull-down. In" 2: 2 pull-down ", 1
PAL 50Hz where 24 film images per second consist of 50 images per second
Or 30 film images per second are converted to NTSC 60 Hz, consisting of 60 images per second. In “3: 2 pulldown”, 1
Twenty-four film images per second are converted to NTSC 60 Hz and thus 60 images per second.

In the case of “2: 2 pulldown”, the film is slightly accelerated to 25 images per second for transmission in the PAL transmission standard, and then each film image or frame is even or odd. It is divided into images or fields, so that two fields A and B, or C and D, or E and F, etc. are obtained. These fields are aligned with each other and occupy the same time position, but are spatially located or separated at different positions, and when viewed vertically, as shown at the top of FIG. It is only shifted.

Each field A, B, C, etc., consists of a plurality of lines, and in the case of PAL,
It consists of 312.5 lines. In FIG. 2, these lines are A ₀ , A ₁ , A ₂ , etc. for field A, B ₀ , B ₁ , B ₂ for field B, and so on. The original film images are displayed sequentially on the monitor and can be reproduced as even and odd fields interlaced.

Similarly, a film having 30 images per second can be transmitted according to the NSTC standard. Each video or frame is again divided into even and odd fields.

In the case of “3: 2 pulldown”, as shown at the top of FIG. 4, a frame consisting of even and odd fields, eg, field A and field B, on the other hand consists of even and odd fields at the same instant. On the other hand, the even field at one instant and the odd field at the other instant, eg, odd field A,
Even field A ', odd field A, even field B, odd field B
', Etc., so in the end 60 images per second are produced again.

Starting from six original video frames, fifteen interlaced fields are created. In sequence, the three pictures or frames are divided into even and odd fields, followed by two split frames consisting of the even fields of one picture and the odd fields of the other pictures, the result being each of five sequential fields. Followed by a single field.

Film transmission, and thus 2: 2 and 3: 2 pull-down video, when properly detected and processed, gives much better quality results than regular video video.

In the above-described deinterlacing, a median filter 1 is used to obtain a correct motion compensation. However, the median filter has the disadvantage that it cannot handle diagonally moving lines. As a result, small steps or jagged so-called "jaggies" are formed on the diagonal.

This is a very troublesome effect. This effect is even more pronounced for 2: 2 and 3: 2 pulldown formats.

In order to significantly reduce this effect in the case of film transmission, de-interlacing must be performed without a median filter.

Therefore, in the first embodiment, the motion detector 11 is used for detecting which field has motion and which field has absolutely no motion.

This is done by observing the interrelationships between these fields.

To do so, pre-filtering is performed as a first step.

The matching continuous fields A (consisting of lines A ₀ , A ₁ , etc.) and B (consisting of lines B ₀ , B ₁ , etc.) occur at the same instant, but convert the film for video transmission. As a result of the vertical division.

Thus, if there is a spatial separation, it is wrong to compare pixel x from line A ₀ with pixel x from line B ₀ . Instead, the result of the median filter 1 described above, ie, the median of the pixels x of A ₀ , A ₁ , and B ₀ , is compared with the pixel x from line B ₀ . The median is because match the pixel x from spatially line B _0.

In this way, it is sufficient to obtain a first refinement in which a possible correlation between field A and field B can be found.

On the one hand, the result of the median filter 1, and on the other hand, the information coming from the next field via 12, respectively, is filtered by a low-pass filter 13 or 14, for example by a relatively inexpensive recursive filter. I do. At this time, some continuous pixels, for example, 32 continuous pixels are continuously averaged according to the following equation. (32 · x _old −1 · x _old + 1 · x _new ) (1/32)

The motion correlation detection is performed at a low frequency and does not rely on sharp transitions or details. This is to make it less dependent on the net bandwidth of the input signal and possible errors in the median filter 1.

Next, for each pixel, the differentiator 15 calculates the absolute difference between the result of the low-pass filter 13 for the result of the median filter and the result of the low-pass filter 14 for the next incoming image.

[0055] The absolute difference successive summed by the adder 16, when it reaches when there is overflow example 2 ⁹ i.e. 512, counter 17 connected to the adder is incremented. This value, or the sum of the absolute differences, is divided by, for example, 512 so that the number indicating the final result is not too large.

This process is carried out for the whole field, but preferably also for the so-called “
"Windowing" limits or locks from a particular start line to a particular stop line and from a particular start pixel to a particular stop pixel.

At the end of each field, the output of counter 17 gives a number indicating in principle the degree of movement between two consecutive video fields.

In a third step, relevant criteria are determined for determining whether there has been actual movement or stillness.

For bright video motion, the result of the sum of absolute differences between two consecutive fields is:
On average, it is larger than in the dark field case. Therefore, the calculation of the threshold as a motion criterion takes into account the luminance value of the current video being processed.

Therefore, the sum of the luminance is calculated simultaneously with the sum of the absolute differences. The threshold used for motion is determined empirically by a fraction of the sum of the luminance values, for example 1/6
4. This threshold can be doubled or halved, depending on the current processing operation.

If one image or field has small details, such as noise or detail structures, different from the next image, this can have a very bad effect on the result of the motion detection.

If the threshold value is too low, even in the case of a momentary stillness between successive images or fields, the result of motion may be given, which is incorrect. Therefore, the information is filtered by the low-pass filters 13 and 14 by a pre-filtering operation.

As will be described below, prior to performing motion detection, if one wants to make the most of edge boost (which may be necessary for quadrupling for configuration reasons), then The details of the detailed structure will play a large role in determining the threshold. Therefore, in determining this threshold, it is necessary to consider the state of the edge boosting device. The greater the edge boost, the greater the final threshold.

This additional adjustment of the threshold value as a correction for the edge boost is carried out simultaneously for normal values, for example 1/64, twice and half the luminance value.

In that case, the threshold is adjusted to the processing mode.

When the processing mode is a detection mode based on continuous monitoring of incoming data, motion detection is set to a nominal sensitivity, for example, 1/64 of the luminance value.

If a film mode or the like is detected and the processing is adapted to it, the motion detection is set to low sensitivity, eg, twice the nominal threshold.

A third mode, the film resynchronization mode (hereinafter, referred to as “the film resynchronization mode”) in which the system is still in the film mode but a momentary motion detection detects an incorrect film phase.
In the case of "film resync mode"), processing returns to standard median processing, and motion detection is set to a high sensitivity, eg, half the nominal threshold, to allow for rapid detection.

The threshold for motion is determined at the end of each field.

In this case, the film mode / video detection itself is performed as follows.

When the result of the counter 17 exceeds the threshold value, there has been motion between consecutive fields, and the comparator 18 of the motion detector 11 connected to the counter 17 outputs a logic “1”. Generate and this value is clocked into the shift register 19.

Conversely, if the sum is less than the threshold, there is a correlation or no motion, and the comparator 18 of the motion detector 11 detects a logic “0”, which is also the same. Next, the shift register 19 is entered.

The shift register 19 is composed of a bit string of a certain length, for example, 11 bit strings, which means that the history of the motion detector 11 can be judged over a certain number of fields, for example, 11 fields.

In film mode / video detection, the detector 20 scans the sequence of the shift register 19 for specific patterns specific to various modes.

The output of the detector 20 is connected through 20 A to a comparator 18 for feedback of the threshold actually determined by said detector 20.

In the case of film processed by 2: 2 pulldown, ie, a film containing 24 pictures per second, the PAL and the transmission standard 50 Hz have a rate of 5 per second.
In the event of motion when converted to 0 video or when a film with 30 video per second is converted to 60 video per second at NSTC 60 Hz, the motion sequence is as shown in FIG. "0101010101 ..."
It is.

Thus, if the detector 20 detects the “01010101...” Sequence, a determination is made that “2: 2 pulldown mode” is occurring.

For a film processed by 3: 2 pulldown, ie, a film containing 24 videos per second is converted to 60 videos per second at NSTC 60 Hz, the motion sequence is shown in FIG. Thus, "10100101001
01. . . , So if the detector 20 detects this sequence,
A determination is made that "3: 2 pulldown mode" is occurring.

When the detector 20 detects the motion sequence “1111111”, a determination is made that the normal video mode to be processed by standard median filtering is taking place.

The system sets the film mode display to high upon detection of a 2: 2 or 3: 2 “pull-down mode”. This means that the motion detector 11 has to be set to the low sensitivity mode and that processing synchronization has to be started. The detection length of the motion detection is also switched to 6 bits instead of a short word length, for example, 11 bits.

The processing synchronization is necessary so that data processing proceeds in a correct manner in response to detection of a motion sequence corresponding to 2: 2 or 3: 2 “pull-down mode”.

For this processing, the median filter 1 is switched off, the matching fields are merged, and the merged video is repeated, as will be described in detail below.

For this synchronization, the device is a synchronization device 2 consisting of two vibration shift registers.
One. One of these registers (hereinafter referred to as "3: 2 pulldown synthesizer" 22) generates a sequence "10100" and repeats it indefinitely. In other words, a 3: 2 pull-down processing sequence is continuously generated. Another register (hereinafter referred to as “2: 2 pulldown synthesizer” 23)
Generate sequence "01" and repeat it. In other words, a 2: 2 pull-down processing sequence is continuously generated.

The output of the synchronizer 21 is connected to the memory controller 9 through 21 A.

When one of the film modes is detected by the detector 20, the associated pull-down synthesizer 22 or 23 is synchronized by the film mode / video detection, eg, by a sync pulse from the detector 20. From that moment on, the processing operation takes place under the control of the working sinker 22 or 23 and proceeds completely synchronous with the incoming film phase.

The working sinker 22 or 23 also serves as a monitor reference for film mode / video detection. If the synthesizer 22 or 23 outputs a 0 and the detector 20 outputs a 1, then errors will penetrate into the film phase (failure of video source, too many vertical on / off structures, video editing situation, illegal "Queue flush", etc.) In such a situation, the device goes into film resync mode, the motion detector 11 is set to very high sensitivity, and the processing is temporarily returned to median filtering. Due to this error detection in the film phase, the output of synchronizer 21 forms part of a feedback loop to detector 20 through 21B.

[0087] These strategies provide ample opportunity to quickly recover a new film phase, even with small and brief movements. If there is too much movement,
There is a risk that the motion detector 11 will saturate and the device will no longer be able to distinguish between film mode and video mode.

In that case, the device switches to normal video mode, ie normal median filtering, and again starts searching for the film mode using the 11-bit word length indicating the history of the detector 11 and the nominal sensitivity. . From this point on, if there is enough movement, the correct film phase is found again in a relatively short time. Motion detector 1
The sensitivity switching of 1 can be prohibited by the interface. In that case, one of the three fixed sensitivity settings is selected.

The film processing itself is performed as follows.

The transition of the motion detector 1 from 0 to 1 means that motion has been detected between consecutive fields. A transition from 1 to 1 means that there is continuous movement between successive fields. A transition from 0 to 0 means stationary between consecutive fields. A transition from 1 to 0 means a transition from motion to stationary between consecutive fields.

If the correct film mode is found and the synthesizer 22 or 23 is in the correct state, the request when the output of the synthesizer 22 or 23 is 1 is to perform a merge between successive fields and The request when the output of 23 is 0 is that a repetition of the merged field should be performed. This is shown in FIG.
4 schematically shows the 2: 2 pull-down, and the lower part of FIG. 4 shows the 3: 2 pull-down.

In the case of 2: 2 pull-down, fields A and B, C and D, etc. are merged, with line B ₀ being between lines A ₀ and A ₁ and line B ₁ being lines A ₁ and A ₂ and so on. The merged fields are each repeated once, as indicated by the arrows at the bottom of FIG.

In the case of 3: 2 pull-down, fields A and A ′, B and B ′, and others are merged.
At this time, the line A '₀Is line A₀And A₁Between the line A '₁Is line A ₁ And A₂And so on. Figure 4 shows the merged fields.
Are repeated one or two times, respectively, as indicated by the arrows below.

In the normal video mode, the result of the median filter 1 is combined with the original lines of the current field, and these lines of the current field are then combined by the memory controller 9 with the next field of the current field. It is merged with the original line, thus producing a 625 or 525 line image. It simply merges the even and odd fields from the same image on the film.

Therefore, the line placed between the continuous lines is not the median but the line of the next field. In the case of 2: 2 pull-down, for example, as shown in FIG. 2, line B ₀ is arranged between lines A ₀ and A ₁ and in the case of 3: 2 pull-down,
For example, the line _A'0 as shown in FIG. 4 is disposed between the line _{A 0} and _A'1.

Therefore, in the case of the film mode, the median filter 1 is used only for measurement, not for processing.

The result of the merger is sent to the display bank 2.

The information is also passed through the median filter 1 to the interpolator 5, so that interpolation is also performed on the merged fields. The result of this interpolation is sent to the display bank 6.

The display banks 2 and 6 are read out as described above.

The repetition of the merged field, indicated by the arrow at the bottom of FIGS. 2 and 4, is performed with the writing to display banks 2 and 6 inhibited, and thus the display bank 2
And 6 are read, the same so-called "old" information is read.

The even and odd fields from the same film image are merged again,
An original film image is obtained and this image is repeated until the next original film image can be constructed by the merger.

As already mentioned, the edges and details of the video image can be selectively enhanced by so-called “edge boost”.

Such enhancement can be performed on the incoming signal by means of two phase linear “finite impulse responses” or FIR with zero coefficients. One of these filters is a bandpass filter having a coefficient “−10 20 −1”, and the other is a high-pass filter having a coefficient “−1 2 −1”. The sample frequency is 16 MHz.

The results of the two filters are merged and added or subtracted, multiplied or multiplied at a fixed rate, to the original video signal. Both filters have, for example, eight settings. This allows various combinations of the two filters.

If an excessively emphasized edge is detected, the effectiveness of the filter can be reduced, if necessary, to prevent deadlock.

The present invention is not limited to the embodiments described above and shown in the figures, and without departing from the scope of the present invention, the method and the method of video processing of the present invention can be modified in various ways. The device can be implemented.

[Brief description of the drawings]

FIG. 1 is a block diagram of an apparatus used to carry out a method according to the present invention.

FIG. 2 is a diagram schematically showing 2: 2 pull-down conversion and processing.

FIG. 3 is a diagram schematically showing a sequence of movement between fields in the case of a film processed by 2: 2 pull-down.

FIG. 4 is a diagram schematically showing 3: 2 pull-down conversion and processing.

FIG. 5 is a diagram schematically showing a motion sequence between fields in the case of a film processed by 3: 2 pull-down.

[Explanation of symbols]

Reference Signs List 1 median filter 2 memory bank (display bank) 3, 4 video FIFO 5 interpolation circuit 6 memory bank (display bank) 7, 8 video FIFO 9 memory controller 10 digital-analog converter 11 motion detector 13, 14 low-pass filter 15 Differentiator 16 Adder 17 Counter 18 Comparator 19 Shift register 20 Detector 21 Synchronizer 22 3: 2 pulldown synthesizer 23 2: 2 pulldown synthesizer A, B, C, D, E, F, A ', B' , C 'fields A, B, C, D, E, F, A', B ', C' with 0-2 subscript lines

Claims (21)

[Claims] 1. A method of video processing, comprising detecting possible movements between successive fields of an even and odd numbered video and determining a mode, ie a video mode or a film mode. Detecting a motion or stationary sequence between consecutive fields, storing this sequence over several fields in memory, and then comparing the sequence with a pattern specific to the mode, a normal video mode If it is detected, median filtering is performed, while if a film mode (2: 2 pulldown or 3: 2 pulldown) is detected, the median filtering is switched off and the same film is synchronized. The matching even and odd fields derived from the video are How again merged until the film image is obtained, the the video, subsequent original film image is characterized in that it consists, repeating until again constituted by the merger. 2. The method according to claim 1, wherein when a film mode is detected, the length of the motion sequence stored in the memory is reduced. 3. The detection of the motion, ie the detection of the sequence of motion and stillness, can be performed by a three-point median filtering operation, after which the result of the median filtering operation and the input information of the following fields are compared. Two low-pass filters, calculate the absolute difference between the results of the two low-pass filters, add the differences, divide the sum, possibly by a number, and compare it to a threshold Method according to claim 1 or 2, wherein the result of this comparison constitutes the sequence of motion and stillness between consecutive fields, stored in memory. 4. A different binary number is stored in a memory depending on whether the sum is greater than or less than the threshold, wherein a sequence of binary numbers in the memory is used to determine the motion and stillness between successive fields. 4. The method of claim 3, comprising constructing a sequence. 5. The method according to claim 3, wherein the threshold value is calculated taking into account the luminance value of the image currently being processed. 6. The sum of the luminances is calculated and the nominal threshold selected is the value of a certain percentage of the sum of the luminances, said nominal value being taken into account for the processing mode and the edge boosting being performed. Method according to claim 5, characterized in that, if so, it is adjusted, in particular twice or half, to take account of the state of the edge boosting. 7. When a film mode is detected, data processing is performed in a 2: 2 mode.
A sequence corresponding to the pull-down mode and / or a sequence corresponding to the 3: 2 pull-down mode is continuously transmitted, and when one of the corresponding modes is detected, transmitted by detecting a motion and still sequence between consecutive fields. By using a vibration shift register to synchronize with the synchronization pulse,
7. The method according to claim 1, wherein the method is synchronized with a film phase. 8. In the case of detection of an asynchronous or error in a film phase, a synchronous mode is implemented for switching to film, the sensitivity of motion detection is set to a high level again, and processing is resumed when the film phase is observed again. 8. The method according to claim 7, wherein the method is temporarily re-performed by median filtering until it is performed. 9. The method according to claim 1, wherein when a film mode is detected and synchronization is being performed, the threshold value is increased so as to reduce the sensitivity of detecting motion or stillness. The method according to any one of 5 to 8. 10. The method according to claim 1, wherein the input signal is a two-phase linear filter with a sum of 0 coefficients, ie a bandpass filter having preferably a coefficient of 10 2 −1 and a
Filtering by means of a high-pass filter having 12-1 to add an edge boost to the signal, after which the results of these filtering operations are merged and increased or decreased at a constant rate. Item 10. The method according to any one of Items 1 to 9. 11. A video signal according to claim 1, wherein a doubling or quadrupling and / or a doubling of the field rate are applied to the video signal to be processed. Method. 12. The interlaced video signal to be processed is converted to a non-interlaced video signal by a three-point median filtering operation, and the result is stored in a memory bank, which is then operated at a higher speed than normal, preferably The method according to claim 11, characterized in that the field rate is up-converted by reading at a constant speed at 2x speed, for example at 100 or 120Hz instead of at a frequency of 50 or 60Hz. 13. An interlaced video signal to be processed is converted to a non-interlaced video signal by a three-point median filtering operation, the result of which is stored in a first memory bank, and at the same time interpolation is applied to these video signals. Done,
The result is stored in the second memory bank, and thereafter, one line is read from one memory bank or one line from another memory bank, and the contents of the first and second memory banks are sequentially read. At a higher than normal speed, preferably at double speed, eg 100 or 120 Hz instead of 50 or 60 Hz
12. The method of claim 11 wherein reading at a constant speed results in a field rate upconversion and shifts outgoing frame pulses by half a line every other frame for correct interlacing. The method described in. 14. Apparatus for use in a method according to any one of the preceding claims, comprising a motion detector (11), a film mode / video mode detector connected to the detector. , A synchronization device for synchronizing the process with the film phase (
21) and an apparatus having a film processor main body. 15. A motion detector (11) connected as input to a median filter (1) having a current field and a next field of the video image and having two low-pass filters (13 and 14). One of these filters is connected to the output of the median filter (1), the other has information of the next input field as input (12), and the motion detector (11) has two low-pass filters. (13 and 14) and a differentiator (1) which calculates the difference between these outputs.
5), an adder (16) connected to the differentiator (15), a counter (17) connected to the adder (16), and a counter (17) connected to the counter.
15. Apparatus according to claim 14, comprising a comparator (18) for comparing the output of (1) with a threshold value. 16. A film mode / video mode detector comprising a shift register (19) in which a comparator (18) over several fields is provided.
16. The apparatus according to claim 15, wherein the result of step (d) is stored, thereby enabling comparison with a pattern specific to a specific mode. 17. The synchronizer (21) has at least one vibration register or so-called “sincer” (22 or 23),
A sequence corresponding to the sequence of motion and still between successive fields for 2 pulldown or 3: 2 pulldown is continuously transmitted and the synchronizer (2
1) preferably has two shift registers or "sincers" (22 and 23), one for 2: 2 pulldown and one for 3: 2 pulldown Apparatus according to any one of claims 14 to 16, characterized in that: 18. The film processor according to claim 14, wherein said film processor body has means for merging aligned fields of a film image and for repeating the merged image. Equipment. 19. Apparatus for video processing, means for using doubling and / or means for performing quadrupling, and for example 50 or 60 Hz
Device having a field rate converter from 100 to 120 or 120 Hz. 20. Apparatus according to any one of claims 14 to 18 and claim 19, comprising a motion detector (11), a film mode / video mode detector,
Apparatus, characterized in that it comprises means for performing doubling and / or means for performing doubling together with a synchronizer (21) and a film processor, and a field speed converter. 21. Apparatus according to claim 15, wherein the means for performing the doubling and / or the means for performing the doubling comprise a median filter (1), the filter comprising a motion detector (1). An apparatus, which also forms part of 11).