KR20000061568A - Method for frame rate conversing of video signal - Google Patents

Abstract

PURPOSE: A method for changing a frame rate of a video signal is provided to reduce variance in displaying time and improve image quality and performance of a transmitter/receiver. CONSTITUTION: The method for changing a frame rate of a video signal comprises steps of repeating a standard field having a small time variance to interpolate a new field, interpolating a new field with an average value of standard fields most adjacent to each other with a display time axis in the center, dividing the standard field into block units to obtain a motion vector and then interpolating a new field using the motion vector, and selecting one out of the interpolating steps to interpolating a new field. In the step of interpolating the new field using the average value of standard fields, if a top field will be displayed based on the display time, the interpolating step is performed using two top fields most adjacent to each other. if a bottom field will be displayed based on the display time, the interpolating step is performed using two bottom fields most adjacent to each other.

Description

Method for frame rate conversing of video signal

The present invention relates to a method for converting a frame rate of a video signal, and more particularly, to a frame rate conversion method for converting a frame rate of 24 frames / second to a frame rate of 30 frames / second.

In general, when displaying the video signal, if the frame rate of the video being transmitted is different from the frame rate to be displayed, it may be natural to achieve the display only by adjusting it. For example, if the frame rate to be transmitted is 24 frames / second and the frame rate to be displayed is 30 frames / second, the frame rate should be changed to 30 frames with only 24 frames of information. Therefore, a pull down method is used for this.

In the 3: 2 pull-down method, as shown in FIG. 1, four fields of information corresponding to two frames of 24 frames / sec are displayed to be displayed in five fields.

At this time, the reason called 3: 2 pull down is as follows. That is, when creating 5 fields with 4 fields of information, if one field of the top or bottom field is repeated for the top and bottom fields corresponding to the 2 fields, the information of the 2 fields becomes 3 fields. If the remaining 2 fields are displayed as they are, finally 5 fields of information are created. When you do this, the first two fields become three fields and the remaining two fields become two fields, so the 2: 2 field turns into a 3: 2 field, so it's called 3: 2 pulldown.

The conventional 3: 2 pull down conversion method is described in detail as follows. In FIG. 1A, four frames (8 fields) are prepared. Each field name is N-1 , N-1 , N , N , N + 1 , N + 1 , N + 2 , N + 2 In this case, when each field undergoes a 3: 2 pull down process, mapping is performed as shown in FIG. The format in which the mapping is made is the same as indicated by the arrows in FIG. 1.

And, when the 3: 2 pulldown is performed, inevitably there is a deviation of the display. Figure 2 shows the number of time deviations before and after this 3: 2 pull down.

Here, the sum of the time deviations is the sum of the absolute values of the temporal differences between the displayed time points between the ten fields after the pull down is performed for the eight fields before the pull down is performed, and the equation Equation 1

Sum of Display Time Deviation = | 1-0.8 | + | 2-1.6 | + | 3-2.4 | + | 4-3.2 | + | 4-3 | + | 6-4.8 | + | 5.6-5 | + | 8-6.4 | + | 7.2-7 | + | 8-8 |

= 0.2 + 0.4 + 0.6 + 0.8 + 1 + 1.2 + 0.6 + 1.6 + 0.2 + 0

= 6.6

That is, as shown in Equation 1, the sum of the display time deviations is 6.6, which means that a total time difference corresponding to 6.6 fields is displayed while displaying 10 fields. This can be regarded as a numerical value where the display screen looks unnatural. For this reason, seeing a video that actually has a 3: 2 pulldown is annoying. That is, flicker phenomenon and unnatural display phenomenon appear.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and an object of the present invention is to provide a frame rate conversion method of a video signal which reduces display time variation by interpolating a new field with a repeating field order different from the conventional one.

Another object of the present invention is to provide a method of converting a frame rate of a video signal by interpolating a new field with an average value of nearest reference fields around the display time axis.

It is still another object of the present invention to provide a method of converting a frame rate of a video signal by dividing two corresponding fields into macroblock units to obtain a motion vector and interpolating a new field based on the value of the motion vector.

It is still another object of the present invention to provide a frame rate conversion method of a video signal interpolating a new field with the most optimal frame rate conversion method among the frame rate conversion methods described above according to the degree of motion of an image.

1 is a diagram illustrating a conventional 3: 2 pull down conversion process.

FIG. 2 is a diagram illustrating an example of temporal deviation in the 3: 2 pull down conversion process of FIG. 1.

3 is a diagram illustrating a 3: 2 pull-down conversion process according to an embodiment of the present invention.

4 is a diagram illustrating an example of temporal deviation in the 3: 2 pull-down conversion process of FIG. 3.

5 is a diagram illustrating a 3: 2 pull-down conversion process according to another embodiment of the present invention.

6 illustrates an interpolation process using reference fields in the pull-down method of FIG. 5.

7 is a diagram illustrating a 3: 2 pull-down conversion process according to another embodiment of the present invention.

8 is a flowchart illustrating a process of selecting any one of the pull-down methods proposed in FIGS. 3 to 7;

The frame rate conversion method of a video signal according to the present invention for achieving the above object comprises the steps of interpolating a new field by repeating a reference field having a smaller time variation, and the nearest reference field around the display time axis. Interpolating a new field with an average value of the fields, obtaining a motion vector by dividing a reference field by a block unit, and then interpolating a new field using the motion vector, and selecting one of the interpolation steps to select a new field. And interpolating.

The field interpolation selection step may select any one of the interpolation steps according to the degree of motion of the image.

Other objects, features and advantages of the present invention will become apparent from the following detailed description of embodiments taken in conjunction with the accompanying drawings.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

The present invention proposes an improved various 3: 2 pull down method and selects the most optimal pull down method according to the degree of motion of a video to convert a frame rate of 24 frames / sec into a frame rate of 30 frames / sec.

First, the various pull-down methods that have been improved are described, and finally, the process of selecting the most optimal pull-down method is described.

3 illustrates a 3: 2 pull-down process according to an embodiment of the present invention, and induces a natural display by reducing a display time variation by changing a repeating field from a conventional method.

That is, the eight fields shown in (a) of FIG. 3 represent the original video at the frame rate of 24 frames / second. The numbers in each field (N-1, N, N + 1, N + 2, ...) and , The symbol represents eight fields that proceed sequentially. The 10 fields shown in (b) of FIG. 3 indicate each field converted to a frame rate of 30 frames / second after the 3: 2 pull-down process is completed.

In the present invention, first, the mapping method when 3: 2 pull down is performed is different. The mapping process from the 8 fields to the 10 fields is the same as shown in FIG. 3 and listed below.

N-1 top <-N-1 , N-1 Bottom 〈-N-1 , N top 〈-N , N bottom 〈-N ,

N + 1 top 〈-N , N + 1 Bottom 〈-N + 1 , N + 2 top <-N + 1 ,

N + 2 Bottom 〈-N + 1 , N + 3 saw <-N + 2 , N + 3 Bottom 〈-N + 2

That is, conventionally, when interpolating N + 2 bottom, N + 2 Was repeated, the present invention is N + 1 By repeating, the time deviation is reduced.

The process of calculating the deviation of the display involved in this process is shown in FIG. 4. In FIG. 4, the sum of the time deviations is the sum of the absolute values of the corresponding temporal differences with respect to the time when the 10 fields after the pull down is displayed for the eight fields before the pull down is performed. , Represented by the equation is equal to Equation 2 below.

Sum of Display Time Deviation = | 1-0.8 | + | 2-1.6 | + | 3-2.4 | + | 4-3.2 | + | 4-3 | + | 6-4.8 | + | 5.6-5 | + | 6.4-6 | + | 7.2-7 | + | 8-8 |

= 0.2 + 0.4 + 0.6 + 0.8 + 1 + 1.2 + 0.6 + 0.4 + 0.2 + 0

= 5.4

As shown in Equation 2, the sum of the display time deviations is 5.4, which means that a total time difference corresponding to 5.4 fields occurs while displaying 10 fields.

Compared with the conventional method, this value has been reduced from 6.6 to 5.4, and thus the degree of disturbance has been improved by at least one field.

FIG. 5 illustrates a 3: 2 pull down process according to another embodiment of the present invention. When a field to be repeated is designated, two top fields or bottoms closest to each other based on the display time axis are not directly repeated. Interpolate new fields using fields and use them instead of repeating fields.

Referring to (a) and (b) of FIG. 5, the upper side shows two reference fields which are the targets when interpolating a new field, and when the display point to be repeated is reached, the two closest fields on the time axis are referred to. After selecting, use the information in the two selected fields. Here, assuming that the field to be repeated is N + 1 top, the two reference fields are each N as shown in FIG. And N + 1 Becomes In addition, when the field to be repeated is N + 1 bottom, the two reference fields are N as shown in (b) of FIG. 5. And N + 1 Becomes

In this case, since the field to be interpolated is located at the center of the other two reference fields, an average may be obtained for each pixel. The value of the pixel corresponding to the 6-5 position of (d) of Figure 6 is the value of the 6-1 position of (a) and the value of the 6-4 position of (c) as shown in Equation 3 below It is an average.

In addition, when focusing on the displayed time point, the field to be repeated can obtain information from the nearest field regardless of whether it is a top or bottom field. In FIG. 6D, the value of the 6-5 position may be obtained by using the bottom values at the positions 6-2 and 6-3 of FIG. 6B.

In this case, the final calculated value at position 6-5 can be obtained as shown in Equation 4 below.

In this case, the α value is a value that adjusts to reproduce the most natural image experimentally.

However, when the motion or the image is severely changed, the above method may not be applied well. In this case, the motion vector between two reference fields is estimated for two positions to be restored, and the final value is calculated based on the motion vector.

FIG. 7 illustrates a 3: 2 pull down process using this method. Instead of interpolation, two corresponding fields are divided into macroblock units to obtain a motion vector as in a moving picture expert group (MPEG) video compression algorithm. To interpolate between both fields based on the value.

That is, when there are two reference fields 1 and 2 for constructing a new interpolation field, the value of the position to be obtained in the reference field moves to different reference fields on the corresponding positions of the reference fields 1 and 2. Calculate the vector The method of calculating a motion vector defines a macro block around a predetermined position 7-1 and finds a motion vector up to a position 7-4 in which the macro block has a minimum error to another reference field. This time, we estimate the motion vector in the opposite direction, find the motion vector from the position (7-2) to the position (7-3) with the minimum error in the same way, and then the value of the position indicated by these two motion vectors. Interpolate new fields using averaging

Considering all the above methods, the interpolation field can be obtained by selecting one of the proposed methods or by selecting one using arbitrary selection conditions.

8 is a flowchart illustrating a process of selecting an optimal pull down method among the pull down methods proposed in the present invention. First, a correlation is calculated for two reference fields (step 801). That is, when the difference between two reference fields is calculated and this calculated value is m1, the value of m1 is small when the two reference fields are similar to each other, and the value of m1 is large when they are different from each other.

Accordingly, the correlation value m1 is compared with the first threshold th1 (step 802). If the correlation value m1 is smaller than the first threshold value th1, it is determined that there is almost no change in the motion or the image, and the field is repeated as shown in FIG. Interpolate the new field with step 803.

If it is determined in step 802 that the correlation value m1 is greater than the first threshold value th1, the correlation value m1 is again compared with the second threshold value th2 to determine the degree of movement (step 804).

If it is determined in step 804 that the correlation value m1 is greater than the second threshold value th2, it is determined that the motion or the change of the image is very severe, and the new field is interpolated using the motion estimation as shown in FIG. 7 (step 809).

On the other hand, if it is determined in step 804 that the correlation value m1 is smaller than the second threshold value th2, the movement is not very large but there is some degree. In this case, as shown in FIG.

At this time, since the value of α needs to be defined, another correlation value m2 is obtained (step 805).

That is, the correlation value m2 is the two reference fields and the other one (the field closest to the point of time to be interpolated, and if both reference fields are tops, they are bottom fields therebetween. Field is equivalent).

The calculated correlation value m2 is compared with the third threshold th3 (step 806).

If it is determined in step 806 that the correlation value m2 is smaller than the third threshold value th3, the value α is greater than 0 and less than 1 (0 <α <1). Interpolate the fields (step 807).

If it is determined in step 806 that the correlation value m2 is greater than the third threshold value th3, the α value is 1, and in this case, the new field is interpolated using the average of the reference fields as shown in Equation 3 above (step 808). Here, the first to third thresholds th1 to th3 are determined through experiments and may vary depending on the designer.

This 3: 2 pull down method is widely used in the transmitter because it is useful for converting a film source made at 24 frames / second to a typical TV interlace signal of 30 frames / second. Therefore, applying such a method at the transmitting end can produce a natural, high quality image just before transmitting a video signal or performing video compression.

In parallel with this, the receiver can be effectively used. The reason for this is that when a video is received using a standard such as MPEG, a video that is initially 24 frames / sec is directly compressed without sending a change in the frame rate. to be.

In this situation, the receiver can obtain satisfactory results for both the compression ratio and the naturalness of the image by using the 3: 2 pull-down method proposed in the present invention.

That is, the 3: 2 pulldown method according to the present invention can be effectively used for both the transmitter and the receiver.

As described above, according to the method of converting a frame rate of a video signal according to the present invention, a new field is interpolated by performing field repetition in a different order from that in the conventional way, that is, to reduce time variation, or adjacent reference fields based on display time. By interpolating a new field with an average value of these fields or interpolating a new field using motion estimation, there is an effect of reducing flicker and unnatural display phenomenon.

In addition, natural image quality may be obtained by interpolating a new field by selecting any one of the pull-down methods described above according to the degree of motion or the change of the image.

In particular, the present invention can be used when converting the film source before transmitting at the transmitter side, and may be applied to restore the film mode and finally just before the final display when transmitting the video by compressing by the MPEG method. , Or improve the performance of the receiver.

Claims (5)

Interpolating a new field by repeating the reference field having the smaller time deviation; Interpolating a new field with the average of the nearest reference fields around the display time base; Dividing a reference field by a block to obtain a motion vector, and then interpolating a new field using the motion vector; And selecting any one of the interpolation steps to interpolate a new field. The method of claim 1, wherein the interpolation step using the average value of the reference fields If the top field is displayed based on the display time, interpolation is performed using the two closest top fields. The method of claim 1, wherein the interpolation step using the average value of the reference fields If the bottom field is displayed based on the display time, interpolation is performed using the two nearest bottom fields. The method of claim 1, wherein the interpolation step using the average value of the reference fields A method of converting a frame rate of a video signal, characterized in that interpolation is performed using three top and bottom fields closest to each other based on a display time. The method of claim 1, wherein the field interpolation selection step is And selecting one of the interpolation steps according to the degree of motion of the image.