KR0157467B1 - Moving image encoding method and device - Google Patents

Abstract

The present invention relates to a video encoding method and apparatus according to a buffer state.

According to the present invention, a function of varying an area of a DCT coefficient to be encoded regardless of a frame type according to the fullness of a buffer, a function of encoding only a DC component in an I frame, and no DCT coefficient in a B or P frame A function of transmitting only the same vector, and a function of limiting the number of DCT coefficients to be coded regardless of the frame type in the scan order of the scan-pattern.

Thus, not only the buffer overflow is prevented, but also the effect of restoring data without damaging the image quality is provided.

Description

Video encoding method and device according to buffer status

1 is a block diagram showing a video encoding apparatus according to a buffer state according to the present invention.

FIG. 2 is a conceptual diagram illustrating region classification according to buffer fullness in the apparatus of FIG.

3 (a) to 3 (c) are conceptual views illustrating a video encoding method for preventing overflow of a buffer in a FIG. 1 device.

* Explanation of symbols for main parts of the drawings

12 low pass filter 13 quantization unit

19: variable length encoding unit 20: buffer

The present invention relates to a video encoding method and apparatus according to a buffer state. More particularly, when a buffer overflow occurs, an overflow occurs in a buffer by limiting the number of coefficients to be encoded to reduce the amount of data generated. The present invention relates to a video encoding method and an apparatus according to a buffer state.

Recently, in systems such as HDTV, HD VTR, digital VTR, digital camcorder, multimedia, videophone, video conference, etc., video signals and audio signals are encoded into digital signals, transmitted or recorded on media, and decoded again. Playback is mainly used. In such a digital video processing system, the IBBPBBPI... Code as recommended by the moving picture expert group (MPEF). If a GOP (Group of Picture) structure is present, the decoding side can perform random access processing by this GOP structure. More specifically, in the MPEGII video compression encoder, frames are classified into Intra (I), Bidirection (B), and Predictive (P) types, and the order thereof is IBBPBB. I'm in order. The frames from the first I frame to just before the next I frame are called GOPs and are usually composed of 12 to 15 frames.

In order to perform encoding and decoding, each frame is divided into a macroblock of 16 × 16 pixels, and the macroblock is divided into four blocks of 8 × 8 pixels. In the encoding apparatus, a motion vector and a quantization level are determined in macroblock units.

In general, the encoding system encodes and transmits image data to be currently encoded, or encodes a difference signal between frame data using a very high correlation between screens of the image data. That is, the motion vector is detected by the amount of motion from the previous screen, and the discrete cosine transform (DCT) of the difference signal with the image data obtained using the detected motion vector is performed. These transform coefficients are quantized by the quantization level determined by the buffer state and the image characteristics of the input image data. The quantized coefficients are scanned according to the scan order of a given scan-pattern and encoded in the form of [run, level]. When scanning an N × N block, start with low frequency components and scan with high frequency components to encode run and level pairs. Here, runs are the number of zeros between nonzero coefficients in N × N block quantized coefficients, and the level corresponds to the absolute value of the nonzero coefficient. As described above, the data encoded in the form of [run, level] is subjected to a variable length encoding process to obtain final encoded image data.

To some extent in the rate control section that determines the quantization level, the overflow of the buffer can be prevented by varying the quantization level according to the state of the buffer. The amount of generation increases rapidly, causing an overflow in the buffer. Therefore, conventionally, when there is a risk of overflow, the frame is skipped without being encoded so that there is no bit generation amount. The decoding side displays the skipped frame by replacing the previous frame.

This can prevent the overflow of the buffer by greatly reducing the amount of bits generated. However, since the information on the frame is not encoded and transmitted at all, not only does it cause damage to the image quality during playback, but also a significant problem when there is important information in the skipped frame. do.

Accordingly, an object of the present invention is to divide the buffer into three state areas so as to solve the above-mentioned problem, and to prevent overflow as well as to prevent overflow by limiting coefficients to be forcibly encoded when the buffer is in the overflow risk area. The present invention provides a method of encoding corresponding information in a frame without a frame.

Another object of the present invention is to provide an apparatus for implementing a video encoding method according to the buffer state described above.

The video encoding method according to the buffer state of the present invention for achieving the above object, the step of dividing the buffer into an underflow risk region, a normal operation region, an overflow risk region according to the fullness of the buffer, the buffer overflow risk Filtering the conversion coefficients of the predetermined block by limiting the filtering band when in the region, quantizing the conversion coefficient of the predetermined block or the filtered conversion coefficient, and limiting the number of scans when the buffer is in the overflow risk region. And scanning the quantized coefficients by a limited number in a preset scan-pattern and encoding the same.

In accordance with another aspect of the present invention, there is provided a video encoding apparatus according to a buffer state. A low pass filter for filtering, a quantization unit for quantizing a transform coefficient or a filtered transform coefficient of a predetermined block, and a limited number of scans according to a fullness of the buffer to limit the number of quantized coefficients according to a preset scan-pattern It includes a variable length encoding unit for encoding the data generated by scanning.

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

1 is a block diagram showing an embodiment of a video encoding apparatus according to a buffer state of the present invention. In the apparatus shown in FIG. 1, the input digital image data is divided into small blocks (e.g., 8x8 or 16x16) of a predetermined size. Then, the motion estimation unit 17 detects a motion vector MV for predicting a motion from the portion of the image data of the previous screen stored in the frame memory 16 corresponding to the currently input block data with respect to the input block data. Motion Estimation is performed. The motion compensator 18 performs motion compensation by extracting the block of the previous screen closest to the current block by using the motion vector MV obtained from the motion estimation unit 17. The encoding process then processes only the currently input block (intra mode) or processes the difference signal of the compensating block with the current input block (inter mode). This encoding process is performed by the N × N converter 11, the quantizer 13, and the variable length encoder 19, and the inverse quantizer 14 to store the image data of the previous frame in the frame memory 16. ) And an N × N inverse transform unit 15.

In addition, the buffer 20 is controlled to maintain a constant bit rate of a signal that is encoded and transmitted. In order to prevent overflow or underflow of the buffer, a quantization level for controlling the amount of data generated by the quantization unit 13 is generated in units of slices according to the data accumulation state of the buffer. The quantization level in slice units is input to the ratio controller 21. The ratio controller 21 receives the input image data in the unit of block and judges the image characteristic, and varies the quantization level in the slice unit according to the image characteristic to generate the quantization level in the macroblock unit to the quantization unit 13. Supply. The above description is an operation description of a general video encoding apparatus, and an operation description for preventing an overflow of the buffer 20 is as follows.

First, the buffer 20 is divided into three areas as shown in FIG. 2 according to the data accumulation state. First, underflow risk area; Second, normal operating area; Third is the overflow risk area. The region classification is determined to an appropriate size according to the size of the buffer 20 and the total number of image pixels to be encoded. In particular, when the buffer 20 is in the overflow risk area, the encoding process will be described according to the frame type.

In the case of an I frame that encodes the input image itself among the frame types, a transform coefficient (shown in FIG. The conversion coefficient to be encoded is limited by passing the filtering band determined according to the state of 20). That is, it can be expressed as follows.

Y (U, V) = X (V, U) · H (U, V)

Here, X (U, V) is a conversion coefficient in the block, and the energy of the conversion coefficient is mainly concentrated toward the low frequency side. H (U, V) is a filtering band as follows.

H (U, V) = 1 V N _U , V N _V

= 0 other range

N _U and N _V have a value between 0 and 7, depending on the buffer status. Y (U, V) is the output value passing through the filtering band.

If both N _U and N _V have a value of 0, only the DC component passes and is encoded and transmitted. In this case, the buffer state exceeds 50% of the overflow risk area. In addition, the number of quantized coefficients to scan in a given scan-pattern is limited according to the buffer state. For example, suppose that in the scan-patterns shown in FIGS. 3 (b) and 3 (c), the number of bits to be encoded is limited to five by the buffer state. If the quantized coefficients in the block are referred to as coordinates Z (U, V), in scan-pattern 1 such as the third (b) diagram, [Z (0,0), Z (1,0), Z (0,1) ), Z (0,2), Z (1,1)], and five quantized coefficients at the coordinate positions are scanned and encoded. In scan-pattern 2 as shown in FIG. 3 (c), [Z (0,0), Z (0,1), Z (0, X), Z (0,3), Z (1,0)], etc. The five quantized coefficients at the coordinate positions are scanned and encoded.

On the other hand, in the case of B and P frames among the frame types, intra macro blocks in B and P frames are encoded in the same manner as I frames. The same is true for the inter macroblock, but the filtering coefficient is limited when the filtering band is limited through the low pass filter 12. Since the DC component does not exist in the inter mode, the data to be filtered when the filtering band N _U or N _V is 0. There will be no. If there is no data to be filtered, the quantizer 13 and the variable length encoder 19 do not operate. However, in the inter mode, even if only the information of the moving vector (MV) can be restored, the macroblock data can be restored on the decoder side, so that only the moving vector (MV) obtained by the tracking unit 17 is transmitted when there is no data to be filtered.

As described above, the video encoding method and apparatus according to the buffer state of the present invention divides the buffer into three areas according to the data accumulation state, compared to the conventional method of skipping without encoding a frame when there is a risk of overflow in the buffer. If the buffer is in the overflow risk area among the three areas, it is possible not only to prevent the overflow of the buffer but also to improve the image quality by transmitting the DC value or the dynamic vector or limiting the number of quantized coefficients to be scanned according to the frame type. Has the effect of restoring without damage.

Claims (8)

An apparatus for encoding a continuous array of I-frames, P-frames, and B-frames, comprising: a buffer for maintaining a constant bit rate of a transmission signal; A low pass filter configured to filter a conversion coefficient of a predetermined block by limiting a filtering band according to the fullness of the buffer; A quantization unit for quantizing a transform coefficient or a filtered transform coefficient of a predetermined block; And a variable length encoding unit configured to limit the number of scans according to the fullness of the buffer and to encode data generated by scanning the quantized coefficients by a limited number according to a preset scan-pattern. The video encoding apparatus of claim 1, wherein the buffer is divided into an underflow risk region, a normal operation region, and an overflow risk region according to the size of the buffer and the total number of pixels of an image to be encoded. The video encoding apparatus of claim 2, wherein the low pass filter filters only the DC component of the conversion coefficient when the frame type to be encoded is an I frame and the buffer is in the overflow risk region. The video encoding apparatus of claim 2, wherein the low pass filter comprises P frames and B frames to be encoded, and does not filter a conversion coefficient when the buffer is in an overflow risk region. The video encoding apparatus of claim 4, wherein a motion vector is transmitted instead of the transform coefficient filtering. A method of encoding according to a state of a buffer for maintaining a bit rate of a transmission signal when encoding a continuous array of I-frames, P-frames, and B-frames, wherein the buffer is underflowed according to the fullness of the buffer. Dividing into a risk area, a normal operation area, and an overflow risk area; Filtering the conversion coefficient of a predetermined block by limiting a filtering band when the buffer is in an overflow risk area; Quantizing a transform coefficient of the block or the filtered transform coefficient; And limiting the number of scans when the buffer is in the overflow risk area, and scanning and encoding the quantized coefficients by a limited number in a preset scan-pattern. 7. The video encoding method according to claim 6, wherein the filtering step includes filtering the DC component of the transform coefficient during video encoding of an I frame. The video encoding method according to claim 6, wherein in the filtering step, a motion vector is transmitted instead of filtering a transform coefficient when encoding P frames and B frames.