JP3090763B2 - Video encoding device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a video encoding apparatus having a motion compensation predictor and an in-loop filter, and more particularly to a technique for controlling on / off of an in-loop filter.

[0002]

2. Description of the Related Art In highly efficient coding of a moving image signal,
As a method of increasing the compression ratio, a motion-compensated inter-frame prediction coding (hereinafter, simply referred to as motion-compensated prediction) method is known. In this motion-compensated prediction, the screen is divided into blocks of a certain size (for example, blocks of 16 pixels × 16 lines), and for each block, a block having the smallest prediction error is searched from the prediction screen. The resulting motion vector is sent to the receiving side as information representing the motion of the object on the screen.

[0003] This motion-compensated prediction is usually performed on a prediction screen composed of real pixels, and an integer pixel vector (ie, a vector having no decimal part) is obtained as a motion vector. Such a method of performing prediction in units of integer pixels is called integer pixel motion compensation prediction.

As will be described later, a virtual half pixel (half pel) between real pixels is obtained from surrounding real pixels by interpolation, and the virtual pixel (interpolated pixel) is also used as a prediction screen. Some motion compensation is used, which is called half-pixel (half-pel) motion compensated prediction.

FIG. 7 shows CCITT recommendation H.264. 1 shows a moving image encoding device conforming to H.261. This moving picture coding apparatus uses integer pixel motion compensation prediction as a prediction method, and further includes an in-loop filter in addition to this function.

In FIG. 7, reference numeral 1 denotes a subtractor for taking a difference between an input screen and a prediction screen and outputting a prediction error signal, 2 a DCT circuit for performing a discrete cosine transform on the prediction error signal, and 3 a discrete cosine transform. A quantizer for quantizing the prediction error signal, 4 is an inverse quantizer for inversely quantizing the output signal of the quantizer 3, 5 is an IDCT circuit for performing an inverse discrete cosine transform of the inversely quantized signal, and 6 is an IDCT circuit. An adder for adding a prediction screen to the output signal of the circuit 5 and outputting a locally decoded signal; 7 a frame memory for holding a locally decoded screen in units of frames; and 8 a variable delay circuit for variably delaying the output signal of the frame memory , 9 are in-loop filters for filtering the predicted values output from the variable delay circuit 8, and 13 is for comparing the input screen with the locally decoded screen of the frame memory 7, and operating based on the result. A motion compensation prediction unit that controls the delay amount of the variable delay circuit 8 generates a vector MV.

The in-loop filter 9 reduces the pulse-like signal which is difficult to encode as a prediction error signal by applying a low-pass filter to the locally decoded signal to which the quantization error has been added. Yes, coding efficiency can be improved by controlling ON / OFF of the filter in the loop.

FIG. 8 shows an example of the filter characteristics of the in-loop filter 9. Here, filtering is performed on a block of 8 pixels × 8 lines. Explaining this filtering process, the area (a) at the four corners of the block
Pixels use their own values as they are. Each pixel in the area (b) of the middle six pixels on the first line and the eighth line weights its own pixel by し, and assigns 1 /
The weighted values of 4 are summed to obtain the value of the own pixel after the filtering process. Each pixel in the area (c) of the middle six lines of the first pixel column and the eighth pixel column sums the value obtained by weighting its own pixel by し and weighting the adjacent upper and lower pixels by １ ／. Then, the value of the own pixel after the filtering process is set. Each pixel in the area (d) of 6 pixels × 6 lines inside the block has a weight of 1/4 for its own pixel, a weight of 1/8 for pixels on the left, right, upper and lower sides, and a pixel on the four diagonally adjacent corners , And the values weighted by 1/16 are summed to obtain the value of the own pixel after the filtering process.

There are several ways to control the on / off of the in-loop filter when obtaining the predicted value in the moving picture coding apparatus shown in FIG. As one method of obtaining a sub-optimal prediction value, first, the motion compensation prediction unit 13 determines an optimal integer pixel vector (that is, performs integer pixel motion compensation prediction) as a motion vector. On the other hand, the motion compensation prediction unit 13 creates both predicted values when the in-loop filter is turned on (when the filter operation is performed) and when the filter is turned off (when the filter operation is not performed). Of the input screen, the respective prediction errors are obtained, and the one with the smaller prediction error is selected. If the selected prediction error is the one when the filter in the loop is on, the filter 9 in the loop is actually used. It is controlled to be turned on, and if it is off, it is controlled to turn off the in-loop filter 9.

In the case of the "motion-compensated prediction + intra-loop filter" method, the motion-compensated prediction unit 13 detects a motion vector of an integer pixel and then applies a further intra-loop filter to a prediction block having the motion vector. It is necessary to perform an operation, and an operation circuit for the operation is required.

Therefore, there is an RM8 system as a method for simplifying the on / off control of the in-loop filter. In this manner, the time the motion vector (MV _x, MV _y) obtained by the motion compensation prediction unit 13 is not (0, 0) (i.e., when there is movement), is that to turn the loop filter . This eliminates the need for the in-loop filter operation in the motion compensation prediction unit 13 for the on / off determination of the in-loop filter. However, since the on / off determination is performed without actually performing the filter operation. However, as compared with the above-described method of performing the on / off determination by performing the in-loop filter operation, good characteristics cannot be obtained in the S / N ratio and the like.

Incidentally, the motion compensation prediction method is not limited to the above-described integer pixel motion compensation prediction method, and a virtual 1/2 pixel (half pel) between real pixels is replaced with surrounding real pixels. In addition, there is also motion compensation in which a virtual pixel (interpolated pixel) is also used as a prediction screen, obtained by interpolation, and this is called 1/2 pixel (half-pel) motion compensation prediction. In this 1/2 pixel motion compensation prediction, the motion vector is obtained not only in integer pixel increments but also in 0.5 pixel increments. This 1/2 pixel motion compensation prediction is MPEG
The method is adopted.

FIG. 9 shows a schematic configuration example of a moving picture coding apparatus according to the 1/2 pixel motion compensation prediction method. In the figure, a subtractor 1, a quantizer 3, an inverse quantizer 4, an adder 6,
The frame memory 7 and the variable delay circuit 8 are the same as those in the embodiment of FIG. 7 described above. The difference is that the motion compensation prediction unit performs the motion compensation prediction with 1/2 pixel. 13 'and does not have the in-loop filter 9.

FIG. 10 (I) shows the concept of a prediction screen for this 1/2 pixel motion compensation prediction. In the figure, a triangle mark is an actual integer pixel, and a cross mark is a half pixel obtained by interpolation. (Interpolated pixel). As shown in FIG. 10 (II), when the values of adjacent actual integer pixels are A, B, C, and D, the calculation method for obtaining the half pixel is 1
The values a, b, c, d, and e of the / 2 pixel (interpolated pixel) are obtained according to the following equations. a = (A + B) / 2 b = (A + c) / 2 c = (A + B + C + D) / 4 d = (B + D) / 2 e = (C + D) / 2

[0015]

As described above, the on / off control method of the in-loop filter in the conventional moving picture coding apparatus is performed when the in-loop filter operation is performed for on / off determination. However, there is a problem that the device scale increases due to the necessity of the arithmetic circuit described above, and the characteristic deteriorates when the on / off determination is performed only based on whether or not there is a motion in the motion vector.

When the device is manufactured by LSI or the like, the above-described H.264 standard is used. It is desirable to create a general LSI that can be applied to both the H.261 system and the MPEG system, instead of separately manufacturing them.

The present invention has been made in view of such a problem, and an object of the present invention is to provide the receiving side with integer pixels.
A moving picture coding device that transmits a stepped motion vector
By properly combining the half-pixel motion compensation prediction method and the in-loop filter method, it is possible to perform an approximate operation without specially performing an in-loop filter operation for ON / OFF determination of the in-loop filter. Accordingly, the on / off control of the in-loop filter can be performed, and good characteristics can be obtained without increasing the number of arithmetic circuits.

[0018]

FIG. 1 is an explanatory view of the principle according to the present invention. In FIG. 1, reference numeral 21 denotes a subtractor for calculating a difference between an input screen and a prediction screen and outputting a prediction error signal;
Is a quantizer for quantizing the prediction error signal, 23 is an inverse quantizer for inversely quantizing the output signal of the quantizer 22, and 24 is a local decoded signal obtained by adding a prediction screen to the inversely quantized prediction error signal. , A frame memory for holding the locally decoded signal in frame units, a variable delay circuit for variably delaying the output signal of the frame memory, a filter for filtering the predicted value output from the variable delay circuit , The input screen and the frame memory 25
A motion compensation prediction unit that generates a motion vector MV and controls the variable delay circuit 26 based on the result of comparison with the locally decoded picture of A determination unit that determines on / off of the in-loop filter 27 based on the motion vector from the motion compensation prediction unit 28 and outputs an on / off control signal.

The moving picture coding apparatus according to the present invention, as one mode, performs a motion compensation prediction using 1/2 pixel to perform 1/2.
The half-pixel motion compensation prediction unit 28 that obtains a motion vector of the size of each pixel, the in-loop filter 27 that filters the prediction screen, and the motion vector obtained by the half-pixel motion compensation prediction unit 28 is 1 A determination unit 29 that outputs an on / off control signal for turning on the in-loop filter 27 at the time of a / 2 pixel vector to the in-loop filter 27 ;
1/2 pixel increment obtained by the 1/2 pixel motion compensation prediction unit
Integer pixel increments based on motion vectors of nominal size
And a transmission unit for transmitting the motion vector of

As another mode, the moving picture coding apparatus of the present invention performs a half-pixel motion compensation frame-based prediction using a half pixel to obtain a motion vector having a size of a half pixel. A compensation prediction unit, a filter in a loop for applying a filter to a prediction screen, a motion vector obtained by the half-pixel motion compensation prediction unit is a half-pixel vector, and a norm of the motion vector is a predetermined threshold. ON / OFF to control the in-loop filter when larger
OFF control signal determined in ,, the half pixel motion compensation prediction unit and the determination unit outputs the loop filter 1 /
Integer obtained based on a motion vector with a size of 2 pixels
And a transmission unit for transmitting the motion vector in pixel units to the receiving side .

In the above moving picture coding apparatus, the transmission
Is the half-pixel increment obtained by the half-pixel motion compensation prediction unit.
Truncated the fractional part of the motion vector
It can be configured to convert the motion vector into an integer .

[0022]

According to the present invention, the on / off control of the filter is performed by regarding the calculation for obtaining a 1/2 pixel in the 1/2 pixel motion compensation prediction section 28 as a filter operation in a loop for approximately determining on / off. I do. Considering this, 1 /
When a two-pixel vector is obtained as a motion vector in the two-pixel motion compensation prediction unit 28, the two-pixel motion compensation prediction unit 28 performs a loop when comparing a block in which the in-loop filter operation has been performed on a predicted block with a block in which no prediction is performed. Since it can be considered that the block subjected to the inner filter operation is approximately the same as that in which the prediction error is determined to be smaller, in this case, the determination unit 29 determines the prediction error, and Is turned on. Also, the size of the half-pixel unit of the half-pixel motion compensation prediction unit
Then, a motion vector in increments of integer pixels is obtained based on the motion vector , and transmitted to the receiving side.

The condition for turning on the in-loop filter 27 depends not only on whether the above-mentioned motion vector is a half-pixel vector or an integer pixel vector, but also on the condition that the value of the motion vector is a predetermined value . Control may be performed to turn on the in-loop filter when it is greater than the threshold (ie, only when there is motion), thereby allowing for more appropriate on / off control.

In the transmission section, the 1/2 pixel motion compensation prediction section 2
Trim the decimal part of the motion vector in 1/2 pixel increments found in 8.
Discard the motion vector in increments of integer pixels
The motion vector thus obtained is (0,0)
Approaching a vector, reducing the amount of information to send.
Can be reduced.

[0025]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 2 shows the configuration of a moving picture coding apparatus as one embodiment of the present invention. This embodiment is described in H. "Motion compensation prediction + filter in loop" so that it can be applied to the H.261 system
Has the function of In FIG. 2, a subtractor 1, a DCT
Circuit 2, quantizer 3, inverse quantizer 4, IDCT circuit 5,
The adder 6, the frame memory 7, the variable delay circuit 8, and the in-loop filter 9 have the same functions as those described in the conventional example.

The difference from the conventional example is that the motion compensation prediction unit calculates a motion vector in half-pixel units and performs 1/2 motion compensation prediction.
The on / off of the in-loop filter 9 is determined based on a motion vector (a motion vector having a size of a half pixel) output from the / 2 motion compensation prediction unit 10, and the on / off is determined.
And a filter determination unit 11 for sending an off control signal to the in-loop filter 9.

In the filter determination section 11, the motion vector from the 1/2 motion compensation prediction section 10 is a 1/2 pixel vector (that is, a vector having a size of an integer + 0.5).
The control is performed so that the in-loop filter 9 is turned on when the image data is of the type, and turned off when the image is an integer pixel vector (ie, a vector having an integer size). FIG. 3 shows a configuration example of the filter determination unit 11. Here, the motion vector (MV _x , MV _y ) is composed of m bits, and the least significant bit LSB is the decimal part. Filter determiner 11 fractional part of the X-direction motion vector MV _x (MV
_x -LSB) and fractional part of the Y-direction motion vector MV _y (M
V _y -LSB) is input to each of the OR gates 111.
The output signal of the OR gate 111 is an on / off control signal of the in-loop filter 9.

The filter judging section 11 further includes an integer conversion circuit for converting the input motion vector MV in half-pixel units into a motion vector (IMV) in integer-pixel units. The necessity of this circuit is based on CCITT Recommendation H.264. 2
This is because in 61, the motion vector sent to the receiving side is an integer pixel vector. FIG. 4 shows a configuration example of this integer conversion circuit. This integer conversion circuit performs integer conversion by truncating the decimal part, and this integer conversion circuit is prepared for each of motion vectors in the X direction and the Y direction.

This integer conversion circuit comprises an adder 112 and a selector 113 as shown in the figure. The adder 112
Least significant bit L of input motion vector (m bits)
"1" is added to SB (decimal part) and output. The selector 113 converts the input motion vector (m bits) and the output signal from the adder 112 into input terminals IN0 and IN0, respectively.
The input terminal IN1 is switched to select the input terminal IN0 when the bit is “0” and the input terminal IN1 when the bit is “1” by the most significant bit MSB (sign bit) of the motion vector MV. As a signal, the upper (m-1) bits of the motion vector MV (that is, an integer part excluding the decimal part) are output.

Thus, for example, when the input motion vector is a binary number “000000001” corresponding to a decimal number +3.5.
When the value is 1.1 ", the value is selected by the selector 113, and the value" 00000011 "(=
+3) is output. On the other hand, when the binary number “11111100.1” corresponding to the decimal number −3.5 is output, the LS
A value “11111” obtained by adding “1” to B by the adder 112
101.0 "is selected by the selector 113 and its LSB
Is removed, and the value “11111101” (= −3) is output. As described above, this integer conversion circuit performs integer conversion in the direction close to the (0, 0) vector, and as a result, the number of bits of the motion vector to be transmitted is reduced as much as possible, that is, the amount of transmission information can be reduced as much as possible. Like that.

As described above, the operation of the apparatus according to the present embodiment is as follows. The motion vector MV calculated by the 1/2 motion compensation prediction unit 10 at every 1/2 pixel is calculated by dividing the motion vector MV in at least one of the X direction and the Y direction. If the motion vector is an integer pixel vector, the filter determination unit 11 determines the filter vector, outputs a filter-on control signal to turn on the in-loop filter 9, and if the motion vector is an integer pixel vector, turns off the filter. The control signal is output to turn off the in-loop filter 9.

As described above, the on / off control of the in-loop filter is performed depending on whether the motion vector is a 1/2 pixel vector or an integer pixel vector. The reason for performing the on / off operation is that the function for calculating the in-loop filter (corresponding to FIG. 8) is also a filter function (FIG. 10) for obtaining 1/2 pixel.
8) are both low-pass filter functions.
Fig. 1 instead of the filter function of the filter operation in the loop
By performing on / off determination using a filter function of 1/2 pixel operation of 0 as an approximate function of a filter in a loop,
This is because an approximately quasi-equivalent operation can be performed.

That is, the fact that the motion vector obtained by the 動 き motion compensation prediction unit 10 is a 画素 pixel vector means that a prediction block of an actual integer pixel (that is, a block in which no operation is performed). In other words, the prediction error of a half-pixel prediction block (that is, a block to which an approximate in-loop filter function is applied by a half-pixel calculation) obtained by interpolation has a smaller prediction error,
It is considered that this is approximately the same as the case where it is determined that the prediction error is smaller in the block that has been subjected to the in-loop filter by the conventional method of performing on-off filtering by performing an in-loop filter operation on the predicted block. Therefore, if the on / off control is performed so that the in-loop filter is turned on when the obtained motion vector is a 1/2 pixel vector, it is expected that the coding characteristics will be improved.

FIG. 5 shows a transmission rate S by computer simulation when the moving picture coding apparatus of the present embodiment is used and when the above-described moving picture coding apparatus of the RM system is used.
The N ratio characteristics are shown. In the figure, the horizontal axis represents the transmission rate (× 64 kbps) and the vertical axis represents the S / N ratio (dB). It is of the RM8 type. As is clear from this figure, the method according to the present invention is about 0.2 times less than the RM8 method.
The S / N ratio is improved by .about.0.5 dB, and the effect of the improvement increases particularly as the transmission rate increases. This is because the RM8 system does not perform any in-loop filter operation to determine the on / off state of the filter, while the RM8 system performs this operation approximately by half-pixel operation.
Accordingly, the S / N ratio is improved.

The apparatus of this embodiment is based on H.264. Although the circuit configuration can be applied to the H.261 system, since it has a 1/2 motion compensation prediction unit, it can be easily adapted to the MPEG system by only slightly changing the circuit configuration. H. This is advantageous in that a general-purpose circuit that can support both the H.261 system and the MPEG system can be created with a single LSI.

In implementing the present invention, various modifications are possible. FIG. 6 shows one of such modifications. This example is a modification of the configuration of the filter determination unit of the above-described embodiment. As shown, the X-direction motion vector M
Comparator 114 for comparing the V _x and a predetermined threshold value TH1
If, Y motion vector MV _y a predetermined threshold TH2
And a comparator 115 comparing the X-direction motion vector MV _x
(MV _x −LSB) and Y direction motion vector MV
an OR gate 116 which fractional portion of the _y (MV _y -LSB) is input, the comparators 114, 115 and OR gate 1
An AND gate 117 which receives 16 output signals and outputs an on / off control signal.

This modification is different from the above-described embodiment in that the R
The concept of the M8 system is also adopted, and the magnitude of the motion vector (MV _x , MV _y ) is set to a predetermined threshold value (TH
(1, TH2) or less, the filter is not turned on. That is, the X direction of the motion vector MV, Y
At least one of the magnitudes in the direction is the respective threshold value T
When it is larger than H1 and TH2 and is a 1/2 pixel vector, the control is performed so that the in-loop filter 9 is turned on.

In this modification, whether or not the magnitude of the motion vector exceeds the threshold value is determined by each X-direction component M
V _x and the Y-direction component MV _y are the respective threshold TH _1,
Although it is determined separately whether or not it exceeds TH2, the present invention is not limited to this. For example the absolute value of the motion vector ^{_{| MV | = (MV x 2}} + MV y 2) 1/2 may be performed by whether exceeds a predetermined threshold. In other words, the condition for turning on the in-loop filter may be a time when a 1/2 pixel vector is obtained as the motion vector MV and the norm of the motion vector MV is larger than a predetermined threshold.

[0039]

As described above, according to the present invention, a motion vector having a value in increments of an integer pixel is transmitted to a receiving side.
In a moving picture coding apparatus of this type, the on / off control of the in-loop filter can be performed using the 1/2 pixel motion compensation prediction circuit as it is, so that the in-loop filter operation for the filter on / off determination is specially performed. You do not have to do it. Further, the obtained characteristics are approximately obtained by performing on / off determination by performing an in-loop filter operation, so that the characteristics are improved as compared with the conventional RM8 system, and particularly when the transmission rate is high, the coding efficiency is greatly improved. Become.

Since the apparatus of the present invention includes a half-pixel motion compensation prediction circuit therein, 2
Not only can it be applied to the "motion compensation prediction + filter in loop" method such as the 61 method, but also the MP
The present invention can be easily applied to the EG method and the like, so that the versatility of the device can be increased, which is advantageous when implementing an LSI.

[Brief description of the drawings]

FIG. 1 is an explanatory view of the principle according to the present invention.

FIG. 2 is a block diagram showing a moving picture coding apparatus as one embodiment of the present invention.

FIG. 3 is a block diagram illustrating a configuration of a filter determination unit in the embodiment device.

FIG. 4 is a block diagram illustrating an integer conversion circuit of a filter determination unit in the apparatus according to the embodiment.

FIG. 5 is a diagram showing S / N ratio characteristics by computer simulation of in-loop filter control in a comparison between the method of the present invention and the RM8 method.

FIG. 6 is a block diagram illustrating a configuration of a filter determination unit as another embodiment of the present invention.

FIG. 7 shows a conventional CCITT recommendation H.264; FIG. 1 is a block diagram illustrating a moving image encoding device using the H.261 system.

FIG. 8 is a diagram showing an example of an in-loop filter characteristic in a conventional device.

FIG. 9 is a diagram showing a moving picture encoding device based on the MPEG system.

FIG. 10 is a diagram showing the concept of a prediction screen for 1/2 pixel motion compensation prediction.

[Explanation of symbols]

 Reference Signs List 1 subtracter 2 DCT (discrete cosine transform) circuit 3 quantizer 4 inverse quantizer 5 IDCT (discrete inverse cosine transform) circuit 6 adder 7 frame memory 8 variable delay circuit 9 loop filter 10 1/2 pixel motion compensation Prediction unit 11 filter determination unit 13 motion compensation prediction unit 111, 116 OR gate 112 adder 113 selector 114, 115 comparator 117 AND gate

Continued on front page (72) Inventor Katsuhiro Eguchi 3-22-8 Hakata-ekimae, Hakata-ku, Fukuoka City, Fukuoka Prefecture Inside Fujitsu Kyushu Digital Technology Co., Ltd. (72) Inventor Tsutomu Fujigo 1015 Kamiodanaka, Nakahara-ku, Kawasaki City, Kanagawa Prefecture Address Fujitsu Limited (72) Inventor Kiichi Matsuda 1015 Kamiodanaka, Nakahara-ku, Kawasaki City, Kanagawa Prefecture Fujitsu Limited (56) References JP-A-5-244585 (JP, A) Journal of the Institute of Image Electronics Engineers of Japan, Vol. 20 No. 4, p. 354 (58) Field surveyed (Int. Cl. ⁷ , DB name) H04N ^7/ 24-7/68 G06T 9/00

Claims (3)

(57) [Claims] 1. A motion vector having a size of 1/2 pixel is calculated by performing motion compensation prediction using 1/2 pixel.
A pixel motion compensation prediction unit, an in-loop filter for applying a filter to a prediction screen, and an on switch for turning on the in-loop filter when the motion vector obtained by the 画素 pixel motion compensation prediction unit is a 画素 pixel vector A decision unit for outputting a control signal to the in-loop filter, and a 1/2 pixel obtained by the 1/2 pixel motion compensation prediction unit.
Integer pixel increments calculated based on the motion vector of the increment size
A moving image encoding apparatus comprising: a transmitting unit that transmits only a motion vector to a receiving side . 2. A half-pixel motion-compensated prediction unit for performing motion-compensated inter-frame prediction using half-pixels to obtain a motion vector having a size of 1/2 pixel, and a filter in a loop for filtering a prediction screen When a motion vector is 1/2 pixel vector determined by the half-pixel motion compensation prediction unit, and the motion vector Roh
Determining portion and said half-pixel motion compensation pre that Lum outputs an ON / OFF control signal for controlling turn on the loop filter when greater than a predetermined threshold value in the filter in the loop
The motion vector in the size of 1/2 pixel obtained by the measuring unit
Receives motion vector in integer pixel increments based on torque
Moving picture coding apparatus and a transmission unit for transmitting to the side. 3. The transmission unit is configured to convert a motion vector in 1/2 pixel increments obtained by the 1/2 pixel motion compensation prediction unit into a motion vector in integer pixel increments by truncating its decimal part. The moving picture coding apparatus according to claim 1.