JP3223043B2 - Video encoding device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION The present invention is video conferencing, video telephone, storage media, wireless image transmission, to video encoding KaSo location to be used for such image compression of the video of the broadcast.

[0002]

2. Description of the Related Art FIG. 4 shows a configuration of a conventional moving picture coding apparatus. In FIG. 4, reference numeral 101 denotes a frame drop unit which receives a video signal 151 and a frame drop instruction signal 164 as input, and outputs a video signal 155 after frame drop. 102
Is a frame memory, which receives a reproduction signal 160 and outputs a prediction signal 152. A motion vector detection unit 103 receives the prediction signal 152 and the video signal 155 after the frame has been dropped, and outputs a motion vector 153. A motion compensation unit 104 receives the motion vector 153 and the prediction signal 152, and outputs a motion-compensated prediction signal 154. 1
A subtractor 05 receives a prediction signal 154 after motion compensation and a video signal 155 after frame dropping as inputs, and outputs a difference signal 156.
Is the output. Reference numeral 106 denotes a DCT unit which receives the difference signal 156 as an input and outputs a DCT-difference signal 157. 1
A quantization unit 07 receives the DCT-differential signal 157 and the remaining amount 163 of the transmission buffer as inputs and outputs a quantization coefficient 158. An IDCT unit 108 receives a quantization coefficient 158 as an input and outputs a reproduction difference signal 159. An adder 109 receives the motion compensated prediction signal 154 and the reproduction difference signal 159 and outputs a reproduction signal 160. 11
Reference numeral 0 denotes a variable-length encoding unit that receives a quantization coefficient 158 as an input and outputs an encoded signal 161. Reference numeral 111 denotes a transmission buffer which receives the coded signal 161 as an input, and
62 and the remaining transmission buffer amount 163 are output. Reference numeral 112 denotes a frame drop control unit which receives the transmission buffer remaining amount 163 as an input and outputs a frame drop instruction signal 164.

Next, the operation of the conventional moving picture coding apparatus will be described. In FIG. 4, the frame dropping unit 101 performs frame dropping when the frame dropping instruction signal 164 is positive,
In the case of a negative value, no frame drop is performed, and the video signal 1 after the frame drop is performed.
Output as 55. The frame memory 102 stores the reproduction signal 160, and stores the prediction signal 152 in the next processing frame.
Output as The motion vector detection unit 103 detects a motion vector in a predetermined detection range from the prediction signal 152 and the video signal 155 after the frame is dropped, and outputs a motion vector 153. Based on the motion vector 153, the motion compensation unit 104 outputs a motion-compensated prediction signal 154 whose coordinates are shifted by the amount of the motion vector. The subtractor 105 subtracts the motion-compensated prediction signal 154 from the frame-dropped video signal 155, and outputs a difference signal 156. DCT section 106 converts difference signal 156 into DCT (Di
and outputs a DCT-differential signal 157. The quantization unit 107 quantizes the DCT-processed difference signal 157 using a quantization step size determined by the remaining transmission buffer amount 163, and outputs the result as a quantization coefficient 158. The IDCT unit 108 converts the quantization coefficient 158 into an IDCT (Inverse Discrete Cosine).
Transform) and outputs it as a reproduction difference signal 159.
The adder 109 adds the motion compensated prediction signal 154 and the reproduction difference signal 159, and outputs the result as a reproduction signal 160. The variable-length encoding unit 110 encodes the quantized coefficient 158 and outputs the encoded signal 161. A transmission buffer 111 temporarily stores the encoded signal 161, outputs the encoded signal 162, and outputs the code amount remaining in the transmission buffer as a transmission buffer residual amount 163.
The frame drop control unit 112 outputs a frame drop instruction signal 164 when the transmission buffer remaining amount 163 exceeds a certain threshold.

FIG. 5 shows an example in which no frame is dropped, and FIG. 6 shows an example in which a frame is dropped. In FIG. 1 and 2 are encoded without dropping frames. Frame No. In 3, the transmission buffer exceeds the threshold value for dropping one frame, and issues a frame drop instruction signal for one frame. In the frame drop control unit, the frame N
o. 3 is dropped, and frame no. 4 is encoded.
Therefore, the frame No. During the period 3, the moving picture coding apparatus is stopped. Frame No. In No. 4, since the time given to the encoding is only 1/30 second, the motion detection range cannot be changed.

As described above, in the above-described conventional moving picture coding apparatus, since the motion vector detection range is fixedly determined in advance, it is possible to detect a motion vector in that range and perform coding of a moving picture.

[0006]

However, in the above-described conventional moving picture coding apparatus, when a frame drop occurs, the coding apparatus stops until the frame drop is released, and the time given to motion detection increases. There was a problem that the image quality could not be improved by utilizing this. Despite the time given for motion detection depending on the image size and minimum frame interval, the detection range of the motion vector is fixed, and the image quality is improved by taking advantage of the increased time given for motion detection. There was a problem that it could not be done.

An object of the present invention is to solve such a conventional problem, and an object of the present invention is to provide a moving picture coding apparatus capable of improving image quality.

[0008]

In order to achieve the above object, the present invention provides a detection method for determining a detection range according to an image size.
The motion range detection unit determines the motion vector based on the determined detection range.
And a motion vector detecting unit for performing the detection.
Section shows the detection vector of the motion vector when the image size is reduced.
It is characterized in that motion detection is performed with more complements.

[0009]

Therefore, according to the present invention, the remaining amount of the transmission buffer is detected, the number of dropped frames is determined from the remaining amount of the transmission buffer, and the number of detected vectors is determined according to the image size, the minimum frame interval, and the number of dropped frames. By determining and detecting a motion vector in the determined detection range, a motion vector corresponding to a large motion can be correctly detected, and image quality can be improved.

[0010]

FIG. 1 shows the configuration of an embodiment of the present invention.
In FIG. 1, reference numeral 1 denotes a frame drop unit, which receives a video signal 51 and a frame drop instruction signal 64 as input and outputs a video signal 55 after frame drop. Reference numeral 2 denotes a frame memory which receives a reproduction signal 60 and outputs a prediction signal 52. Reference numeral 3 denotes a motion vector detection unit which receives the prediction signal 52, the video signal 55 after frame dropout, and the detection range 67, and outputs a motion vector 53. Reference numeral 4 denotes a motion compensator, which receives a motion vector 53 and a prediction signal 52 as input, and outputs a motion-compensated prediction signal 54. Reference numeral 5 denotes a subtractor, which is a prediction signal 54 after motion compensation.
And the video signal 55 after frame drop is input, and the difference signal 56
Is the output. Reference numeral 6 denotes a DCT unit which receives a differential signal 56 as an input and outputs a DCT-differential signal 57. Reference numeral 7 denotes a quantization unit which receives the DCT-differential signal 57 and the remaining amount of the transmission buffer 63 as inputs, and outputs a quantization coefficient 58. 8 is ID
In the CT unit, a quantization coefficient 58 is input and a reproduction difference signal 59 is input.
Is output. An adder 9 receives the motion compensated prediction signal 54 and the reproduction difference signal 59 and outputs a reproduction signal 60. Reference numeral 10 denotes a variable-length encoding unit, and a quantization coefficient 58
, And outputs an encoded signal 61. A transmission buffer 11 receives an encoded signal 61 as input, and outputs an encoded signal 62 and a remaining transmission buffer 63. Reference numeral 12 denotes a frame drop control unit, which receives the transmission buffer remaining amount 63 and outputs a frame drop instruction signal 64.

Reference numeral 13 denotes a component detection range determination unit which is a feature of the present embodiment.

Next, the operation of the above embodiment will be described.
In FIG. 1, the frame-dropping unit 1 performs frame-dropping when the frame-dropping instruction signal 64 is positive, and does not perform frame-dropping when the signal is negative, and outputs the video signal 55 after frame-dropping. The frame memory 2 stores the reproduction signal 60 and outputs it as a prediction signal 52 in the next processing frame. The motion vector detection unit 3 detects a motion vector in the detection range from the prediction signal 52, the video signal 55 after frame dropout, and the detection range 67, and outputs a motion vector 53. Based on the motion vector 53, the motion compensation unit 4 outputs a motion-compensated prediction signal 54 whose coordinates are shifted by the amount of the motion vector. The subtractor 5 calculates the motion compensated prediction signal 54 from the video signal 55
Is subtracted, and a difference signal 56 is output. The DCT unit 6 converts the difference signal 56 into a DCT (Discrete Cosine Transform).
Then, the difference signal 57 subjected to DCT is output. Quantization unit 7
Quantizes the DCT-differential signal 57 using a quantization step size determined by the transmission buffer residual amount 63, and outputs it as a quantization coefficient 58. The IDCT unit 8 converts the quantization coefficient 58 into an IDCT (Inverse DiscreteCo
sine Transform) and outputs it as a reproduction difference signal 59. The adder 9 adds the motion-compensated prediction signal 54 and the reproduction difference signal 59 and outputs the result as a reproduction signal 60. The variable-length encoding unit 10 encodes the quantization coefficient 58 and outputs it as an encoded signal 61. 11 is a transmission buffer,
The coded signal 61 is temporarily stored, output as a coded signal 62, and the amount of code remaining in the transmission buffer is output as a remaining transmission buffer amount 63. The piece drop control unit 12
When the transmission buffer remaining amount 63 exceeds a certain threshold value, a frame drop instruction signal 64 is output.

FIG. 2 shows an example of dropping frames. Frame N
o. It is encoded without dropping one or two frames. Frame N
o. In the case of 3, the transmission buffer exceeds the threshold value for dropping one frame, and indicates a frame drop instruction signal for one frame. In the frame drop control unit 12, the frame No. 3 without dropping frames. 4 is dropped. Therefore, the frame No.
No. 3 can encode using the time of 2 frames.

The frame drop detection range determination unit 13 determines a detection range from the image size 65, the maximum frame rate 66, and the frame drop instruction signal 64, and outputs a detection range 67.
FIG. 3 shows an example of this determination method. For example, the image size is FCIF, and the maximum frame rate is 30 frames /
In the case of seconds, if there is no dropped frame, the coding frame interval is 1
/ 30, and if there is a dropped frame, it becomes 2/30, 3/30, 4/30,... According to the indicated amount of the dropped frame. Since the motion detection unit has the ability to detect 225 vectors when the encoding frame interval is 1/30 in FCIF, the encoding frame interval 2/30, In 3/30 and 4/30, the detected vectors are also doubled, tripled and quadrupled.

As described above, the moving picture coding apparatus of this embodiment detects the remaining amount of the transmission buffer, determines the number of dropped frames from the remaining amount of the transmission buffer, and determines the image size, the minimum frame interval, and the number of dropped frames. By determining the number of detection vectors in response to the motion vector and detecting the motion vector in the determined detection range, a motion vector corresponding to a large motion can be correctly detected, and the image quality can be improved.

[0016]

According to the present invention, as is apparent from the description of the above embodiment , the detection range for determining the detection range according to the image size is determined.
Motion vector detection in the determined detection range
A motion vector detection unit for performing the detection,
Is a motion vector detection candidate when the image size is reduced.
The motion detection is performed by increasing the number of motions, and the motion vector corresponding to the large motion can be correctly detected, and the image quality can be improved.

[Brief description of the drawings]

FIG. 1 is a block diagram of a moving picture coding apparatus according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram of frame dropping in the embodiment.

FIG. 3 is an explanatory diagram of a method of determining a motion detection range in the embodiment.

FIG. 4 is a block diagram of a conventional moving picture encoding device.

FIG. 5 is an explanatory diagram in a case where there is no frame drop in a conventional example.

FIG. 6 is an explanatory diagram when a frame is dropped in a conventional example.

[Explanation of symbols]

 1 Frame Dropping Unit 2 Frame Memory 3 Motion Vector Detection Unit 4 Motion Compensation Unit 5 Subtractor 6 DCT Unit 7 Quantization Unit 8 IDCT Unit 9 Adder 10 Variable Length Encoding Unit 11 Transmission Buffer 12 Frame Dropping Control Unit 13 Detection Range Determination Part 51 Video signal 52 Prediction signal 53 Motion vector 54 Prediction signal after motion compensation 55 Video signal after frame dropout 56 Difference signal 57 DCT-differential signal 58 Quantization coefficient 59 Reproduction difference signal 60 Reproduction signal 61 Encoding signal 62 Code Signal 63 Remaining amount of transmission buffer 64 Frame drop instruction signal 65 Image size 66 Maximum frame rate 67 Detection range

Claims (1)

(57) [Claims] 1. A detection range is determined according to an image size.
Detection range determination unit, and motion vector in the determined detection range
A motion vector detecting section for performing detection;
The fixed part detects motion vectors when the image size is reduced.
A moving image characterized by performing motion detection by increasing the number of candidates
Encoding device.