JP2000253404A - Device and method for encoding moving image and moving image processing system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a moving image encoding device, a moving image encoding method and a moving image processing system capable of generating an encoded bit stream capable of executing trans-coding only by the regular rearrangement of data in the encoded bit stream on the encoding side by specifying the format of a decodable code by the information of a decoder which is supplied from the receiving side, stopping part of an encoding format function in accordance with the format of the decodable code and encoding a moving image to a decodable code on the decoding side. SOLUTION: A moving image is encoded to a code to be simply trans-coded to a decoding format of the decoding side by specifying the format of a decodable code on the decoding side by the information (receiving side information) of a decoder which is supplied, e.g. from the receiving side and stopping part of the functions of a quantizer 3 and a movement compensating predictor 9 on the encoding side in according with the format of the decodable code.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a moving picture coding apparatus for coding a moving picture on the coding side of a moving picture processing system for processing moving pictures in different coding formats on the coding side and the decoding side. And a moving image encoding method and a moving image processing system using the same.

[0002]

2. Description of the Related Art In recent years, international standardization of moving picture coding formats has been actively pursued. Such moving picture coding formats currently include ISO (International Org).
announcement for Standardisa
ION) / IEC (International E)
electrotechnical Commissio
n) MPEG-4 (Moving Picture Ex.), which is being standardized as 14496-2
parts group Phase-4), ITU-T (Telecommuni) for videoconferencing, etc.
station Standardization Se
ctor of International Tel
E. communication Union)
261 and H.E. 263 baselines and the like.

FIG. 5 is a block diagram showing a configuration example of a conventional moving picture coding apparatus for coding a moving picture in the MPEG-4 format, for example. FIG. 6 shows the motion compensation predictor 109 of FIG.
FIG. 7 is a block diagram showing a detailed configuration example of FIG.
3 is a block diagram illustrating a detailed configuration example of a quantizer 103 of FIG.

In FIG. 1, reference numeral 1 designates intra-coding or inter-coding, and selectors 11 and 1 are selected accordingly.
2 is a control unit for performing interlock control. Reference numeral 2 denotes a DCT unit that performs a discrete cosine transform on an image signal, and reference numeral 4 denotes a variable-length encoder that performs variable-length encoding on quantized data.

A quantizer 103 quantizes DCT coding coefficients obtained by performing discrete cosine transform on an image signal. For example, intra-AC coefficient adaptive prediction quantization is used for quantization of MPEG-4 DCT encoded coefficients. The quantizer 103 shown in FIG.
This performs C coefficient adaptive prediction quantization.

[0006] In FIG. 7, when quantizing the AC component of each block of the DCT coding coefficient of the intra-coded block, 41 performs prediction from the AC component of the adjacent block.
An AC coefficient predictor 42 calculates the difference between the actual value and the predicted value. Reference numeral 42 denotes a comparator for comparing the amount of code between when the intra AC coefficient prediction is performed and when the intra AC coefficient prediction is not performed. It is a quantization calculator that performs quantization on the smaller of the code amount, either the difference value or the actual value, based on the comparison result.

Reference numeral 5 denotes an inverse quantizer for inversely quantizing the data quantized by the quantizer 103, reference numeral 6 denotes an inverse DCT unit for performing inverse discrete cosine transform of the inversely quantized data, and reference numeral 7 denotes a selector. When the image signal of the previous frame is supplied from the pixel 12, the sum of the image signal and the data from the inverse DCT unit 6 is calculated and output for each pixel, and the image signal of the previous frame is not supplied from the selector 12. In case, reverse D
An adder 8 outputs the data from the CT unit 6 as it is. A frame memory 8 stores at least one frame of image signal.

A motion detection unit 109 detects a motion for each predetermined block based on the image signal of the previous frame stored in the frame memory 8 and the image signal of the current frame among the moving image input signals, and corresponds to the block. This is a motion compensation predictor that reads out an image signal from the frame memory 8 and outputs it. For example, motion compensation prediction in MPEG-4 format is int
er4V motion compensation and the like are employed. The motion compensation predictor 109 in FIG.
It can execute 4V motion compensation and the like.

In FIG. 6, reference numeral 21 designates a control signal for specifying a search range of a motion vector according to a parameter fcode specified based on the MPEG-4 format, and whether or not to round up a fractional part of a prediction signal to be generated. The control unit supplies a control signal or the like to the integer pixel motion compensation predictor 22, the half-integer pixel motion compensation predictor 23, and the motion compensation predictor 24 for inter4V. Reference numeral 23 denotes a half-integer pixel motion compensation predictor that generates a signal and detects a motion vector in half-integer pixel units around a motion vector detected in integer pixel units to generate a prediction signal. 24, the macro block is divided into four blocks, and a set of motion vectors is detected for each of the 8 × 8 pixel blocks. inter to generate a prediction signal to
A 4V motion compensation predictor 25 compares a prediction signal generated by the integer pixel motion compensation predictor 22 and a half integer pixel motion compensation predictor 23 with a prediction signal generated by the inter4V motion compensation predictor 24. , A comparator which selects and outputs the better one of the predetermined evaluation values.

Reference numeral 10 denotes a moving picture input signal and a motion compensation predictor 1
Numeral 13 denotes a differentiator for calculating and outputting a difference of each pixel from a prediction signal from the pixel 09, and dividing the moving image input signal into blocks of a predetermined number of pixels (for example, 8 pixels vertically and horizontally). Is a pre-processing unit that executes.

Next, the operation will be described. The control unit 1
The selectors 11 and 12 are interlocked to perform intra-frame (intra) coding at certain intervals, and to perform inter-frame prediction (inter) coding using motion compensation prediction for the other.

In intra coding, each frame of a moving image input signal is divided into blocks of a predetermined size (for example, 8 × 8 pixels vertically and horizontally) by a preprocessing unit 13. The block of the moving image input signal is supplied to the DCT unit 2 via the selector 11 and is converted into a DCT coded coefficient. The DCT coded coefficient is quantized by the quantizer 103 and is converted into a quantization index. Then, by the variable length encoder 4,
The Huffman code is assigned to the quantization index and output as an encoded bit stream.

The quantization index is inversely quantized by an inverse quantizer 5 and further subjected to an inverse discrete cosine transform by an inverse DCT unit 6, and then a frame memory 8 is used as a part of a prediction reference frame used in inter-frame motion compensation prediction. Is stored.

On the other hand, in the inter coding, each frame of the moving image input signal is divided into macroblocks of a predetermined size (for example, 16 × 16 pixels in length and width) by the preprocessing unit 13. Then, the motion compensation predictor 109 refers to the prediction reference frame stored in the frame memory 8 and detects a motion vector for each macroblock as information representing the motion of the object on the screen. Is generated.

The difference between each macroblock of the moving picture input signal and the corresponding prediction signal is calculated by a differentiator 10, and the difference is converted by a DCT unit 2 to obtain a DCT.
A T-coded coefficient is generated. Then, the DCT coded coefficients are quantized by the quantizer 103, and the Huffman code is assigned to the generated quantization index and motion vector by the variable length coder 4, and is output as a coded bit stream.

As in the case of the intra coding, the quantization index is inversely quantized by the inverse quantizer 5 and further subjected to inverse discrete cosine transform by the inverse DCT unit 6. And
An adder 7 adds a prediction signal to the data after the inverse discrete cosine transform, and the calculation result is stored in the frame memory 8 as a part of the next prediction reference frame.

Next, a detailed operation of the motion compensation predictor 109 in FIG. 6 will be described. The control unit 21 uses the MPEG-4
A control signal for specifying a search range of a motion vector according to a parameter fcode specified based on the format, a control signal for specifying whether or not to round up a fractional part of a generated prediction signal, and the like are converted to an integer pixel motion compensation predictor 22. , A half-integer pixel motion compensation predictor 23 and a motion compensation predictor 24 for inter4V.

The integer pixel motion compensation predictor 22 detects a motion vector in integer pixel units to generate a prediction signal, and the half-integer pixel motion compensation predictor 23 generates a prediction signal around the motion vector detected in integer pixel units. , A motion vector is detected in half-integer pixel units to generate a prediction signal. On the other hand, in
The ter4V motion compensation predictor 24 divides the macroblock into four blocks, detects a set of motion vectors for each of the 8 × 8 pixel blocks in the vertical and horizontal directions, and generates a prediction signal. Motion compensation). Then, the comparator 25 compares the prediction signal generated by the integer pixel motion compensation predictor 22 and the half-integer pixel motion compensation predictor 23 with the motion compensation predictor 24 for inter4V.
Are compared, and the better one of the predetermined evaluation values is selected and output. As described above, in the ordinary coding apparatus, the motion compensation prediction is first performed by the integer pixel motion compensation predictor 22 in units of integer pixels in order to reduce the amount of computation.
In many cases, motion compensation is performed on the periphery by the half-integer pixel motion compensation predictor 23 in half-integer pixel units.

As described above, according to the MPEG-4 format, the detection of a motion vector and the inter
4V motion compensation is feasible. In addition, MPEG-4
According to the format, the search range of the motion vector can be appropriately set, and the reference block to be compared when generating the prediction signal extends from the inside of the frame to the outside without being limited to the inside of the reference image frame. (This motion compensation is called unrestricted motion compensation). Further MP
According to the EG-4 format, it is possible to appropriately set whether to round down or round up the fractional part of the prediction signal for a half-integer pixel.

On the other hand, H. H.261 format and H.264 format. In a moving picture coding apparatus that performs coding in the H.263 baseline format, the range of motion prediction is limited to ± 15, and a block to be referred to for motion compensation is limited to the inside of the reference picture frame. H. 261 and H.E. In the 263 baseline format, no inter4V motion compensation is performed, and one set of motion vectors is calculated for a macroblock having a size of 16 × 16 pixels.

H. According to the H.261 format, the motion compensation prediction is performed on a prediction screen composed of real pixels, the motion vector is an integer pixel vector (that is, a vector having no decimal part), and the motion compensation prediction is performed in integer pixel units. (This motion compensation is called integer pixel motion compensation prediction).

Further, H. In the H.263 baseline format, similarly to the MPEG-4 format, the value of a virtual pixel between real pixels is calculated from the values of surrounding real pixels by interpolation, and the value of the virtual pixel (interpolated pixel) is also used as a prediction signal. Motion compensation is performed in units of half-integer pixels using this method (this motion compensation is called half-integer pixel motion compensation prediction). At this time, the color difference components are also stored in MPEG-4 format or H.264 format. In the 263 baseline format, a motion vector is detected in units of half integer pixels.
H. In the H.261 format, detection is performed in integer pixel units. Also,
H. In the H.263 baseline format, if the predicted signal for the interpolated half-integer pixel is not an integer, the fractional part is always rounded up.

As described above, the presence or absence of various functions in motion compensation differs depending on the coding format used.

Next, a detailed operation of the quantizer 103 shown in FIG. 7 will be described. When quantizing the AC component of each block of the DCT coding coefficient of the intra-coded block, the AC coefficient predictor 41 performs prediction from the AC component of the adjacent block and calculates the difference between the actual value and the predicted value. I do. Then, the comparator 42 compares the code amount when the intra AC coefficient prediction is performed with the code amount when the intra AC coefficient prediction is not performed.
Performs quantization on the difference value or the actual value, whichever is smaller, based on the comparison result of the comparator 42.
As described above, by selecting the one with the smaller code amount, the coding efficiency can be increased.

As described above, the moving picture coding formats which are international standards have different functions, respectively.
When transmitting and receiving a moving image signal in a communication system that uses different moving image coding formats at the transmitting side encoder and the receiving side decoder, the coded bit stream is processed when processing the moving image signal. A transcoder for conversion is required.

FIG. 8 is a block diagram showing an example of a moving image processing system for processing moving images of different encoding formats on the encoding side and the decoding side. In the moving image processing system shown in FIG. 8, an MPEG-4 encoder 201 that encodes a moving image in the MPEG-4 format and outputs it as an MPEG-4 bit stream is used as a transmitting side encoder. H. H.261 encoded in H.261 format
H.261 decoding a moving image from a H.261 bit stream 2
61 decoder 203 is used. In this case,
The transcoder 202 converts the MPEG-4 bit stream into an H.264 image. 261 bit stream. In the transcoder 202, the moving image is once decoded from the MPEG-4 bit stream by the MPEG-4 decoder 211, and the decoded moving image H.261 by the H.261 encoder 212. 261 bit stream.

[0027]

Since the conventional moving picture processing system is configured as described above, in the transcoder, the inverse DCT is used for decoding the encoded bit stream in the MPEG-4 format from the transmission side. And a frame memory are required. In order to encode a moving image in the H.261 format, a DCT unit, an inverse DCT unit, a frame memory, a motion compensation predictor, and the like are required, and it is difficult to reduce the circuit scale and cost of the transcoder. However, there is a problem that it is difficult to reduce the processing time for transmitting and receiving a moving image due to the occurrence of delay time due to the decoding process and the encoding process.

The present invention has been made in order to solve the above-described problems, and specifies the format of a code that can be decoded based on, for example, information of a decoder supplied from a receiving side.
Depending on the format of the code that can be decoded, a part of the function of the coding format on the encoding side is stopped, and the moving image is encoded into a code that can be decoded on the decoding side. Obtaining a moving picture coding apparatus, a moving picture coding method, and a moving picture processing system that generate a coded bit stream on the coding side capable of performing transcoding only by regular rearrangement of data in the bit stream With the goal.

[0029]

A moving picture coding apparatus according to the present invention comprises: a code form specifying means for specifying a code form which can be decoded on the decoding side; An encoding unit for stopping a part of the function of the encoding format on the encoding side according to the code format and encoding the moving image into a code that can be decoded on the decoding side.

In the moving picture coding apparatus according to the present invention, the coding means has a function of coding a moving picture in one of the MPEG formats.

In the moving picture coding apparatus according to the present invention, when it is specified that the decoding format on the decoding side is to be decoded without using inter4V motion compensation, inter
The 4V motion compensation function is stopped to encode a moving image.

The video encoding apparatus according to the present invention, when it is specified that the decoding format on the decoding side is to be decoded based on a motion vector only within a predetermined pixel range,
The function of searching for a motion vector outside a predetermined pixel range is stopped to encode a moving image.

The video encoding apparatus according to the present invention, when it is specified that the decoding format on the decoding side is to perform motion compensation on the basis of a motion vector in units of integer pixels and to decode the half-integer pixels, The function of detecting a unit motion vector is stopped to encode a moving image.

The moving picture coding apparatus according to the present invention, when it is specified that the decoding format on the decoding side is to be decoded without relying on motion compensation by a motion vector from outside the frame, In this case, the function of detecting a motion vector is stopped to encode a moving image.

In the moving picture coding apparatus according to the present invention, it is specified that the decoding format on the decoding side performs motion compensation for each integer pixel and decodes, and the coding format on the coding side is a half integer pixel. In the case of encoding with motion compensation for each pixel, the function of searching for a motion vector in one of the odd-numbered pixel and the even-numbered pixel is stopped, and the motion vector is searched for in the other pixel. In this case, a moving image is encoded.

In the moving picture coding apparatus according to the present invention, when the decoding format on the decoding side performs motion compensation for each half-integer pixel and decodes and interpolates the half-integer pixel value from the integer pixel value. When it is specified that the fractional part of the value of the half-integer pixel is rounded up, the function of other motion compensation prediction is stopped, and the value of the half-integer pixel is used when interpolating the half-integer value from the integer pixel value. The motion compensation image is encoded by performing motion compensation prediction by rounding up the decimal part of.

The moving picture coding apparatus according to the present invention, when the decoding format on the decoding side is determined to be a code that has been subjected to the intra AC coefficient adaptive predictive quantization and cannot be inversely quantized, the intra AC coefficient adaptive The function of predictive quantization is stopped, quantization is performed, and a moving image is encoded.

In the moving picture coding method according to the present invention, the step of specifying the format of the code that can be decoded on the decoding side, and the step of specifying the coding format on the coding side according to the specified format of the code that can be decoded. Stopping a part of the function and encoding the moving image into a code that can be decoded on the decoding side.

A moving picture processing system according to the present invention comprises a code format specifying means for specifying a code format that can be decoded on the decoding side, and a code corresponding to the format of the decodable code specified by the code format specification device. A moving image encoding apparatus including an encoding unit that stops a part of the function of the encoding format on the encoding side, encodes the moving image to a code that can be decoded on the decoding side, and outputs an encoded bit stream. A video decoding device that decodes an encoded bit stream in a format that can be decoded on the decoding side, and based on the difference in the data order of the encoded bit stream depending on the decoding format on the encoding side and the decoding side, The encoded bit stream output by the video encoding device is rearranged and converted into an encoded bit stream in a format that can be decoded on the decoding side, and the converted encoded bit stream is converted. In which and a transcoder supplied to the video decoding apparatus.

[0040]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below. Embodiment 1 FIG. In the first embodiment, the MPEG-4 format is used as the encoding format on the encoding side (transmission side), and H.264 is used as the encoding format on the decoding side (reception side). H.261 format or H.261 format. Although the H.263 baseline format is exemplified, the types of the encoding format on the encoding side (transmission side) and the encoding format on the decoding side (reception side) are not limited to these.

FIG. 1 is a block diagram showing a configuration example of a moving picture coding apparatus according to Embodiment 1 of the present invention. FIG. 2 is a block diagram showing a detailed configuration example of the motion compensation predictor 9 in the moving picture encoding device of FIG. 1, and FIG. 3 is a block diagram showing a detailed configuration example of the quantizer 3 of FIG. .

In the drawing, reference numeral 1 designates intra-coding or inter-coding, and the selectors 11 and 1 are selected accordingly.
2 is a control unit for performing interlock control. Reference numeral 2 denotes a DCT unit that performs a discrete cosine transform on an image signal, and reference numeral 4 denotes a variable-length encoder that performs variable-length encoding on quantized data.

Numeral 3 denotes a format in which the image signal can be inversely quantized on the decoding side in accordance with the information on the receiving side.
This is a quantizer that quantizes the T encoding coefficient. MPEG-
Intra-AC coefficient adaptive prediction quantization is used for quantization of the four types of DCT coding coefficients. Quantizer 3 of FIG.
Is for performing intra AC coefficient adaptive prediction quantization in accordance with the MPEG-4 format.

The receiving side information is, for example, an ITU-T
Recommendation H. H.245 format or the like, and is supplied as a control signal for notifying the receiving capability and the like exchanged between the transmitting side and the receiving side. Also, for example, when a telephone line is used for communication, the information on the receiving device and the telephone number of the receiving side are recorded in advance, and the telephone number of the receiving side is supplied as the receiving side information. Is also good.

In FIG. 3, when quantizing the AC component of each block of the DCT coding coefficient of the intra-coded block, prediction is performed from the AC component of the adjacent block.
An AC coefficient predictor 42 calculates the difference between the actual value and the predicted value. Reference numeral 42 denotes a comparator for comparing the amount of code between when the intra AC coefficient prediction is performed and when the intra AC coefficient prediction is not performed. It is a quantization calculator that performs quantization on the smaller of the code amount, either the difference value or the actual value, based on the comparison result. Reference numeral 71 denotes a control unit (code format specifying unit) that controls the switch 72 based on the reception side information and switches whether to supply the DCT coding coefficient to the AC coefficient predictor 41.

Reference numeral 5 denotes an inverse quantizer for inversely quantizing the data quantized by the quantizer 3, reference numeral 6 denotes an inverse DCT unit for performing inverse discrete cosine transform on the inversely quantized data,
When the image signal of the previous frame is supplied from the selector 12, the sum of the image signal and the data from the inverse DCT unit 6 is calculated and output for each pixel, and the image signal of the previous frame is output from the selector 12. When not supplied, an adder that outputs the data from the inverse DCT unit 6 as it is, and 8 is a frame memory that stores at least one frame of image signal.

Numeral 9 denotes a format in which motion can be compensated on the decoding side in accordance with the information on the receiving side. This is a motion compensation predictor that performs motion detection for each predetermined block based on the block, and reads out and outputs an image signal corresponding to the block from the frame memory 8. M
PEG-4 type motion compensation prediction employs inter4V motion compensation or the like. The motion compensation predictor 9 shown in FIG. 2 is capable of executing inter4V motion compensation or the like according to the MPEG-4 format.

In FIG. 2, reference numeral 21 denotes a control signal for designating a search range of a motion vector according to a parameter fcode designated based on the MPEG-4 format, and whether or not to round up a fractional part of a prediction signal to be generated. A control unit that supplies a control signal or the like to be specified to the control unit 51. Reference numeral 22 denotes an integer pixel motion compensation predictor that detects a motion vector in integer pixel units and generates a prediction signal. Reference numeral 23 denotes a detection unit in integer pixel units. Is a half-integer pixel motion compensation predictor that detects a motion vector in units of half-integer pixels around the generated motion vector and generates a prediction signal. Is a motion compensation predictor for inter4V that detects a set of motion vectors for each block and generates a prediction signal. Prediction signal generated by the measuring unit 22 and half-integer pixel motion compensation predictor 23 and inter4
This is a comparator that compares the prediction signals generated by the V motion compensation predictor 24 and selects and outputs the better one of the predetermined evaluation values.

Reference numeral 51 denotes a switch 52 based on information on the receiving side.
And the selector 53, the integer pixel motion compensation predictor 22, the half-integer pixel motion compensation predictor 23, and
After switching the use / non-use of the 4V motion compensation predictor 24 and converting the control signal from the control unit 21 based on the receiving side information, the integer pixel motion compensation predictor 2
2. Half-integer pixel motion compensation predictor 23 and inter4
A control unit (code format specifying unit) that supplies the motion compensation predictor 24 for V.

Reference numeral 10 denotes a moving picture input signal and a motion compensation predictor 9
13 is a differentiator for calculating and outputting a difference for each pixel from a prediction signal from the video signal, and 13 performs preprocessing such as dividing a moving image input signal into blocks of a predetermined number of pixels (for example, 8 pixels vertically and horizontally). This is a preprocessing unit to be executed.

The control section 1 to the preprocessing section 13 constitute an encoding means.

Next, the operation will be described. In the first embodiment, for example, the telephone number of the receiving side or the H.264 standard is used. 245
A control signal exchanged between the transmitting side and the receiving side based on the format is supplied to the control unit 71 of the quantizer 3 and the control unit 51 of the motion compensation predictor 9 as receiving side information.

The control unit 51 of the motion compensation predictor 9 specifies an encoding format that can be decoded on the decoding side based on the information on the receiving side, and controls the switch 52 and the selector 53 in accordance with the information to perform integer pixel motion compensation. The use / non-use of the predictor 22, the half-integer pixel motion compensated predictor 23, and the inter4V motion compensated predictor 24 are switched.

For example, H. When the coding format of the code that can be decoded on the decoding side (that is, the decoding format on the decoding side) does not perform the inter4V motion compensation prediction as in the H.261 format, the control unit 51 sets the switch 52 to In the off state, the motion compensation predictor 24 for the inter4V is not used, and the motion compensation prediction is performed for each macroblock. That is, in this case, the inter4V motion compensation prediction is not performed. On the other hand, when the decoding format on the decoding side performs inter4V motion compensation prediction, such as the MPEG-4 format, for example, the control unit 51 sets the switch 52 to the ON state and sets the inter4V motion compensation predictor 24. To execute the inter4V motion compensation prediction. Thus, use / non-use of the inter4V motion compensation predictor 24 is switched according to the reception side information.

Further, the control unit 51 outputs a control signal for limiting the motion vector search range to a predetermined range in accordance with the receiving side information, by the integer pixel motion compensation predictor 22 and the integer.
The motion vector is supplied to the r4V motion compensation predictor 24 to search for a motion vector outside a predetermined range.

For example, H. When the decoding format on the decoding side, such as the H.261 format, specifies that the search range of the motion vector is limited to ± 15, the control unit 51
From the motion vector search range.
The control signal for restricting the control signal to a range of 15 pixels is an integer pixel motion compensation predictor 22 and an inter4V motion compensation predictor 2
4 to stop the function of searching for a motion vector outside the range of ± 15 pixels, and to search for a motion vector within the range of ± 15 pixels.

When the decoding format on the decoding side specifies that the chrominance component is to be decoded based only on the motion vector in units of integer pixels, the control unit 51 Change the control signal and search for the motion vector. For the luminance component, the even-numbered (or odd-numbered)
The control signal to be performed only for the pixel is supplied, and the motion vector of the color difference component is searched in such a manner. Thereby, a motion vector of a color difference component in an integer pixel unit is detected.

When it is determined that the decoding format on the decoding side does not execute decoding based on a motion vector from outside the frame, the control unit 51
And a control signal for stopping the function of detecting a motion vector from outside the frame is supplied. Thereby, a motion vector is detected only within a frame.

For example, H. When it is specified that the decoding format on the decoding side does not perform the half-integer pixel motion compensation prediction like the H.261 format, the control unit 51 causes the selector 53 to select the comparator 25 side and The pixel motion compensation prediction unit 23 is not used, and the integer pixel motion compensation prediction is performed using only the integer pixel motion compensation prediction unit 22. On the other hand, if the decoding format on the decoding side performs half-integer pixel motion compensation prediction, such as the MPEG-4 format, the control unit 51 causes the selector 53 to select the half-integer pixel motion compensation predictor 23 Is selected and the half-integer pixel motion compensation prediction is performed using the half-integer pixel motion compensation predictor 23. The half-integer pixel motion compensation predictor 2 according to the function of the decoding side specified based on the information of the receiving side as described above.
3 is switched between use and non-use.

The control unit 51 supplies a "rounding_type" flag having a value based on the receiving side information to the half-integer pixel motion compensation predictor 23 as a control signal, and uses a function of always rounding up a decimal part of the prediction signal. Specify whether or not.

For example, H. If the decoding format on the decoding side, such as the H.263 baseline format, is specified to always round up a small part of the prediction signal calculated by interpolation from real pixels in half-integer pixel motion compensation prediction, the control unit 51 Supplies a "rounding_type" flag with a value of 0 to restrict the fractional part of the prediction signal calculated by interpolation to always be rounded up.

On the other hand, the control unit 71 of the quantizer 3 controls the switch 72 based on the information on the receiving side to switch between using and not using the AC coefficient predictor 41. For example, H. 261
When it is determined that the decoding format on the decoding side does not perform the AC coefficient prediction in the inverse quantization like the format, the control unit 71 sets the switch 72 to the OFF state and sets the AC
The coefficient predictor 41 is not used, and the AC coefficient prediction in the quantization is stopped.

In this way, the various functions of the motion compensation predictor 9 and the quantizer 3 are stopped according to the information on the receiving side. After the use / non-use of various functions is set, encoding of a moving image is executed.

That is, first, the control unit 1 selects the selector 1
1 and 12 are interlocked, and intra-frame (intra) coding is performed at certain intervals, and inter-frame prediction (inter) coding using motion compensation prediction is performed for the other.

In the intra coding, each frame of the moving image input signal is divided into blocks of a predetermined size (for example, 8 × 8 pixels vertically and horizontally) by the preprocessing unit 13. The block of the moving image input signal is supplied to the DCT unit 2 via the selector 11 and is converted into a DCT coding coefficient. The DCT coding coefficient is quantized by the quantizer 3 and converted into a quantization index. At this time, quantization is performed by a preset function. Then, the Huffman code is assigned to the quantization index by the variable length encoder 4 and output as an encoded bit stream.

After the quantization index is inversely quantized by the inverse quantizer 5 and further subjected to inverse discrete cosine transform by the inverse DCT unit 6, the frame memory 8 is used as a part of the prediction reference frame used in the inter-frame motion compensation prediction. Is stored.

On the other hand, in the inter coding, each frame of the moving image input signal is divided into macroblocks of a predetermined size (for example, 16 × 16 pixels in length and width) by the preprocessing unit 13. Then, the motion compensation predictor 9 refers to the prediction reference frame stored in the frame memory 8 to detect a motion vector for each macroblock as information representing the motion of the object on the screen, Is generated. At this time, motion compensation prediction is performed using a preset function.

Then, the difference between each macroblock of the moving picture input signal and the corresponding prediction signal is calculated by the differentiator 10, and the difference is converted by the DCT unit 2 to obtain a DCT.
A T-coded coefficient is generated. Then, the DCT coded coefficients are quantized by the quantizer 3, the Huffman code is assigned to the generated quantization index and the motion vector by the variable length coder 4, and the resultant is output as a coded bit stream. At this time, similarly, the quantization is executed by a preset function.

As in the case of the intra coding, the quantization index is inversely quantized by the inverse quantizer 5 and further subjected to inverse discrete cosine transform by the inverse DCT unit 6. And
An adder 7 adds a prediction signal to the data after the inverse discrete cosine transform, and the calculation result is stored in the frame memory 8 as a part of the next prediction reference frame.

The coding of the moving picture is executed in this manner. If only the use / non-use of various functions of the motion compensation prediction is set based on the receiving side information,
The quantizer 3 may be a conventional quantizer 103. Similarly, when only the use / non-use of various quantization functions is set based on the reception side information, the motion compensation predictor 9 may be the conventional motion compensation predictor 109.

As described above, according to the first embodiment, the format of a code that can be decoded is specified based on the receiving side information.
Depending on the format of the code that can be decoded, some of the functions of the coding format on the encoding side are stopped and the moving image is encoded into a code that can be decoded on the decoding side. This makes it possible to generate an encoded bit stream capable of performing transcoding only by regular rearrangement of data in the encoded bit stream.

Also, according to the receiving side information,
Since the function of motion detection that calculates a motion vector in units of blocks defined by motion compensation is stopped, in
coded bits that can be transcoded into a coded bit stream that can be decoded by a decoder that does not decode a motion vector in units of blocks specified by ter4V motion compensation by only regular rearrangement of data The effect that a stream can be generated is obtained.

Further, since the search range of the motion vector is limited according to the information on the receiving side, the search range of the motion vector on the encoding side is larger than the search range of the motion vector allowed on the decoding side. Also, the coded bits that can be transcoded into a coded bit stream that can be decoded by a decoder that decodes based on the motion vector in the search range only by regular rearrangement of data The effect that a stream can be generated is obtained.

Further, since the function of motion detection for calculating a motion vector in half-integer pixels in accordance with the information on the receiving side is not used, decoding is performed by a decoder that does not decode motion vectors in half-integer pixels. An effect is obtained that an encoded bit stream can be generated in which transcoding into a possible encoded bit stream can be performed only by regular rearrangement of data.

Further, since the function of detecting a motion vector from outside the frame according to the information on the receiving side is not used, a code that can be decoded by a decoder that decodes based only on the motion vector within the frame is used. This makes it possible to generate an encoded bit stream that can be transcoded into an encoded bit stream only by regular rearrangement of data.

Further, since the search for the motion vector for the odd-numbered or even-numbered pixels is stopped for the luminance value component according to the receiving side information, the color difference component is determined based on the motion vector in integer pixel units. An effect is obtained that an encoded bit stream can be generated that can be transcoded into an encoded bit stream that can be decoded by a decoder that performs decoding only by regular rearrangement of data. Can be

Further, according to the information on the receiving side, "round"
Since the value of the “ing_type” flag is changed to set whether to always round up the minority part of the prediction signal, an encoded bit stream that can be decoded by a decoder that always rounds up the minority part of the prediction signal This makes it possible to generate an encoded bit stream that can be transcoded only by regular rearrangement of data.

Further, since the prediction of the intra AC coefficient in the quantization is stopped according to the information on the receiving side, the transcoding to the coded bit stream that can be decoded by a decoder that does not perform the intra AC coefficient prediction is performed. Thus, it is possible to generate an encoded bit stream that can be performed only by the regular rearrangement.

Embodiment 2 FIG. 4 is a block diagram showing a configuration example of a moving image processing system according to Embodiment 2 of the present invention. In the figure, reference numeral 81 denotes a transmitting side encoder (moving image encoding apparatus) to which the moving image encoding apparatus according to the first embodiment is applied, and reference numeral 82 denotes codes of the transmitting side encoder 81 and the receiving side decoder 83. Based on the difference in the data order of the encoded bit stream depending on the encoding format,
, Rearranges the data in the coded bit stream, converts the coded bit stream into a coded bit stream in a format that can be decoded by the receiving side decoder 83, and supplies the converted coded bit stream to the receiving side decoder 83. Rearranger 9
1 is a transcoder, and 83 is a receiving side decoder (moving image decoding device) for decoding an encoded bit stream of a predetermined encoding format.

Next, the operation will be described. For example, H. 2
The receiver-side decoder 83 outputs the receiver-side information such as the coding format of the code that can be decoded based on the 45-format.
Feed to 1.

The transmission-side encoder 81 executes the encoding of the moving image based on the supplied reception-side information as described above (Embodiment 1), and converts the generated encoded bit stream into a transcoded image. It is supplied to the coder 82. The data reordering unit 91 of the transcoder 82 changes the order of the data in the supplied encoded bit stream, converts the data into an encoded bit stream in a format that can be decoded by the receiver-side decoder 83, and performs encoding after the conversion. The bit stream is supplied to the receiving side decoder 83. Then, the receiving side decoder 83
Decodes a coded bit stream of a predetermined coding format from the transcoder 82 and outputs a moving image.

As described above, according to the second embodiment, the transmitting side (encoding side) encoding in a predetermined encoding format
When performing communication between a receiving side (decoding side) that decodes in a different decoding format and a function that is provided on the transmitting side but not provided on the receiving side, Did not use that function to perform encoding, and rearranged the data to convert the coded bit stream into a coded bit stream in the decoding format on the decoding side. The effect is obtained that the processing and circuit scale required for transcoding and the required time can be reduced.

In the above embodiment, an encoder of the MPEG-4 format is used as an encoder, and H.264 is used as a decoder. H.261 format or H.261 format. The H.263 decoder has been described by way of example, but the encoding and decoding formats applicable to the present invention are not limited to these. For example, MPEG-1 format, MPEG-2 format And so on.

In the above embodiment, integer pixel motion compensation prediction and half-integer pixel motion compensation prediction are performed in series, and integer pixel motion compensation prediction and inter4V motion compensation prediction are performed in parallel. However, the present invention can be applied. The form of the motion compensation prediction process is not limited to this.

The present invention is not limited to the above-described embodiment, but also includes those which are easily changed or modified by those skilled in the art.

[0086]

As described above, according to the present invention, the code format that can be decoded on the decoding side is specified, and the coding format on the coding side is specified according to the specified code format that can be decoded. Since a part of the functions is stopped and the moving image is encoded to a code that can be decoded on the decoding side, transcoding can be performed only by regular rearrangement of data in the encoded bit stream. The effect is that a possible encoded bit stream can be generated.

According to the present invention, when it is specified that the decoding format on the decoding side is to be decoded without using the inter4V motion compensation, the function of the inter4V motion compensation is stopped to encode the moving picture. So that
coded bits that can be transcoded into a coded bit stream that can be decoded by a decoder that does not decode a motion vector in block units specified by inter4V motion compensation, by only regular rearrangement of data There is an effect that a stream can be generated.

According to the present invention, when it is specified that the decoding format on the decoding side is to be decoded based on a motion vector only within a predetermined pixel range, the motion vector is determined outside the predetermined pixel range. Since the search function is stopped and the moving image is encoded, even when the search range of the motion vector on the encoding side is larger than the search range of the motion vector allowed on the decoding side, the search range Generating a coded bit stream that can be transcoded into a coded bit stream that can be decoded by a decoder that performs decoding based on the motion vector by only regular reordering of data. There is an effect that can be.

According to the present invention, when it is specified that the decoding format on the decoding side is to perform motion compensation based on a motion vector in integer pixel units and to perform decoding, a motion vector detection in half integer pixel units is performed. Function is stopped and the moving picture is coded, so that the transcoding into a coded bit stream that can be decoded by a decoder that does not decode the motion vector in half-integer pixels is performed regularly. There is an effect that it is possible to generate an encoded bit stream that can be performed only by simple rearrangement.

According to the present invention, when it is determined that the decoding format on the decoding side is to be decoded without relying on motion compensation using a motion vector from outside the frame, a motion vector from outside the frame is detected. Is configured to stop the function of encoding the moving image, so that the decoder that decodes based on the motion vector in the frame regularly transcodes the data into an encoded bit stream that can be decoded. There is an effect that it is possible to generate an encoded bit stream that can be performed only by simple rearrangement.

According to the present invention, it is specified that the decoding format on the decoding side performs motion compensation for each integer pixel for decoding, and the coding format on the encoding side performs motion compensation for each half integer pixel. In the case of encoding, the function of searching for a motion vector in one of the odd-numbered pixel and the even-numbered pixel is stopped, and a motion vector is searched for in the other pixel to encode a moving image. For example, for a chrominance component, transcoding into a coded bit stream that can be decoded by a decoder that performs decoding based on a motion vector in units of integer pixels is performed using a regular arrangement of data. There is an effect that it is possible to generate an encoded bit stream that can be performed only by permutation.

According to the present invention, when the decoding format on the decoding side is motion compensated for each half-integer pixel and decoded, and when the value of the half-integer pixel is interpolated from the value of the integer pixel, the decoding format of the half-integer pixel is If it is specified to round up the decimal part of the value, stop the motion compensation prediction function and round up the decimal part of the half-integer pixel value when interpolating the half-integer pixel value from the integer pixel value Since it is configured to encode a moving image by performing motion compensation prediction, it is necessary to transcode into a coded bit stream that can be decoded by a decoder that always rounds up a small part of a prediction signal calculated by interpolation. There is an effect that an encoded bit stream that can be performed only by dynamic rearrangement can be generated.

According to the present invention, when the decoding format on the decoding side is determined to be a code that has been subjected to intra AC coefficient adaptive predictive quantization and cannot be inversely quantized, the function of intra AC coefficient adaptive predictive quantization is performed. Is stopped, quantization is performed, and a moving image is coded. Therefore, transcoding into a coded bit stream that can be decoded by a decoder that does not perform intra AC coefficient prediction in quantization is performed according to data rules. There is an effect that an encoded bit stream that can be performed only by dynamic rearrangement can be generated.

According to the present invention, the code format specifying means for specifying the format of a code that can be decoded on the decoding side, and the code format on the coding side according to the format of the code that can be decoded specified by the code format specifying device. Stop some features of the encoding format,
A moving image encoding apparatus including an encoding unit that encodes a moving image to a code that can be decoded on the decoding side and outputs an encoded bit stream, and decodes an encoded bit stream in a format that can be decoded on the decoding side. A video decoding device that performs encoding, and arranges the data of the encoded bit stream output by the video encoding device based on the difference in the data order of the encoded bit stream depending on the decoding format on the encoding side and the decoding side. In other words, a transcoder that converts the encoded bit stream into a format that can be decoded on the decoding side and supplies the encoded bit stream after conversion to the video decoding device is provided. There is an effect that required processing and circuit scale, and required time can be reduced.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration example of a video encoding device according to a first embodiment of the present invention.

FIG. 2 is a block diagram illustrating a detailed configuration example of a motion compensation predictor in the video encoding device of FIG.

FIG. 3 is a block diagram illustrating a detailed configuration example of a quantizer in FIG. 1;

FIG. 4 is a block diagram illustrating a configuration example of a moving image processing system according to a second embodiment of the present invention;

FIG. 5 is a block diagram illustrating a configuration example of a conventional moving image encoding device that encodes a moving image in, for example, the MPEG-4 format.

6 is a block diagram illustrating a detailed configuration example of a motion compensation predictor in FIG.

FIG. 7 is a block diagram illustrating a detailed configuration example of the quantizer of FIG. 5;

FIG. 8 is a block diagram illustrating an example of a moving image processing system that processes moving images in different encoding formats on the encoding side and the decoding side.

[Explanation of symbols]

Reference Signs List 1 control unit (encoding unit), 2 DCT unit (encoding unit), 3 quantizer (encoding unit), 4 variable length encoder (encoding unit), 5 inverse quantizer (encoding unit) , 6
Inverse DCT unit (encoding unit), 7 adder (encoding unit), 8 frame memory (encoding unit), 9 motion compensation predictor (encoding unit), 10 difference unit (encoding unit), 11, 12 Selector (encoding unit), 13 preprocessing unit (encoding unit), 51, 71 control unit (encoding format identifying unit), 81 transmission side encoder (moving image encoding device), 82 transcoder, 83 reception side Decoder (moving image decoding device).

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Etsuhisa Yamada 2-3-2 Marunouchi, Chiyoda-ku, Tokyo Inside Mitsubishi Electric Corporation (72) Inventor Kotaro Asai 2-3-2 Marunouchi, Chiyoda-ku, Tokyo Inside Rishi Electric Co., Ltd. (72) Inventor Yuri Hasegawa 2-3-2 Marunouchi, Chiyoda-ku, Tokyo Sanritsu Electric Co., Ltd. (72) Shunichi Sekiguchi 2-3-2 Marunouchi, Chiyoda-ku, Tokyo Mitsubishi Within Electric Co., Ltd. (72) Inventor Fuminobu Ogawa 2-3-2 Marunouchi, Chiyoda-ku, Tokyo F-term (reference) 5C059 KK06 KK11 MA00 MA23 MC11 MC38 ME02 NN03 NN15 NN28 PP04 SS07 TA45 TA62 TB07 TC18 TC27 TC45 UA02 UA05 UA33 5J064 AA03 BA01 BB03 BB13 BC01 BC14 BC16 BD03

Claims (11)

[Claims] 1. A moving image encoding apparatus for encoding a moving image on the encoding side of a moving image processing system for processing moving images in different encoding formats on an encoding side and a decoding side, wherein: Code format specifying means for specifying the format of a code that can be decoded on the side, and a part of the function of the coding format on the coding side according to the format of the code that can be decoded specified by the code format specifying device And a coding unit for coding the moving image into a code that can be decoded on the decoding side. 2. The moving picture coding apparatus according to claim 1, wherein said coding means has a function of coding the moving picture in one of MPEG formats. 3. The encoding means stops the function of the inter4V motion compensation when it is determined that the format of the code that can be decoded on the decoding side does not depend on the inter4V motion compensation. 3. The moving picture coding apparatus according to claim 1, wherein the moving picture is coded. 4. When the encoding unit determines that the format of a code that can be decoded on the decoding side is to be decoded based on a motion vector only within a predetermined pixel range, the predetermined pixel range is determined. 3. The moving picture coding apparatus according to claim 1, wherein the function of searching for a motion vector is stopped outside and the moving picture is coded. 5. When the encoding unit determines that the code format that can be decoded on the decoding side is to be motion compensated and decoded based on a motion vector in integer pixel units, the encoding unit performs decoding in half-integer pixel units. 3. A moving picture is coded by stopping the function of detecting a motion vector.
The moving picture encoding device according to the above. 6. The encoding means, when it is determined that the format of a code that can be decoded on the decoding side is to be decoded without using motion compensation by a motion vector from outside the frame, 3. The moving picture coding apparatus according to claim 1, wherein the function of detecting a motion vector is stopped to code the moving picture. 7. The encoding means specifies that a code format that can be decoded on the decoding side is to be decoded by motion compensation for each integer pixel, and that the encoding format on the encoding side is half integer pixel. In the case of encoding with motion compensation, the function of searching for a motion vector in one of the odd-numbered pixel and the even-numbered pixel is stopped, and a motion vector is searched for in the other pixel. The moving picture coding apparatus according to claim 1 or 2, wherein the moving picture is coded. 8. The encoding unit according to claim 1, wherein the code format that can be decoded on the decoding side is decoded by performing motion compensation for each half-integer pixel, and interpolating the half-integer pixel value from the integer pixel value. If it is specified that the fractional part of the half-integer pixel value is to be rounded up, the other motion compensation prediction function is stopped, and when interpolating the half-integer pixel value from the integer pixel value, the half-integer pixel value 3. The moving picture coding apparatus according to claim 1, wherein the moving picture is coded by performing motion compensation prediction in which a decimal part is rounded up. 9. The coding means, when it is specified that the code form decodable on the decoding side is a code that cannot be inversely quantized by the intra AC coefficient adaptive prediction quantization, 3. The moving picture coding apparatus according to claim 1, wherein the function of quantization is stopped, quantization is performed, and the moving picture is coded. 10. A moving picture coding method for coding a moving picture on the coding side of a moving picture processing system for processing moving pictures in different coding formats on the coding side and the decoding side, Identifying the format of the code that can be decoded on the side, and stopping a part of the function of the encoding format on the encoding side according to the identified format of the decodable code, Encoding the moving image into a decodable code. 11. A moving image processing system for processing moving images of different encoding formats on an encoding side and a decoding side, comprising: a code format identification unit for identifying a format of a code that can be decoded on the decoding side; A part of the function of the encoding format on the encoding side is stopped according to the format of the decodable code specified by the code format identification unit, and the moving image is converted to a code decodable on the decoding side. A moving image encoding device including an encoding unit that outputs an encoded bit stream, a moving image decoding device that decodes an encoded bit stream in a format that can be decoded on the decoding side, and the code Based on the difference in the data order of the encoded bit stream according to the decoding format of the decoding side and the decoding side,
The data of the coded bit stream output by the video coding apparatus is rearranged and converted into a coded bit stream in a format that can be decoded on the decoding side, and the coded bit stream after the conversion is decoded by the video decoding. A moving image processing system comprising: