JP2002016929A - Encoder and encoding method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an encoder which can prevent overflow of a buffer and can realize encoding of VOP within the predetermined period even if the processing time required for the encoding each macro-block is not constant. SOLUTION: The maximum amount of code Tmax is set for every VOP, and an output of an encoding means and an output of a fixed code output means are switched to provide an output in order to prevent that amount of code Sc of VOP does not exceed the maximum amount of code Tmax.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an encoding apparatus and an encoding method for encoding a video signal in real time, for example, for a portable telephone, a TV telephone system, and the like.

[0002]

2. Description of the Related Art FIG. 15 shows, for example, "All about MPEG-4" (Industry Research Council) p. 39 to p. FIG. 16 is a block diagram of the conventional encoding device shown in FIG. 40, FIG. 16 is an explanatory diagram showing an input signal of the conventional encoding device, FIG. 17 is an explanatory diagram showing a configuration of a bit stream, and FIG. FIG. 4 is an explanatory diagram showing a position (arrangement) of a video packet on a screen (displayed state).

[0005] In FIG. 15, reference numeral 1 denotes a subtractor having an external input signal (a luminance signal and a chrominance signal in the example shown in FIG. 1) as a first input, and the output of the subtracter 1 is a DC signal.
T (Discrete Cosine Transform. Discrete Cosine)
e Transform) means 2 and a quantizer 3, which are input to a DC / AC predictor 4 and an inverse quantizer 6 for predicting a quantized value of a direct current (DC) component and an alternating current (AC) component. The output of the DC / AC predictor 4 is provided to a first input of a variable length coding unit 5, and the variable length coding unit 5 outputs a bit stream.

On the other hand, the output of the inverse quantizer 6 to which the output of the quantizer 3 is input is passed through an inverse DCT means 7 to an adder 8.
Is given to the first input. The output of the adder 8 is a memory 9
, And the output of the memory 9 is provided to a first input of the predicted image creating means 10 and a first input of the motion detecting means 11.

An external input signal is supplied to a second input of the motion detecting means 11, and an output of the motion detecting means 11 is supplied to a second input of the predicted image creating means 10 and a motion vector predictor 12.

[0006] The output of the motion vector predictor 12 is provided to a second input of the variable length coding means 5. Further, the output of the predicted image creation means 10 is the second input of the subtractor 1 and the adder 8
Is given to the second input.

Next, the operation will be described. First, the video signal is divided into macroblocks, which are basic processing units, as shown in FIG. 16, and is input as an external input signal (the external input signal here is basically input as a macroblock, Even if a macroblock is input, a means for generating a macroblock may be provided at a preceding stage to convert the macroblock into a macroblock.)

When the input video signal is 4: 2: 0,
The 16 pixels × 16 lines of the luminance signal (Y) have the same size on the screen as the 8 pixels × 8 lines of the two color difference signals (Cb, Cr). Accordingly, one macroblock is composed of six blocks of 8 pixels × 8 lines (4 blocks for luminance signals and 2 blocks for color difference signals, that is, 6 blocks).

[0009] Here, it is assumed that a video object plane (VOP; unit image) input as an external input has a rectangular shape and is the same as a frame.

Each block is a discrete cosine transform (DCT)
And then quantized by the quantization means 3. The quantized DCT coefficient is converted to DC,
After predicting the coefficient of each AC component, variable-length coding is performed together with additional information such as a quantization parameter.

This is intra coding (sometimes referred to as intra-frame coding). A VOP to which intra coding is applied to all macroblocks is an I-VOP
(Intra-VOP).

On the other hand, the quantized DCT coefficient is subjected to inverse quantization by the inverse quantization means 6 and inverse DCT by the inverse DCT means 7 to be decoded, and the decoded image is stored in the memory 9. The decoded image stored in the memory 9 is used when performing inter-coding (sometimes referred to as inter-frame coding).

In the case of inter coding, the motion detecting means 1
In step 1, a motion vector indicating the motion of a macroblock input as an external input signal is detected. The motion vector indicates a position in the decoded image stored in the memory 9 where the error from the input macroblock is minimized.

The predicted image creating means 10 includes a motion detecting means 11
A predicted image is created based on the motion vector detected in.

Subsequently, a difference signal between the input macroblock and the predicted image created by the predicted image creation means 10 is obtained, DCT is applied to the difference signal by the DCT means 2, and quantization is performed by the quantization means 3. Perform the conversion.

The quantized DCT coefficients are subjected to variable-length coding together with predictive-coded motion vectors and additional information such as quantization parameters. Also, the quantized D
The CT coefficient is inversely quantized by the inverse quantization means 6 and the inverse DC
After performing the inverse DCT in the T means 7, the result is added to the predicted image by the adder 8 and stored in the memory 9.

The inter coding includes unidirectional prediction in which prediction is performed only from a VOP that is temporally earlier in the display order of images, and bidirectional prediction in which prediction is performed from both a VOP that is temporally earlier and a VOP that is later. . VOP encoded by unidirectional prediction is P
-VOP (Predictive VOP), and a VOP encoded by bidirectional prediction is called a B-VOP (Bidi
reactionally Predictive VO
P).

Next, the configuration of the bit stream output from the variable length coding means 5 will be described with reference to FIG. The bit stream of one VOP is shown in FIG.
And one or more video packets.

Here, one video packet is composed of encoded data of one or more macroblocks,
For the first video packet of a VOP, the VO
A P header is added, and finally, a stuff bit for byte alignment is added (FIG. 17B).

In the case of the second and subsequent video packets, Resyn for detecting the head of the video packet is provided at the head.
A cMarker and a video packet header are added, and finally a stuff bit is added (FIG. 17C).

The stuff bit is added to the end (break) of the video packet in units of 1 to 8 bits in order to adjust byte alignment at the end of the video packet. The meaning is distinguished.

Also, an arbitrary number of stuffings can be included in a video packet as shown in FIG. For example, in the case of MPEG4 Video, this stuffing is called a stuffing macroblock, and can be included in an arbitrary video packet in the same manner as a macroblock. This stuffing is discarded (not substantially used) on the decoding device side.

The stuffing here is used as a word such as 9 bits or 10 bits for increasing the code amount, and is byte alignment (for example, adjusting the end of a video packet).
The stuff bit is used irrespective of the stuff bit and is inserted between macro blocks.

The number of macroblocks that can be included in one video packet is arbitrary. However, in consideration of error propagation, it is generally preferable that the configuration is such that the code amount of each video packet is substantially constant. ing. When the code amount of the video packet is almost constant as described above,
The area occupied by one video packet in one VOP is not constant as shown in FIG.

Next, referring to FIG.
The details of the operation of the predictor 4 will be described (here, the Y component of the macroblock will be described). As mentioned above,
The DC / AC predictor 4 predicts the DC and AC component coefficients of the quantized DCT coefficients output from the quantizer 3 in the case of intra coding. In the case of the inter coding, the DC component and the AC component are not predicted, and the quantizer 3
And outputs the quantized DCT coefficient output from the variable length encoding means 5 as it is. In this case, DC / AC prediction is separately performed on the luminance signal Y and the color difference signal C.

In the following, prediction of DC and AC components in the case of intra coding will be described. The quantized DCT coefficients of the block currently being encoded are represented by Fx (i,
j) (0 ≦ i ≦ 7, 0 ≦ j ≦ 7), and quantized DCT coefficients of the block on the left of this block are Fa (i, j).
(0 ≦ i ≦ 7, 0 ≦ j ≦ 7) and the quantized DCT coefficients of the upper adjacent block are represented by Fc (i, j) (0 ≦ i ≦ 7, 0 ≦
j ≦ 7) and the quantized DCT coefficient of the upper left block is Fb (i, j) (0 ≦ i ≦ 7, 0 ≦ j ≦ 7),
First, the DC of the quantized DCT coefficient of the upper left block
The component Fb (0,0) and the quantized D of the block on the left
The prediction direction is determined from the DC component Fa (0,0) of the CT coefficient and the DC component Fc (0,0) of the quantized DCT coefficient of the upper adjacent block.

For example, if the quantization step width of the DC component of the block on the left is Qda, the quantization step width of the DC component of the upper left block is Qdb, and the quantization step width of the DC component of the upper left block is Qdc. , Fa (0,0) = Fa (0,0) × Qda fb (0,0) = Fb (0,0) × Qdb fc (0,0) = Fc (0,0) × Qdc DC components fa (0,0), fb
(0,0) and fc (0,0), respectively, and | fa (0,0) -fb (0,0) | <| fb (0,
If the relationship 0) −fc (0,0) | is established, it is considered that the correlation in the vertical direction is strong, so the DC component fc of the upper adjacent block after inverse quantization is obtained.
Prediction is performed from (0,0), and if the above relationship does not hold, it is considered that the correlation in the left-right direction is strong. Therefore, prediction is performed from the DC component fa (0,0) of the block on the left after dequantization. Do.

When predicting the DC component from the upper adjacent block, Px (0,0) = Fx (0,0) -fc (0,0) / Q
dx, and when predicting the DC component from the block on the left, Px (0,0) = Fx (0,0) -fa (0,0) / Q
The predicted DC component Px (0,0) is obtained as dx. Here, Qdx is the quantization step width of the DC component of the current block, and the above division is calculated by, for example, rounding off.

Subsequently, the AC component is predicted using the above-described DC component prediction direction. That is, assuming that the quantization parameter of the block on the left is Qpa, the quantization parameter of the block on the upper side is Qpc, and the quantization parameter of the current block is Qpx, when the DC component is predicted from the block on the upper side, , AC component prediction, Px (i, 0) = Fx (i, 0) − (Fc (i, 0) ×
Qpc) / Qpx (i = 1,..., 7).

When the DC component is predicted from the block on the left, the AC component is predicted by Px (0, j) = Fx (0, j)-(Fa (0, j) ×
Qpa) / Qpx (j = 1,..., 7) to determine the predicted AC component Px (i, 0) or Px (0, j). However, the above division is
For example, it is assumed that the calculation is performed by rounding.

After independently predicting the AC component for the six blocks constituting one macroblock,
Whether or not to predict the AC component is determined for each macroblock as described below (determined for each macroblock depending on which block the prediction is performed on).

Here, as an index indicating whether it is better to keep the original video signal as it is (no prediction of the AC component is performed) or to perform the prediction, the AC of the block is used as an index.
The prediction judgment index SB is obtained as follows. For example, for each of the six blocks constituting one macroblock, when that block (the block for which the AC prediction determination index SB is to be obtained) is predicted from the block on the upper side,

[0033]

(Equation 1)

When the AC prediction judgment index SB is obtained from the block, and the block is predicted from the block on the left side,

[0035]

(Equation 2)

The AC prediction decision index SB is obtained by the above operation, and the sum SBS of the AC prediction decision indexes SB of the six blocks constituting one macroblock (A obtained for each block)
C sum), if SBS ≧ 0, the AC component is predicted. Otherwise, A
Do not predict the C component.

When predicting the AC component, ac_
When pred_flag = 1, and when the AC component is not predicted, ac_pred_flag = 0,
After the pred_flag is added as additional information for each macroblock, each macroblock is
By encoding.

For a block belonging to a macroblock of ac_pred_flag = 1, if that block is predicted from the upper neighboring block,

[0039]

(Equation 3)

As a result, when Ox (i, j) is obtained and the block is predicted from the block on the left,

[0041]

(Equation 4)

Thus, Ox (i, j) is obtained. Also,
For a block belonging to a macroblock of ac_pred_flag = 0,

[0043]

(Equation 5)

As a result, Ox (i, j) is obtained,
(I, j) is output to the variable length encoding means 5 as an output of the DC / AC predictor 4.

In the above prediction, the current block is located at the left end of the unit image (when the unit image is one screen,
In the case of the block at the left end of this one screen), the block adjacent to the left of the current block and the upper left block do not exist, and thus the DC component fa after inverse quantization used in the above prediction
Let the values of (0,0) and fb (0,0) be a predetermined constant β. Further, in this case, AC components Fa (i, j) and Fb (i, j) ((i, j) ≠ (0,
0)) is set to 0.

Here, the predetermined constant β is, for example, DC
The DCT coefficient output from the T means 2 is set to an intermediate value in the range of the value of the DC component. That is, when the DC component output from the DCT means 2 can take a value from 0 to 2047 in 11 bits, β = 1024.

Similarly, in the above prediction, if the current block is the block at the upper end of the unit image (the upper end of this one screen when the unit image is one screen), the block immediately above the current block And the upper left block does not exist, the DC component f after inverse quantization used in the above prediction
The values of c (0,0) and fb (0,0) are converted to the above constant β
And AC components Fc (i, j), Fb (i, j)
((I, j) ≠ (0, 0)) is set to 0.

Further, in the above-mentioned prediction, when the block on the left of the current block belongs to a video packet different from the current block, the DC component fa (0,0) after dequantization used in the above-mentioned prediction is converted to the above-mentioned block. And the AC component Fa (i, j) ((i, j) ≠ (0,0)) is set to 0.

Similarly, in the above-described prediction, if the block immediately above the current block belongs to a video packet different from the current block, the DC component fc (0,0) after the inverse quantization used in the prediction is used. The AC component Fc (i, j) ((i, j) ≠ (0,0)) is set to 0 with the above constant β.

In the above-mentioned prediction, when the upper left block of the current block belongs to a video packet different from the current block, the DC component fb (0,0) after dequantization used in the prediction is converted to the above-mentioned block. A constant β is set, and an AC component Fb (i, j) ((i, j) ≠ (0, 0)) is set to 0.

As described above, in the DC / AC predictor 4, the DC / AC predictor 4
By not referring to the coefficients of the component and AC component,
Even when an error is mixed in the transmitted bit stream, the error propagation based on the DC / AC prediction is configured to be contained in the video packet.

[0052]

In the conventional encoding apparatus as described above, processing for avoiding overflow of the transmission buffer and underflow of the VBV buffer which is a virtual buffer on the receiving side is sufficiently considered. It was not.

Normally, the quantization parameter used in the quantizer 3 is adjusted to increase or decrease the code amount. However, the quantization parameter is maximized (the code amount generated by performing the coarsest quantization is suppressed). However, no consideration has been given to the processing in the case where the transmission buffer overflows.

When the rate of the input VOP is F (1
/ Sec), it is required to encode all macroblocks constituting one VOP in 1 / F (sec) or in a shorter time.

However, for example, the motion detecting means 11
Is configured to adaptively change the search range of the motion vector according to the motion of the object in the VOP,
The time required for the motion detection means 11 to detect the motion vector of each macroblock changes for each macroblock, and therefore, the processing time of one VOP is not constant.
In such a case, control for encoding all macroblocks constituting one VOP within a predetermined time has not been considered in the related art.

The present invention has been made to solve the above problems, and proposes an encoding apparatus and an encoding method capable of effectively avoiding overflow of a transmission buffer and underflow of a VBV buffer. is there.

The present invention also provides an encoding apparatus and an encoding method capable of terminating encoding of one VOP within a predetermined time even when the time required for encoding one macroblock is not constant. is there.

[0058]

An encoding apparatus according to the present invention encodes an external input signal in units of macroblocks and outputs a code corresponding to the external input signal, and a code for a unit image. Fixed code output means for outputting a fixed code predetermined for each encoding type; storage means for accumulating the code output from the encoding means or the fixed code output from the fixed code output means; Code amount control means for controlling either the code or the fixed code to be stored in the storage means in accordance with the code amount of the code coded by the conversion means.

Further, the code amount control means in the encoding apparatus according to the present invention obtains the maximum code amount that can be generated by encoding the unit image for each unit image, and based on the maximum code amount, For each macro block constituting an image, control is performed such that either the code output from the encoding means or the fixed code output from the fixed code output means is stored in the storage means.

Further, the code amount control means in the encoding apparatus according to the present invention is characterized in that the maximum code amount Tmax is obtained based on the following equation. Tmax ≦ Bs−B where Tmax is the maximum code amount, Bs is the capacity of the storage unit (capacity allowed in the storage unit), and B is the occupation amount in the storage unit.

Further, the code amount control means in the coding apparatus according to the present invention is characterized in that the maximum code amount Tmax is obtained based on the following equation. Tmax ≦ vbv_bits + Rp where Rp = R / F, where Tmax is the maximum code amount, Rp is the number of bits read from the storage unit in the unit image, R is the bit rate read from the storage unit, and F is the unit image to be coded. , Vbv_bits is the occupation amount of the VBV buffer in the immediately preceding unit image.

Further, the code amount control means in the encoding apparatus according to the present invention is characterized in that the maximum code amount Tmax is obtained based on the following equation or a value that gives a result equivalent thereto. Tmax ≦ min (vbv_bits + Rp, Bs−
B) where Rp = R / F, where Tmax is the maximum code amount, Rp is the number of bits read from the storage unit in the unit image, R is the bit rate read from the storage unit, and F is the unit image to be coded. The rate, vbv_bits are the occupancy of the VBV buffer in the immediately preceding unit image, Bs is the capacity of the storage means (capacity allowed in the storage means), and B is the occupancy of the storage means.

Further, the encoding apparatus according to the present invention is characterized in that the bit rate R read from the storage means is variable.

Further, the code amount control means in the coding apparatus according to the present invention includes a code amount mb_bit of a code outputted from the coding means for the current macroblock, and a code amount of one of the current macroblocks constituting the unit image. The code amount Sc up to the previous macroblock, the maximum code amount Tmax of the unit image, the number of macroblocks M to be coded following the current macroblock constituting the unit image, fixed for the unit image The relationship between the code length L of the fixed code output from the code output means and the code amount α of an additional code generated in video packet units constituting the unit image is Sc + mb_bit + M × L + α> Tmax (where α ≧ 0) ), The fixed code output from the fixed code output means is stored in the storage means for the current macroblock. It is characterized by controlling.

Further, the code amount control means in the encoding apparatus according to the present invention comprises: a code amount mb_bit of a code output from the encoding means for the current macroblock; The code amount Sc up to the previous macroblock, the maximum code amount Tmax of the unit image, the number M of macroblocks to be coded following the current macroblock constituting the unit image,
The relationship between the code length L of the fixed code output from the fixed code output means for the unit image and the code amount α of the additional code generated in video packet units constituting the unit image is Sc + mb_bit + M × L + α> When Tmax (where α ≧ 0), control is performed such that the fixed code output from the fixed code output means is stored in storage means for the current macroblock,
For the M macroblocks to be coded following the current macroblock constituting the unit image, control is performed so that the fixed code output from the fixed code output unit is stored in the storage unit. It is characterized by the following.

Further, the code amount control means in the coding apparatus according to the present invention is arranged such that when the relation of Sc + mb_bit + M × L + α> Tmax is satisfied and the coding type of the unit image is intra, the current macroblock From a new video packet.

Further, the fixed code output means in the coding apparatus according to the present invention is characterized in that it outputs a code corresponding to a macroblock having the minimum code amount in the coding type of a unit image.

An encoding device according to the present invention encodes an external input signal in units of macroblocks, and outputs a code corresponding to the external input signal, and a predetermined coding unit for each coding type of a unit image. Fixed code output means for outputting the fixed code, storage means for storing the code output from the coding means or the fixed code output from the fixed code output means, and a first macro constituting a unit image And a code amount control means for controlling either the code or the fixed code to be stored in the storage means based on an elapsed time from the input of the block.

An encoding apparatus according to the present invention encodes an external input signal in units of macroblocks, and outputs a code corresponding to the external input signal, and a predetermined unit for each encoding type of a unit image. Fixed code output means for outputting the fixed code, storage means for storing the code output from the coding means or the fixed code output from the fixed code output means, Code amount control for storing either the code or the fixed code in the storage means based on the code amount of the code and the elapsed time since the leading macroblock constituting the unit image is input. Means.

The encoding method according to the present invention encodes an external input signal in units of macroblocks, and outputs a code corresponding to the external input signal, and determines in advance for each encoding type of a unit image. A fixed code output step of providing the fixed code, a storage step of storing the code obtained in the coding step or the fixed code given from the fixed code output step, and the code coded by the coding step. And a code amount control step of controlling either the code or the fixed code to be stored in the storage step in accordance with the code amount.

In the code amount control step in the encoding method according to the present invention, the maximum code amount that can be generated by encoding the unit image is obtained for each unit image, and the unit code is calculated based on the maximum code amount. It is characterized in that, for each macroblock constituting an image, control is performed such that either the code given from the encoding step or the fixed code given from the fixed code output step is stored in the storage step.

The code amount control step in the encoding method according to the present invention is characterized in that the maximum code amount Tmax is obtained based on the following equation. Tmax ≦ Bs−B where Tmax is the maximum code amount, Bs is the capacity allowed in the accumulation step, and B is the occupation amount in the accumulation step.

The code amount control step in the encoding method according to the present invention is characterized in that the maximum code amount Tmax is obtained based on the following equation. Tmax ≦ vbv_bits + Rp where Rp = R / F, where Tmax is the maximum code amount, Rp is the number of bits read in the storage step in the unit image, R is the bit rate of the read in the storage step, and F is the unit image to be coded. , Vbv_bits is the occupation amount of the VBV buffer in the immediately preceding unit image.

Further, the code amount control step in the encoding method according to the present invention is characterized in that the maximum code amount Tmax is obtained based on the following equation or a value having a result equivalent thereto. Tmax ≦ min (vbv_bits + Rp, Bs−
B) where Rp = R / F, where Tmax is the maximum code amount, Rp is the number of bits read in the storage process in the unit image, R is the bit rate of read in the storage process, and F is the rate of the unit image to be coded. , Vbv_bits are the occupancy of the VBV buffer in the immediately preceding unit image, Bs is the capacity allowed in the accumulation step, and B is the occupancy in the accumulation step.

Further, in the encoding method according to the present invention, the read bit rate R in the storage step is variable.

The code amount control step in the coding method according to the present invention includes the code amount mb_bit of the code given from the coding step for the current macroblock, and one of the current macroblocks constituting the unit image. The code amount Sc up to the previous macroblock, the maximum code amount Tmax of the unit image, the number M of macroblocks to be coded following the current macroblock constituting the unit image, a fixed code for the unit image The relationship between the code length L of the code output by the output step and the code amount α of the additional code generated in video packet units constituting the unit image is Sc + mb_bit + M × L + α> Tmax (where α ≧ 0). In some cases, for the current macroblock, control is performed so that the fixed code given from the fixed code output step is accumulated in the accumulation step. It is characterized by that.

Further, the code amount control step in the encoding method according to the present invention includes the code amount mb_bit of the code given from the encoding step for the current macroblock, one of the current macroblocks constituting the unit image. The code amount Sc up to the previous macroblock, the maximum code amount Tmax of the unit image, the number M of macroblocks to be coded following the current macroblock constituting the unit image, The relationship between the code length L of the fixed code given from the fixed code output step and the code amount α of the additional code generated in video packet units constituting the unit image is Sc + mb_bit + M × L + α> Tmax (where α ≧ 0), the fixed code provided from the fixed code output step is stored for the current macroblock. Control to accumulate in the step, and also for the M macroblocks to be encoded following the current macroblock constituting the unit image, the fixed code given from the fixed code output step It is characterized in that accumulation is controlled in the accumulation step.

In the code amount control step in the coding method according to the present invention, when the relationship of Sc + mb_bit + M × L + α> Tmax is satisfied and the coding type of the unit image is intra, the current macro block From a new video packet.

Further, the fixed code output step in the coding method according to the present invention is characterized in that a code corresponding to a macroblock having a minimum code amount in a unit image coding type is output.

The encoding method according to the present invention encodes an external input signal in units of macroblocks and outputs a code corresponding to the external input signal, and determines in advance for each encoding type of a unit image. A fixed code output step of providing the fixed code, an accumulation step of accumulating the code given from the encoding step or the fixed code given from the fixed code output step, and a top macroblock constituting the unit image. A code amount control step of controlling either the code or the fixed code to be stored in the storage step based on an elapsed time from the input.

The encoding method according to the present invention encodes an external input signal in units of macroblocks and outputs a code corresponding to the external input signal, and determines in advance for each encoding type of a unit image. A fixed code output step of outputting the fixed code, a storage step of storing the code given from the encoding step or the fixed code given from the fixed code output step, and the encoding step performed by the encoding step. A code amount control step of controlling either the code or the fixed code to be stored in the storage step based on a code amount of a code and an elapsed time since a leading macroblock constituting the unit image is input; And characterized in that:

[0082]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the present invention will be specifically described with reference to the drawings showing the embodiments. Embodiment 1 FIG. FIG. 1 shows an encoding apparatus according to Embodiment 1 of the present invention. In the figure, reference numeral 1 denotes a subtractor having an external input signal as a first input. The output of the subtracter 1 is passed through a DCT means 2 and a quantizer 3 to a DC / AC predictor 4 and an inverse quantizer 6. Is entered. The output of the DC / AC predictor 4 is provided to a first input of the variable length coding means 5a.

On the other hand, the output of the inverse quantizer 6 is supplied to the first input of the adder 8 through the inverse DCT means 7. The output of the adder 8 is provided to a first input of a memory 9, and the output of the memory 9 is provided to a first input of a predicted image creating means 10 and a first input of a motion detecting means 11.

An external input signal is supplied to a second input of the motion detecting means 11, and an output of the motion detecting means 11 is supplied to a second input of the predicted image creating means 10 and a motion vector predictor 12. The output of the predicted image creating means 10 is provided to a second input of the subtractor 1 and a second input of the adder 8.

The output of the motion vector predictor 12 is given to the second input of the variable length coding means 5a. In addition,
The encoding means includes the above-described subtractor 1 to which the external input signal is input, and the variable-length encoding means 5a to which the variable-length code corresponding to the external input signal is output (of course, here). The configuration shown in is only an example,
A known configuration that can perform encoding corresponding to an external input signal can be used. ).

The first output of the variable length coding means 5a is provided to a first input of a temporary buffer 101, and the second output of the variable length coding means 5a is provided to an input of a code amount control means 102.

The output of the fixed code output means 104 is given to the second input of the temporary buffer 101,
A first output of the code amount control means 102 is given to a third input of 01. The output of the temporary buffer 101 is provided to a first input of the transmission buffer 103 (here, the temporary buffer 101 or the transmission buffer 103 corresponds to a storage unit).

The second output of the code amount control means 102 is provided to a second input of the memory 9, and the third output of the code amount control means 102 is provided to a second input of the transmission buffer 103. The output of the transmission buffer 103 is output (transmitted) as a bit stream.

The output (transmitted) bit stream is received and decoded on the decoding device side.

Next, the operation will be described. First, the video signal is divided into macroblocks, which are basic processing units, as shown in FIG. 16, and input to the subtractor 1 and the motion detecting means 11 as input macroblocks. For example, when the input video signal is 4: 2: 0, the luminance signal (Y) 1
6 pixels x 16 lines are two color difference signals (Cb, Cr)
8 pixels x 8 lines of the same size on the screen
Six blocks of 8 pixels × 8 lines constitute one macro block.

When performing intra coding, each block is quantized after DCT. The quantized DCT coefficients are subjected to coefficient prediction in a DC / AC predictor 4 and then variable-length encoded by variable-length encoding means 5a together with additional information such as quantization parameters. Quantized D
The CT coefficient is inversely quantized by the inverse quantizer 6, and the inverse D
The inverse DCT is performed by the CT means 7 and decoded. The decoded image output from the inverse DCT means 7 is stored in the memory 9.

In the case of inter coding, the motion detecting means 1
In step 1, a motion vector indicating the motion of the input macroblock is detected. The motion vector indicates a position in the decoded image stored in the memory 9 where the error from the input macroblock is minimized.

The predicted image creating means 10 includes the motion detecting means 1
A predicted image is created based on the motion vector detected by step (1). Next, a difference between the input macroblock and the predicted image is obtained, DCT is performed on the difference signal by the DCT unit 2, and quantization is performed by the quantizer 3.

The quantized DC output from the quantizer 3
The T coefficient is variable-length coded by the variable-length coding unit 5a together with the coefficient predicted by the DC / AC predictor 4, the motion vector predictively coded by the motion vector predictor 12, and additional information such as a quantization parameter. You.
The quantized DCT coefficient is subjected to inverse quantization by the inverse quantizer 6 and inverse DCT by the inverse DCT means 7, and is then added to the predicted image output from the predicted image creation means 10, and is stored in the memory 9. Is stored.

Next, the operation of the variable length coding means 5a will be described in detail. The variable-length encoding unit 5a encodes the quantized DCT coefficient and additional information for each macroblock (encoding process), writes the encoded information into the temporary buffer 101, and outputs the code amount to the code amount control unit 102.

For example, in the case of an I-VOP of MPEG4,
First, the AC component of the DCT coefficient of each block output from the DC / AC predictor 4 is converted into one by a method such as zigzag scanning.
A dimension scan is performed, and run-length encoding is performed to encode a combination of the number of 0s and non-zero coefficients. The run-length coded AC component data of each block is written to the temporary buffer 101.

As shown in FIG. 2A, the coefficient data of each block is followed by a macroblock type (MTYPE) indicating intra / inter, etc., and whether or not each block of color difference has a non-zero AC coefficient. mcbpc that collectively encodes cbpc, dq indicating a quantization parameter
ant, the DC component of the DCT coefficient of each block, ac_pred_flag indicating whether AC prediction has been performed,
Cbpy indicating whether there is a non-zero AC coefficient in each block of Y is sequentially encoded and written into the temporary buffer 101.

Note that the total of these code amounts is output to the code amount control means 102 for each macroblock.

Similarly, in the case of the MPEG-4 P-VOP, the data encoded in the order shown in FIG.

The code amount control means 102 converts the macroblocks based on the code amount of each macroblock output from the variable length coding means 5a such that the length of each video packet becomes equal to or less than a predetermined value. Summary, temporary buffer 1
01 to the transmission buffer.

For example, in the case of MPEG4, FIG.
As shown in FIG. 3B, a header is added to the head of the video packet, and the video packet is rearranged in the specified bit stream and transferred.

The code amount control means 102 controls the transmission buffer 103 so as not to cause overflow or to prevent VBV (Video Buffering VeV) from occurring.
(a virtual buffer required for receiving a bit stream on the receiving side (the required capacity is described in, for example, a header in a transmission bit stream). Usually, a capacity for at least the I-VOP is set. The maximum code amount Tmax is set for each VOP so that the underflow does not occur, and the variable-length coding unit 5 sets the maximum code amount Tmax so that the code amount of the VOP does not exceed Tmax.
a or the output of the fixed code output unit 104
A fixed code to be written in the temporary buffer 101 is selected.

The maximum code amount Tmax here means that the transmission buffer 103 does not cause overflow,
This can be said to be the upper limit value of the code amount for preventing the VBV buffer from underflowing.

The details of the operation will be described below. Before starting coding of each VOP, the code amount control means 102 obtains the maximum code amount Tmax of the VOP. For example, the capacity of the transmission buffer 103 is Bs (bits), and the current remaining amount of the transmission buffer 103 (that is, the transmission buffer or the VB
Of data that is stored in a storage unit such as a V buffer and is not read out from the storage unit such as the transmission buffer or the VBV buffer (remains (is stored) in a storage unit such as a transmission buffer or a VBV buffer). This is the amount (remaining capacity), and such an amount of data is generally expressed as a buffer occupancy or an occupancy. Hereinafter, it is simply referred to as an occupancy. ) Is B (bits), it is sufficient to set the code amount of the VOP to Bs−B or less in order to prevent the transmission buffer 103 from overflowing. Therefore, the maximum code amount Tmax is set to Tma
x ≦ Bs−B may be set.

When managing the VBV buffer,
When the read bit rate of the transmission buffer 103 is R (bi
ts / sec), and the rate of the VOP to be encoded is F (1 /
sec), the transmission buffer 103 during one VOP period
Is Rp = R / F, and the number of bits received by the VBV buffer during one VOP period is also Rp.

Therefore, the occupancy of the VBV buffer in the decoding time of the VOP immediately before the current VOP is represented by vbv
_Bits (bits), in order to prevent the VBV buffer from underflowing, set the code amount of the VOP to v
bv_bits + Rp or less. That is, the maximum code amount Tmax may be set as Tmax ≦ vbv_bits + Rp.

Therefore, the code amount control means 102 sets each VO
Before starting encoding of P, the maximum code amount Tm of the VOP
ax is defined as Tmax = min (vbv_bits + Rp, Bs−
B) Set as (Min (a, b) indicates that the smaller of a and b is used as its value).

The occupation amount vbv_b of the VBV buffer
The “its” is for estimating the amount of occupancy on the receiving side. However, when an underflow occurs on the receiving side, it is not necessary to manage the underflow of the VBV buffer when taking measures such as delaying the decoding time. Thus V
When it is not necessary to manage the underflow of the BV buffer, it is sufficient to set Tmax = Bs-B.

Since the occupation amount B of the transmission buffer 103 changes over time, the value of the maximum code amount Tmax also changes over time, but this maximum code amount Tmax is calculated for each VOP. .

Next, the code amount control means 102 obtains the code amount of the current VOP for each macroblock, and, according to the flowcharts shown in FIGS. From the output code or the fixed code output from the fixed code output means 104, either the code to be stored in the temporary buffer 101 or the fixed code is selected (the code is controlled so as to be selected and stored). Step), either the code or the fixed code is accumulated in the temporary buffer 101 (accumulation step. The accumulation step may include transmission from the temporary buffer 101 to the transmission buffer 103).

FIG. 4 shows that the current VOP is P-VOP.
FIG. 5 shows a flowchart in the case where the encoding type is inter, and FIG. 5 shows a flowchart in the case where the current VOP is an I-VOP (the encoding type is intra).

(Regarding code amount control in case of P-VOP) First, the operation of the code amount control 102 in the case of P-VOP will be described. In the case of a P-VOP, the variable length coding means 5a
Is coefficient data of each block as shown in FIG.
not_coded, mcbpc, motion vector, cb
Although py and dquant are output for each macroblock, these codes do not always exist. For example, the coefficient data of each block is all 0, and the motion vector is (0, 0). If n
There is only one bit of ot_coded = 1. This is the code having the minimum code amount in the P-VOP macroblock.

Therefore, in the case of the P-VOP, the fixed code output means 104 uses not_coded = 1 as the fixed code.
Is output (fixed code output step.
A case where a fixed code is output also for an I-VOP as described later is also referred to as a fixed code output step.) That is, the fixed code output unit 104 outputs the fixed code of the macroblock having the minimum code amount for the current VOP coding type.

For example, if the coefficient data of each block is all 0
And when the motion vector is (0, 0), only one bit of not_coded = 1 exists. In the case of P-VOP, the code length L of the fixed code output from the fixed code output unit 104 is L = 1.

The code amount control means 102 obtains the code amount of the current VOP for each macroblock, and even if the fixed code output from the fixed code output means 104 is selected for all the remaining macroblocks, If the code amount of all the constituent macroblocks exceeds the maximum code amount Tmax of the VOP, the code of the current macroblock is replaced with the fixed code output from the fixed code output unit 104.

That is, assuming that the total number of macroblocks constituting a VOP is A and the macroblock number of the current macroblock is K (where 0 ≦ K ≦ A−1), the macro to be coded next The number M of blocks (the number M of remaining macroblocks) is expressed as M = A−K−1.

From the macroblock of macroblock number 0 constituting the current VOP, the macroblock number K-1
Let the code amount up to the macroblock of Sc be Sc, and let the code amount of the code output by the variable length coding means 5a for the current macroblock (macroblock number K) be mb_bit, the remaining M macroblocks On the other hand, when the fixed code (the code length is L) of the fixed code output unit 104 is selected, the code amount of the entire VOP is Sc + mb_bit + M × L + α. Here, α is a Resync Marker that can occur in a macroblock after the macroblock number K,
Video packet header, stuff bit, motion
_Marker is the code amount of an additional code generated in video packet units such as _marker (herein referred to as additional code amount), and α ≧ 0.

Therefore, if Sc + mb_bit + M × L + α> Tmax, the fixed code output from the fixed code output means 104 is written into the temporary buffer 101 for the current macroblock (macroblock number K). Controls to write the code output from the variable length coding means 5a into the temporary buffer 101 (FIG. 4).

As the value of the additional code amount α, for example, ResyncMarker in P-VOP, video packet header, stuff bit, motion_
Marker has a maximum code amount of Cp (bit),
The length of the predetermined video packet is VPlen (b
it), the amount of code generated in macroblocks subsequent to the current macroblock (macroblock number K) is
Since at least (M + 1) × L and (M + 1) × L / VPlen + 1 video packets can be generated, the additional code amount α
May be α = ((M + 1) × L / VPlen + 1) × Cp.

If the total number A of the macroblocks constituting the VOP is used, then M + 1 ≦ A. Therefore, to simplify the operation, the additional code amount α is given by α = (A × L / VPlen + 1) × Cp may be a value fixed to P-VOP.

For the P-VOP, the fixed code output from the fixed code output means 104 is stored in the temporary buffer 10.
1, the decoded image of the current macroblock stored in the memory 9 in order to forcibly treat it as not_coded (not coded) is stored in the previous VOP stored in the memory 9. The decoded image of the macroblock at the same position is replaced.

That is, in the memory 9, the immediately preceding VOP
Is copied to the decoded image area of the macroblock with the macroblock number K of the current VOP. In the case of P-VOP, the fixed code output from the fixed code output unit 104 is no.
Since t_coded = 1, the immediately preceding V
By copying the decoded image of the OP, a decoded image corresponding to the fixed code output from the fixed code output unit 104 is obtained.

(Regarding control of code amount in case of I-VOP) Next, the operation of the code amount control means 102 in the case of I-VOP will be described. In the case of an I-VOP, the variable-length coding unit 5a uses the AC component data, mcbpc, dquant, DC component, ac_data of each block as shown in FIG.
pred_flag and cbpy are output to each macroblock. However, not all of these codes are necessarily present. For example, when the values of cbpc and cbpy indicated by mcbpc are both 0, No coefficient data exists. Also, when the macroblock type indicated by mcbpc indicates that it does not have dquant, dquant does not exist.

Therefore, in the case of an I-VOP, the fixed code output means 104 determines that the DC and AC components of each block are all 0, dquant = 0, and ac_pred.
A code for a macroblock in which _flag = 0 is output as a fixed code. Note that MPEG2, M
In most existing encoding methods such as PEG4,
Such a code becomes the minimum code amount of the macroblock of the I-VOP.

As in the case of the P-VOP, the code amount control means 102 calculates the code amount of the current VOP for each macroblock, and outputs the fixed code output means 104 for the remaining macroblocks. Even if a code is selected, if the code amount of all the macroblocks constituting the VOP exceeds the maximum code amount Tmax of the VOP,
The code of the current macroblock is output to the fixed code output unit 10.
4 is replaced with the fixed code output from

Here, since it is required that the video packets forming the VOP are not broken down (all the macroblocks forming the video packet are included), it is necessary to generate encoded data corresponding to the remaining macroblocks. However, when the code amount of all the macroblocks constituting the VOP exceeds the maximum code amount Tmax,
Since the replacement with the fixed code is performed as described above, it is necessary to allow for the allowance for the fixed code.

That is, for the current macroblock (macroblock number K), the code amount of the code output from the variable length coding unit 5a is mb_bit, the fixed code output unit 1
Assuming that the code length of the fixed code output from the macroblock 04 is L and the code amount from the macroblock having the macroblock number 0 to the macroblock having the macroblock number K-1 constituting the current VOP is Sc, as shown in FIG. If Sc + mb_bit + M × L + α> Tmax (M = A−K−1), the fixed code output from the fixed code output unit 104 is written to the temporary buffer 101 for the current macroblock (macroblock number K). If not, control is performed so that the code output from the variable length coding means 5a is written into the temporary buffer 101.

Here, α is a Resync Marke which may occur in a macroblock after the macroblock number K.
r, video packet header, stuff bit, dc_m
The code amount (additional code amount) of a code generated in video packet units such as arcer, and α ≧ 0.

As the value of the additional code amount α, for example, ResyncMarker in I-VOP, video packet header, stuff bit, dc_mark
Assuming that the total code amount of er is Ci (bit) at the maximum and the length of a predetermined video packet is VPlen (bit), the code amount generated in a macroblock after the current macroblock (macroblock number K) Is at least (M + 1) × L, and (M + 1) × L / VPlen + 1 video packets can be generated. Therefore, α = ((M + 1) × L / VPlen + 1) × Ci.

If the total number A of the macroblocks constituting the VOP is used, then M + 1 ≦ A. To simplify the operation, α = (A × L / VPlen + 1) × Ci and I-VOP May be fixed.

In the case of an I-VOP, the fixed code output from the fixed code output means 104 for the current macroblock (macroblock number K) is stored in the temporary buffer 10.
1 and the output (fixed code) of the fixed code output means 104 is not selected for the immediately preceding macroblock (macroblock number K-1) (variable length coding means 5a), a new video packet is constructed from the current macroblock, as shown in FIG.

For an I-VOP, ac_pred_fl
Since the DC prediction is performed even when ag = 0, when the DC component stored in the temporary buffer 101 is 0, this corresponds to DC /
The predicted DC component Ox (0, 0,
0) is 0, and does not indicate that the DC component Fx (0,0) output from the quantizer 3 is 0.

Therefore, the fixed code output means 104 determines that the DC component and the AC component of each block are all 0 and d
When a fixed code for a macroblock in which quant = 0 and ac_pred_flag = 0 is output, an image obtained by decoding the fixed code is generally not constant (that is, the fixed code itself is fixed. Also, the value related to the image expression can take an arbitrary value instead of a fixed value.)

However, in the DC / AC predictor 4, D / D is not used between blocks belonging to different video packets.
Since the coefficient β of the C component is not referred to and the constant β which is an intermediate value in the range of the value of the DC component is used as the reference value,
When the fixed code output from the fixed code output unit 104 is selected to be stored in the temporary buffer 101, if the control is performed so that a new video packet is formed from the macroblock, the fixed code output from the fixed code output unit 104 is output. DC component f after inverse quantization of each block represented by the code
x (0,0) becomes fx (0,0) = β.

Therefore, in the case of an I-VOP, when the fixed code output from the fixed code output means 104 is decoded, an image in which all the pixels of the macroblock have a constant γ (a so-called image of the same color or the like on the entire screen). , A solid image). Here, the constant γ is an intermediate value in the range of pixel values of the input macroblock. For example, when an input macroblock can take a value of 0 to 255 with 8 bits,
γ = 128.

When the fixed code output from the fixed code output means 104 is selected to be stored in the temporary buffer 101, as described above, each of the blocks of the macro block (macro block number K) is dequantized. Becomes equal to the constant β, when selecting the output (fixed code) of the fixed code output means 104 in the next macroblock (macroblock number K + 1) of the macroblock, a new video packet is not formed. However, the DC component after the inverse quantization becomes a constant β, and the decoded image becomes an image (solid image) whose pixel values are all constant γ.

Therefore, as shown in FIG. 5, the fixed code output from the fixed code
In the case where the value is stored in 1 and the output (fixed code) of the fixed code output unit 104 is not selected for the previous macroblock, a new video packet is formed from the current macroblock. What is necessary is to control.

When the fixed code output from the fixed code output means 104 is stored in the temporary buffer 101,
The decoded image of the current macroblock stored in the memory 9 is replaced with an image whose pixel values are all constant γ. That is, the constant γ is written in the decoded image area of the current macroblock of the current VOP in the memory 9.

As described above, based on the flowcharts of FIGS. 4 and 5, of the codes output from the variable-length coding means 5a or the fixed codes output from the fixed code output means 104, the code stored in the temporary buffer 101 Is selected, the code amount of each VOP becomes the maximum code amount Tmax.
Can be controlled so as not to exceed.

Also, based on the flowchart of FIG.
By determining whether to construct a new video packet from the current macroblock, a new operation is performed on the decoded image corresponding to the fixed code output from the fixed code output unit 104 even in the case of I-VOP. Without writing to the memory 9, the code amount of the unit image can always be controlled to be equal to or less than the maximum code amount Tmax with a small amount of calculation.

Embodiment 2 In the first embodiment, the code amount control means 102 controls to select the output (fixed code) of the variable length coding means 5a or the fixed code output means 104 based on the flowcharts of FIGS. 4 and 5. The code amount control means 102 outputs the output (fixed code) of the variable length coding means 5a or the fixed code output means 104 based on the flowcharts shown in FIGS.
May be configured to be selected.

FIG. 6 shows that the current VOP is P-VOP.
FIG. 7 shows a flowchart when the current VOP is an I-VOP (the encoding type is intra), and FIG. 7 shows a flowchart when the current VOP is an I-VOP (the encoding type is intra).

(Regarding control of code amount in case of P-VOP) First, the case of P-VOP will be described with reference to FIG.
As in the first embodiment, the code amount control unit 102 sets the code amount of the code output from the variable length coding unit 5a for the current macroblock (macroblock number K) to mb_bit, the fixed code output unit L is the code length of the fixed code output from the macroblock 104 and the macroblock number K is from the macroblock of macroblock number 0 constituting the current VOP.
The code amount up to the -1 macroblock is Sc, and the current VO
Assuming that the number of macroblocks after macroblock K + 1 constituting P is M, Sc + mb_bit + M × L + α> Tmax (1) When the following expression is satisfied, the fixed code output unit 104 outputs the current macroblock (macroblock number K). Control is performed so that the fixed code to be output is written in the temporary buffer 101.

Here, α is a Resync Marke which may occur in a macroblock subsequent to the macroblock number K.
r, video packet header, stuff bit, moti
A code amount (additional code amount) of a code generated in video packet units such as on_marker, and α ≧ 0. As described in the first embodiment, α may be calculated for each macroblock, or may be a fixed value for each VOP coding type.

By the way, when the above equation (1) holds for the current macroblock, the code amount is accumulated, so that the equation (1) may hold for the next macroblock. Is very high.

For example, if the equation (1) holds for the macroblock number K, the code amount Sc ′ of the macroblocks from the macroblock number 0 to the macroblock number K becomes the macroblock number of the macroblock number K−1. By adding the code length L of the fixed code output from the fixed code output unit 104 to the code amount Sc up to, Sc ′ = Sc + L. Here, the code amount m of the code output by the variable length coding means 5a for the macroblock of the macroblock number K
If b_bit is equal to the code amount mb_bit 'of the code output from the variable-length coding means 5a for the macroblock of macroblock number K + 1, and the value of α is the same for both macroblocks, Sc '+ Mb_bit' + (M-1) × L + α = Sc +
mb_bit + M × L + α> Tmax, and the equation (1) holds for the macroblock number K + 1.

Therefore, if equation (1) holds for the macroblock number K, it is assumed that equation (1) holds for the macroblocks after the macroblock number K, and the calculation is omitted. it can.

That is, as shown in FIG. 6, first, a macro block (macro block number K-1) immediately before the current macro block (macro block number K) is set.
It is determined whether or not the output (fixed code) of the fixed code output unit 104 has been selected. If the output (fixed code) of the fixed code output unit 104 has been selected in the immediately preceding macroblock,
Fixed code output means 10 for the current macroblock
4 (fixed code) is stored in the temporary buffer 101.

On the other hand, if the output (fixed code) of the fixed code output means 104 is not selected in the immediately preceding macroblock, the above equation (1) is determined. If the equation (1) holds, the fixed code is used. The output (fixed code) of the output unit 104 is stored in the temporary buffer 101 if the output does not hold.

(Regarding code amount control in case of I-VOP) The same applies to the case of I-VOP. As shown in FIG. 7, first, the macroblock (macroblock number K) immediately before the current macroblock (macroblock number K) is used. It is determined whether or not the output (fixed code) of the fixed code output means 104 has been selected for the macro block number K-1), and the output (fixed code) of the fixed code output means 104 is determined for the immediately preceding macroblock. If selected, the output (fixed code) of the fixed code output unit 104 is stored in the temporary buffer 101 even for the current macroblock. In this case, as described in Embodiment 1,
There is no need to construct a new video packet from the current macroblock.

On the other hand, when the output (fixed code) of the fixed code output means 104 is not selected in the immediately preceding macroblock, the above equation (1) is judged. The output (fixed code) of the output unit 104 is stored in the temporary buffer 101 if the output does not hold. When the above equation (1) holds, a new video packet is formed from the current macroblock.

In the second embodiment, one VOP
When Expression (1) is satisfied for a certain macroblock in the set, the fixed code output means 10 is applied to all the macroblocks after that macroblock constituting the VOP.
4 (fixed code) is written to the temporary buffer 101, so in a macroblock to be coded following the macroblock, the subtracter 1, the DCT means 2, the quantizer 3, the DC / AC Predictor 4, variable length coding means 5a, inverse quantizer 6, inverse DCT means 7, adder 8,
The encoding means including the predicted image creation means 10, the motion detection means 11, and the motion vector prediction means 12 does not need to operate.

Therefore, when the expression (1) is satisfied for a certain macroblock in one VOP, the subtractor 1 and the DCT for the macroblock coded following the macroblock constituting the VOP. Means 2, quantizer 3, DC / AC predictor 4, variable length coding means 5a, inverse quantizer 6, inverse DCT means 7, adder 8, predicted image creating means 10, motion detecting means 11, and motion vector By stopping the operation of the prediction unit 12 (stopping the operation from the macroblock after the encoded macroblock to the end of the VOP), the amount of calculation and the power consumption can be reduced.

Embodiment 3 In the third embodiment,
The code amount control means 102 controls to select the output (fixed code) of the variable length coding means 5a or the fixed code output means 104 based on the flowchart shown in FIG.
FIG. 8 shows a flowchart in the case of P-VOP (coding type is inter).

(Regarding code amount control in case of P-VOP) For example, when the motion detecting means 11 is configured to adaptively change the search range of a motion vector according to the motion of an object in the VOP, The time required for the means 11 to detect the motion vector of each macroblock changes for each macroblock, so that the processing time of one VOP is not constant.

In such a case, in order to encode all the macroblocks constituting one VOP within a predetermined time, if the processing time is reduced, the subtracter 1 and the DCT
Means 2, quantizer 3, DC / AC predictor 4, variable length coding means 5a, inverse quantizer 6, inverse DCT means 7, adder 8, predicted image creating means 10, motion detecting means 11, and motion vector The output (fixed code) of the fixed code output unit 104 is stored in the temporary buffer 101 without performing the calculation of the prediction unit 12.

Accordingly, as shown in FIG. 8, the code amount control means 102 measures the elapsed time from the input of the first macroblock (macroblock number 0) constituting the current VOP, and Exceeds the predetermined processing time Tp, the fixed code output means 10
4 (fixed code) is stored in the temporary buffer 101. Otherwise, the output (fixed code) of the fixed code output unit 104 and the output (fixed code) of the variable length coding unit 5a are determined based on the above equation (1). Select output and temporary buffer 101
To memorize.

The predetermined processing time Tp in this case is set to a maximum of 1 VOP period (1 VO
(Because the processing for P must be performed in one VOP period), when other processing is included in this one VOP period, (1 VOP period-period required for other processing) is given to the processing time Tp. It is the maximum value.

(Regarding Code Control for I-VOP)
When the current VOP is an I-VOP, the code amount control means 102 stores the code stored in the temporary buffer 101 according to the flowchart of FIG. 5 or FIG. 7 as in the first or second embodiment. select.

Embodiment 4 In the fourth embodiment,
The code amount control means 102 controls to select the output (fixed code) of the variable length coding means 5a or the fixed code output means 104 based on the flowchart shown in FIG.
FIG. 9 shows a flowchart in the case of P-VOP (coding type is inter).

(Regarding code amount control in case of P-VOP) That is, the code amount control means 102
As described above, first, the fixed-code output unit 104 outputs the macroblock (macroblock number K-1) immediately before the current macroblock (macroblock number K).
It is determined whether or not the output (fixed code) has been selected, and if the output (fixed code) of the fixed code output unit 104 is selected in the immediately preceding macroblock, the fixed code output is also performed for the current macroblock. Output of means 104 (fixed code)
Is stored in the temporary buffer 101.

Next, as described in the third embodiment, the elapsed time from the input of the first macroblock (macroblock number 0) constituting the current VOP is measured.
If the elapsed time exceeds a predetermined processing time Tp, the output (fixed code) of the fixed code output means 104 is controlled to be stored in the temporary buffer 101; Then, the output (fixed code) of the fixed code output unit 104 and the output of the variable length coding unit 5a are selected and stored in the temporary buffer 101.

(Regarding code amount control in case of I-VOP) When the current VOP is an I-VOP, similarly to the first or second embodiment, according to the flowchart of FIG. 5 or FIG. , The code amount control means 102 selects a code to be stored in the temporary buffer 101.

Embodiment 5 In the first embodiment,
Although the block diagram in which the fixed code output means 104 exists independently is shown, for example, when each means is configured by software, the fixed code output means 10
4 and the variable length coding means 5a may be configured to share a ROM table for performing the respective operations.

That is, as described in the first embodiment, the fixed code output from fixed code output means 104 is I
Since the VOP and P-VOP have one pattern of the macroblock codes, the ROM table is shared by integrating the fixed code output unit 104 and the variable length coding unit 5a. be able to.

FIG. 10 shows an encoding apparatus according to Embodiment 5 of the present invention. In the figure, reference numeral 1 denotes a subtractor having an external input signal as a first input. The output of the subtracter 1 is passed through a DCT means 2 and a quantizer 3 to a DC / AC predictor 4 and an inverse quantizer 6. Is entered. The output of the DC / AC predictor 4 is provided to a first input of the variable length coding means 5b.

On the other hand, the output of the inverse quantizer 6 is supplied to the first input of the adder 8 through the inverse DCT means 7. The output of the adder 8 is provided to a first input of a memory 9, and the output of the memory 9 is provided to a first input of a predicted image creating means 10 and a first input of a motion detecting means 11.

An external input signal is supplied to a second input of the motion detecting means 11, and an output of the motion detecting means 11 is supplied to a second input of the predicted image creating means 10 and a motion vector predictor 12. The output of the predicted image creating means 10 is provided to a second input of the subtractor 1 and a second input of the adder 8.

The output of the motion vector predictor 12 is given to the second input of the variable length coding means 5b.

The first output of the variable length coding means 5b is provided to a first input of a temporary buffer 101, and the second output of the variable length coding means 5b is provided to an input of a code amount control means 102.

A first output of the code amount control means 102 is given to a second input of the temporary buffer 101. The output of the temporary buffer 101 is provided to a first input of the transmission buffer 103.

A second output of the code amount control means 102 is given to a second input of the memory 9 and a third input of the variable length coding means 5b, and a third output of the code amount control means 102 is a transmission buffer. 103 is provided to a second input. The output of the transmission buffer 103 is output (transmitted) as a bit stream
Is done.

The output (transmitted) bit stream is received and decoded on the decoding device side.

Next, the operation will be described. Embodiment 5
The operation of the variable-length coding unit 5b and the temporary buffer 101 is different from that of the first embodiment. The other parts are the same as in the first embodiment, and a description thereof will be omitted.

The variable length coding means 5b first codes the data of each macroblock as in the first embodiment,
The code is written into the temporary buffer 101 as shown in FIG. Here, the head address Ak of the code written for the current macroblock (macroblock number K) is stored. Also, the code amount m of the code generated at this time
b_bit is output to the code amount control means 102.

Next, the code amount control means 102 judges the expression (1), and if the expression (1) is satisfied, returns the write address of the temporary buffer 101 to Ak, and indicates that the fixed code is selected. The signal is output to the memory 9 and the variable length coding means 5b.

Upon receiving a signal indicating that a fixed code is to be selected, the variable-length coding means 5b outputs a fixed code predetermined for each VOP coding type to the temporary buffer 101. At this time, since the write address of the temporary buffer 101 has been returned to Ak, the code of the macroblock number K is overwritten with a fixed code. Therefore,
As in the data structure of the temporary buffer shown in FIG. 11B, a fixed code is written after the code of the macroblock number K-1.

When the memory 9 receives the flag indicating that the fixed code is selected, the constant γ is set in the decoded image area of the macroblock number K in the case of I-VOP as described in the first embodiment. In the case of writing or P-VOP, the VOP is stored in the decoded image area of the macroblock number K.
The decoded image of the macroblock having the macroblock number K of the immediately preceding VOP is copied.

With the above arrangement, in the fifth embodiment, means for coding each macroblock and means for outputting a fixed code prepared for each VOP coding type are provided to the variable length coding means 5b. , And the circuit can be reduced in size.

Embodiment 6 FIG. FIG. 12 shows an encoding apparatus according to Embodiment 6 of the present invention. In the figure, reference numeral 1 denotes a subtractor having an external input signal as a first input, and an output of the subtracter 1 passes through a DCT unit 2 and a quantizer 3 and is inversed to a first input of a variable length encoding unit 5c. Quantizer 6
Given to.

The output of the inverse quantizer 6 is supplied to the first input of the adder 8 through the inverse DCT means 7. Adder 8
Is provided to a first input of a memory 9, and an output of the memory 9 is provided to a first input of a predicted image creating means 10 and a first input of a motion detecting means 11.

An external input signal is supplied to a second input of the motion detecting means 11, and an output of the motion detecting means 11 is supplied to a second input of the predicted image creating means 10 and a motion vector predictor 12. The output of the predicted image creating means 10 is provided to a second input of the subtractor 1 and a second input of the adder 8.

The output of the motion vector predictor 12 is given to the second input of the variable length coding means 5c. In addition,
The encoding means is configured to include the above-described subtractor 1 to which the external input signal is input, to the variable-length encoding means 5c to which the variable-length code corresponding to the external input signal is output (of course, here). The configuration shown in is only an example,
A known configuration that can perform encoding corresponding to the external input signal can be used).

The first output of the variable length coding means 5c is provided to a first input of a temporary buffer 101, and the second output of the variable length coding means 5c is provided to an input of a code amount control means 102.

The output (fixed code) of the fixed code output means 104 is given to the second input of the temporary buffer 101, and the code amount control means 10 is given to the third input of the temporary buffer 101.
Two first outputs are provided. The output of the temporary buffer 101 is provided to a first input of the transmission buffer 103.

The second output of the code amount control means 102 is given to a second input of the memory 9, and the third output of the code amount control means 102 is given to a second input of the transmission buffer 103. The output of the transmission buffer 103 is output (transmitted) as a bit stream.

The output (transmitted) bit stream is received and decoded on the decoding device side.

Next, the operation will be described. Embodiment 6
Differs from the first embodiment in that DC / AC prediction is not performed even when the coding type is intra. That is, the variable length coding unit 5c performs coding using the DCT coefficient output from the quantizer 3. For example, H. In the case of an encoding device conforming to H.263, the DC component is always encoded with 8 bits in the case of an I-VOP.

Therefore, fixed code output means 104 outputs I-
For the VOP, the fixed code of the macro block in which the DC component of each block is 128, the AC components are all 0, and dquant = 0 is output.

In this case, since there is no DC prediction, when the output (fixed code) of the fixed code output means 104 is selected,
There is no need to construct a new video packet from the current macroblock. Therefore, the code amount control means 102
4, 6, 8, or 9 in the case of −VOP.
The output (fixed code) of the fixed code output unit 104 or the output of the variable-length coding unit 5c is selected according to the flowchart of FIG.

In the case of P-VOP, the first embodiment
Similarly to the fourth embodiment, the output (fixed code) of the fixed code output unit 104 or the variable length coding unit 5 according to the flowchart of FIG. 4, FIG. 6, FIG. 8, or FIG.
The output of c is selected and stored in the temporary buffer 101.

Embodiment 7 FIG. In the sixth embodiment, the output of the variable length coding unit 5c or the output (fixed code) of the fixed code output unit 104 is stored in the temporary buffer 101, and is transferred from the temporary buffer 101 to the transmission buffer 103. For example, when the data structure is not a data partition or when it is not necessary to rearrange the data, the output of the variable-length coding unit 5c or the output (fixed code) of the fixed code output unit 104 is directly output.
The input to the transmission buffer 103 may be adopted, and the temporary buffer 101 can be omitted (in this case, the storage means corresponds to the transmission buffer 103).

For example, H. In the case of an encoding device conforming to H.263 (when data partitioning is not performed), the configuration of a bit stream output from the transmission buffer 103 is as shown in FIG. Therefore, as in the case of an MPEG4 data partition (FIGS. 2 (b) and 3 (b)), the code of each macroblock is, for example, mcbppc, dquant and DC for each macroblock.
The components ac_pred_flag and cbpy,
It is not necessary to divide the codes generated for each macroblock into a category, such as coefficient data of each block, since the codes of a plurality of macroblocks are not grouped together for each category. Absent.

That is, the variable length coding means 5c
If the code of the macroblock is output in accordance with the format shown in FIG.
01 becomes unnecessary.

FIG. 14 shows an encoding apparatus according to the seventh embodiment. In the figure, reference numeral 1 denotes a subtractor having an external input signal as a first input, and an output of the subtracter 1 passes through a DCT unit 2 and a quantizer 3 and is inversed to a first input of a variable length encoding unit 5c. It is provided to the quantizer 6.

The output of the inverse quantizer 6 is supplied to the first input of the adder 8 through the inverse DCT means 7. Adder 8
Is provided to a first input of a memory 9, and an output of the memory 9 is provided to a first input of a predicted image creating means 10 and a first input of a motion detecting means 11.

An external input signal is supplied to a second input of the motion detecting means 11, and an output of the motion detecting means 11 is supplied to a second input of the predicted image creating means 10 and a motion vector predictor 12. The output of the predicted image creating means 10 is provided to a second input of the subtractor 1 and a second input of the adder 8.

The output of the motion vector predictor 12 is given to the second input of the variable length coding means 5c.

The first output of the variable length coding means 5c is provided to a first input of the transmission buffer 103, and the second output of the variable length coding means 5c is provided to an input of the code amount control means 102.

The output (fixed code) of the fixed code output means 104 is given to the second input of the transmission buffer 103, and the code amount control means 10 is input to the third input of the transmission buffer 103.
Two first outputs are provided. Code amount control means 1
The second output of 02 is provided to a second input of the memory 9.

The output of the transmission buffer 103 is output (transmitted) as a bit stream. The output (transmitted) bit stream is received and decoded on the decoding device side.

Next, the operation will be described. Embodiment 7
Is a variable length encoding means 5c and a fixed code output means 10
4 is different from the sixth embodiment in that a fixed code is output to the transmission buffer 103. That is, the code amount control means 102
Is the output of the variable length coding unit 5c or the output (fixed code) of the fixed code output unit 104 based on the code amount of the code output from the variable length coding unit 5c, as in the sixth embodiment. Is determined for each macroblock, and control is performed such that the selected one is stored in the transmission buffer 103.

In the first to seventh embodiments, the read rate of the transmission buffer 103 is assumed to be R in the setting of the maximum code amount Tmax for each VOP. Even if it is variable, the transmission buffer 10
Tmax can be set so as not to cause an overflow of 3 or an underflow of the VBV buffer.

When the read rate of the transmission buffer 103 is variable, for example, the maximum rate for transmission is determined, and the type of information to be transmitted (for example, video and This corresponds to the case where the transmission rate is allocated according to the type of voice.

Also in this case, based on the flowcharts of FIGS. 4 to 9, the output of the coding means for coding each macroblock and the fixed code determined for each VOP coding type are selected and stored. By doing so, it is possible to control so that the code amount of each VOP is equal to or less than the maximum code amount Tmax.

In the first to seventh embodiments, the case of the data partition of MPEG4 and Although the case of H.263 has been described as an example, the code amount can be controlled with the same configuration as described above even in the case of a non-data partition or MPEG2.

Further, it is needless to say that the present invention can be applied to the case where the input signal is not 4: 2: 0 or the case where the VOP (unit image) is not rectangular (for example, any shape that the object on the screen can take).

[0208]

Since the present invention is configured as described above, it has the following effects.

An encoding apparatus according to the present invention encodes an external input signal in units of macroblocks, and outputs a code corresponding to the external input signal, and determines in advance for each encoding type of a unit image. Fixed code output means for outputting the fixed code, storage means for storing the code output from the coding means or the fixed code output from the fixed code output means, Code amount control means for controlling either the code or the fixed code to be stored in the storage means in accordance with the code amount of the code. Can be restricted.

[0210] The code amount control means in the encoding apparatus according to the present invention obtains the maximum code amount that can be generated by encoding the unit image for each unit image, and based on the maximum code amount, determines the unit code amount. Each macroblock constituting an image is controlled so that either the code output from the encoding unit or the fixed code output from the fixed code output unit is stored in the storage unit. The generated code amount of an image can be easily limited.

Further, the code amount control means in the encoding apparatus according to the present invention is characterized in that the maximum code amount Tmax is obtained based on the following equation, so that overflow of the storage means can be prevented. Tmax ≦ Bs−B where Tmax is the maximum code amount, Bs is the capacity of the storage means, and B is the occupation amount in the storage means.

Further, the code amount control means in the encoding apparatus according to the present invention is characterized in that the maximum code amount Tmax is obtained based on the following equation, so that underflow of the VBV buffer can be prevented. Tmax ≦ vbv_bits + Rp where Rp = R / F, where Tmax is the maximum code amount, Rp is the number of bits read from the storage unit in the unit image, R is the bit rate read from the storage unit, and F is the unit image to be coded. , Vbv_bits is the occupation amount of the VBV buffer in the immediately preceding unit image.

Further, the code amount control means in the coding apparatus according to the present invention is characterized in that the maximum code amount Tmax is obtained based on the following equation or a value equivalent to the following equation. Overflow and V
Both underflow of the BV buffer can be avoided. Tmax ≦ min (vbv_bits + Rp, Bs−
B) where Rp = R / F, where Tmax is the maximum code amount, Rp is the number of bits read from the storage unit in the unit image, R is the bit rate read from the storage unit, and F is the unit image to be coded. The rate, vbv_bits, is the occupancy of the VBV buffer in the immediately preceding unit image, Bs is the capacity of the storage means, and B is the occupancy of the storage means.

Further, since the bit rate R read from the storage means in the encoding apparatus according to the present invention is variable, it is possible to effectively avoid overflow of the storage means or underflow of the VBV buffer. it can.

The code amount control means in the coding apparatus according to the present invention includes a code amount mb_bit of a code output from the coding means for the current macroblock, a code amount of one of the current macroblocks constituting the unit image. The code amount Sc up to the previous macroblock, the maximum code amount Tmax of the unit image, the number of macroblocks M to be coded following the current macroblock constituting the unit image, fixed for the unit image The relationship between the code length L of the fixed code output from the code output means and the code amount α of an additional code generated in video packet units constituting the unit image is Sc + mb_bit + M × L + α> Tmax (where α ≧ 0) ), The fixed code output from the fixed code output means is stored in the storage means for the current macroblock. Therefore, control can be performed such that the code amount of the unit image is always equal to or less than the maximum code amount Tmax.

Further, the code amount control means in the coding apparatus according to the present invention includes a code amount mb_bit of a code output from the coding means for the current macroblock, and one of the current macroblocks constituting the unit image. The code amount Sc up to the previous macroblock, the maximum code amount Tmax of the unit image, the number M of macroblocks to be coded following the current macroblock constituting the unit image,
The relationship between the code length L of the fixed code output from the fixed code output means for the unit image and the code amount α of the additional code generated in video packet units constituting the unit image is Sc + mb_bit + M × L + α> When Tmax (where α ≧ 0), control is performed such that the fixed code output from the fixed code output means is stored in storage means for the current macroblock,
For the M macroblocks to be coded following the current macroblock constituting the unit image, control is performed so that the fixed code output from the fixed code output unit is stored in the storage unit. Therefore, control can be performed such that the code amount of the unit image is always equal to or less than the maximum code amount Tmax with a small amount of calculation.

Further, the code amount control means in the coding apparatus according to the present invention is arranged such that when the relationship of Sc + mb_bit + M × L + α> Tmax is satisfied and the coding type of the unit image is intra, the current macro block Since a new video packet is constructed from the above, a decoded image corresponding to the code output from the fixed code output means can be easily obtained.

Further, the fixed code output means in the coding apparatus according to the present invention is characterized in that it outputs a code corresponding to a macroblock having the minimum code amount in the coding type of a unit image. It is possible to minimize the number of macroblocks from which the output of the code output unit is selected, and to minimize the deterioration of the decoded image quality.

An encoding apparatus according to the present invention encodes an external input signal in units of macroblocks, and outputs a code corresponding to the external input signal, and a predetermined coding unit for each coding type of the unit image. Fixed code output means for outputting the fixed code, storage means for storing the code output from the coding means or the fixed code output from the fixed code output means, and a first macro constituting a unit image Code amount control means for controlling either the code or the fixed code to be stored in the storage means based on the elapsed time since the block was input. Even if the processing time required to encode each macroblock is not constant, all macroblocks constituting the unit image can be encoded within a predetermined time. That.

An encoding apparatus according to the present invention encodes an external input signal in units of macroblocks, and outputs a code corresponding to the external input signal, and a predetermined unit for each encoding type of a unit image. Fixed code output means for outputting the fixed code, storage means for storing the code output from the coding means or the fixed code output from the fixed code output means, Code amount control for storing either the code or the fixed code in the storage means based on the code amount of the code and the elapsed time since the leading macroblock constituting the unit image is input. Means for avoiding overflow of the storage means or underflow of the VBV buffer, and composes a unit image. All macroblocks can be encoded within a predetermined time.

The encoding method according to the present invention encodes an external input signal in units of macroblocks, and outputs a code corresponding to the external input signal, and determines in advance for each encoding type of a unit image. A fixed code output step of providing the fixed code, a storage step of storing the code obtained in the coding step or the fixed code given from the fixed code output step, and the code coded by the coding step. And a code amount control step of controlling either the code or the fixed code to be stored in the storage step according to the code amount of the macro block. I can do it.

In the code amount control step in the encoding method according to the present invention, the maximum code amount that can be generated by encoding the unit image is obtained for each unit image, and the unit code is calculated based on the maximum code amount. For each macroblock constituting the image, control is performed such that either the code given from the encoding step or the fixed code given from the fixed code output step is stored in the storage step. The generated code amount can be easily limited.

Further, since the code amount control step in the encoding method according to the present invention is characterized in that the maximum code amount Tmax is obtained based on the following equation, overflow of the accumulation step can be prevented. Tmax ≦ Bs−B where Tmax is the maximum code amount, Bs is the capacity allowed in the accumulation step, and B is the occupation amount in the accumulation step.

Further, the code amount control step in the encoding method according to the present invention is characterized in that the maximum code amount Tmax is obtained based on the following equation, so that underflow of the VBV buffer can be prevented. Tmax ≦ vbv_bits + Rp where Rp = R / F, where Tmax is the maximum code amount, Rp is the number of bits read in the storage step in the unit image, R is the bit rate of the read in the storage step, and F is the unit image to be coded. , Vbv_bits is the occupation amount of the VBV buffer in the immediately preceding unit image.

Further, the code amount control step in the encoding method according to the present invention is characterized in that the maximum code amount Tmax is obtained based on the following equation or a value equivalent to the following equation. Overflow and V
Both underflow of the BV buffer can be avoided. Tmax ≦ min (vbv_bits + Rp, Bs−
B) where Rp = R / F, where Tmax is the maximum code amount, Rp is the number of bits read in the storage process in the unit image, R is the bit rate of read in the storage process, and F is the rate of the unit image to be coded. , Vbv_bits are the occupancy of the VBV buffer in the immediately preceding unit image, Bs is the capacity allowed in the accumulation step, and B is the occupancy in the accumulation step.

In the encoding method according to the present invention, the read bit rate R in the storage step is variable, so that it is possible to effectively avoid overflow of the storage step or underflow of the VBV buffer. it can.

In the coding amount control step in the coding method according to the present invention, the code amount mb_bit of the code given from the coding step for the current macroblock, one of the current macroblocks constituting the unit image, The code amount Sc up to the previous macroblock, the maximum code amount Tmax of the unit image, the number M of macroblocks to be coded following the current macroblock constituting the unit image, a fixed code for the unit image The relationship between the code length L of the code output by the output step and the code amount α of the additional code generated in video packet units constituting the unit image is Sc + mb_bit + M × L + α> Tmax (where α ≧ 0). In some cases, for the current macroblock, control is performed so that the fixed code given from the fixed code output step is accumulated in the accumulation step. Therefore, control can be performed such that the code amount of the unit image is always equal to or less than the maximum code amount Tmax.

Further, the code amount control step in the encoding method according to the present invention includes the code amount mb_bit of the code given from the encoding step for the current macroblock, and one of the current macroblocks constituting the unit image. The code amount Sc up to the previous macroblock, the maximum code amount Tmax of the unit image, the number M of macroblocks to be coded following the current macroblock constituting the unit image, The relationship between the code length L of the fixed code given from the fixed code output step and the code amount α of the additional code generated in video packet units constituting the unit image is Sc + mb_bit + M × L + α> Tmax (where α ≧ 0), the fixed code provided from the fixed code output step is stored for the current macroblock. Control to accumulate in the step, and also for the M macroblocks to be encoded following the current macroblock constituting the unit image, the fixed code given from the fixed code output step Since the control is performed so as to accumulate in the accumulation step, it is possible to control the code amount of the unit image to be always equal to or less than the maximum code amount Tmax with a small amount of calculation.

In the coding amount control step in the coding method according to the present invention, when the relation of Sc + mb_bit + M × L + α> Tmax is satisfied and the coding type of the unit image is intra, the current macroblock Since a new video packet is constructed from the above, a decoded image corresponding to the code output from the fixed code output means can be easily obtained.

Further, the fixed code output step in the coding method according to the present invention is characterized by outputting a code corresponding to a macroblock having the minimum code amount in the coding type of a unit image. It is possible to minimize the number of macroblocks from which the output of the code output unit is selected, and to minimize the deterioration of the decoded image quality.

The encoding method according to the present invention encodes an external input signal in units of macroblocks, and outputs a code corresponding to the external input signal. A fixed code output step of providing the fixed code, an accumulation step of accumulating the code given from the encoding step or the fixed code given from the fixed code output step, and a top macroblock constituting the unit image. A code amount control step of controlling either the code or the fixed code to be stored in the storage step based on the elapsed time from the input. Even if the processing time required to encode a block is not constant, all macroblocks constituting a unit image can be encoded within a predetermined time. That.

The encoding method according to the present invention encodes an external input signal in units of macroblocks, and outputs a code corresponding to the external input signal. A fixed code output step of outputting the fixed code, a storage step of storing the code given from the encoding step or the fixed code given from the fixed code output step, and the encoding step performed by the encoding step. A code amount control step of controlling either the code or the fixed code to be stored in the storage step based on a code amount of a code and an elapsed time since a leading macroblock constituting the unit image is input; In order to avoid the overflow of the storage means or the underflow of the VBV buffer, All macroblocks formed can be encoded within a predetermined time.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a first embodiment of the present invention.

FIG. 2 is an explanatory diagram showing states of a temporary buffer and a transmission buffer (in the case of an I-VOP) according to Embodiment 1 of the present invention;

FIG. 3 is an explanatory diagram showing states of a temporary buffer and a transmission buffer (in the case of P-VOP) according to Embodiment 1 of the present invention;

FIG. 4 is a flowchart (for P-VOP) showing Embodiment 1 of the present invention.

FIG. 5 is a flowchart (in the case of an I-VOP) showing the first embodiment of the present invention;

FIG. 6 is a flowchart (for P-VOP) showing Embodiment 2 of the present invention.

FIG. 7 is a flowchart (in the case of an I-VOP) showing Embodiment 2 of the present invention.

FIG. 8 is a flowchart showing a third embodiment of the present invention.

FIG. 9 is a flowchart showing a fourth embodiment of the present invention.

FIG. 10 is a block diagram showing a fifth embodiment of the present invention.

FIG. 11 is an explanatory diagram showing a state of a temporary buffer according to Embodiment 5 of the present invention.

FIG. 12 is a block diagram showing a sixth embodiment of the present invention.

FIG. 13 is an explanatory diagram showing a state of a transmission buffer according to Embodiment 7 of the present invention.

FIG. 14 is a block diagram showing a seventh embodiment of the present invention.

FIG. 15 is a block diagram showing a conventional encoding device.

FIG. 16 is an explanatory diagram showing an input signal to a conventional encoding device.

FIG. 17 is an explanatory diagram showing a configuration of a bit stream in a conventional encoding device.

FIG. 18 is an explanatory diagram showing a position on a screen of a video packet in a conventional encoding device.

FIG. 19 is an explanatory diagram showing DC / AC prediction in a conventional encoding device.

[Explanation of symbols]

5a, 5b, 5c Variable length coding means, 101 temporary buffer, 102 code amount control means, 103 transmission buffer, 104 fixed code output means.

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Junko Kijima 2-3-2 Marunouchi, Chiyoda-ku, Tokyo Inside Mitsubishi Electric Corporation (72) Inventor Mamoru Inamura 2-3-2 Marunouchi, Chiyoda-ku, Tokyo Within Rishi Electric Co., Ltd. (72) Inventor Kazuhiro Sugiyama 2-3-2 Marunouchi, Chiyoda-ku, Tokyo F-term (reference) 5C059 KK35 LA10 MA00 MA04 MA05 MA23 MC11 MC33 MC35 ME05 NN01 PP16 RB02 RB09 RC09 RC12 RC16 SS07 SS10 TA17 TB07 TC20 TC36 UA02 UA32 5J064 BA01 BA09 BA16 BB03 BB12 BB13 BC01 BC16 BD02

Claims (24)

[Claims] An encoding means for encoding an external input signal on a macroblock basis and outputting a code corresponding to the external input signal, and a fixed output for outputting a fixed code predetermined for each encoding type of a unit image. Code output means, storage means for storing the code output from the coding means or the fixed code output from the fixed code output means, and a code amount of the code coded by the coding means. A coding amount control unit that controls to store either the code or the fixed code in the storage unit in response to the code amount. 2. The code amount control means calculates a maximum code amount that can be generated by encoding a unit image for each of the unit images, and for each macro block configuring the unit image based on the maximum code amount. 2. The encoding apparatus according to claim 1, wherein control is performed such that either the code output from the encoding means or the fixed code output from the fixed code output means is stored in the storage means. 3. The code amount control means includes a maximum code amount Tmax.
Is calculated based on the following equation:
An encoding device according to claim 1. Tmax ≦ Bs−B where Tmax is the maximum code amount, Bs is the capacity of the storage means, and B is the occupation amount in the storage means. 4. The code amount control means includes a maximum code amount Tmax.
Is calculated based on the following equation:
An encoding device according to claim 1. Tmax ≦ vbv_bits + Rp where Rp = R / F, where Tmax is the maximum code amount, Rp is the number of bits read from the storage unit in the unit image, R is the bit rate read from the storage unit, and F is the unit image to be coded. , Vbv_bits is the occupation amount of the VBV buffer in the immediately preceding unit image. 5. The code amount control means calculates the maximum code amount Tma based on the following equation or a value equivalent to the following equation:
3. The encoding apparatus according to claim 2, wherein x is obtained. Tmax ≦ min (vbv_bits + Rp, Bs−
B) where Rp = R / F, where Tmax is the maximum code amount, Rp is the number of bits read from the storage unit in the unit image, R is the bit rate read from the storage unit, and F is the unit image to be coded. The rate, vbv_bits, is the occupancy of the VBV buffer in the immediately preceding unit image, Bs is the capacity of the storage means, and B is the occupancy of the storage means. 6. The encoding device according to claim 4, wherein the bit rate R read from the storage means is variable. 7. The code amount control means includes a code amount mb_b of a code output from the coding means for a current macroblock.
it, the code amount Sc up to the macroblock immediately before the current macroblock constituting the unit image, the maximum code amount Tmax of the unit image, the code amount which is encoded following the current macroblock constituting the unit image. Between the number M of macroblocks to be processed, the code length L of the fixed code output from the fixed code output means for the unit image, and the code amount α of the additional code generated in video packet units constituting the unit image. When the relation is Sc + mb_bit + M × L + α> Tmax (where α ≧ 0), control is performed such that the fixed code output from the fixed code output means is stored in the storage means for the current macroblock. 3. The encoding device according to claim 2, wherein: 8. The code amount control means includes a code amount mb_b of a code output from the coding means for the current macroblock.
it, the code amount Sc up to the macroblock immediately before the current macroblock constituting the unit image, the maximum code amount Tmax of the unit image, encoding following the current macroblock constituting the unit image Between the number M of macroblocks to be performed, the code length L of the fixed code output from the fixed code output means for the unit image, and the code amount α of an additional code generated in video packet units constituting the unit image If the relationship of Sc + mb_bit + M × L + α> Tmax (where α ≧ 0), control is performed such that the fixed code output from the fixed code output means is stored in storage means for the current macroblock. With
For the M macroblocks to be coded following the current macroblock constituting the unit image, control is performed so that the fixed code output from the fixed code output unit is stored in the storage unit. The encoding device according to claim 2, wherein: 9. The code amount control means, when the relation of Sc + mb_bit + M × L + α> Tmax is satisfied and the coding type of the unit image is intra, configures a new video packet from the current macroblock. 9. The method according to claim 8, wherein
An encoding device according to claim 1. 10. The fixed code output means outputs a code corresponding to a macroblock having a minimum code amount in a coding type of a unit image.
An encoding device according to claim 1. 11. An encoding means for encoding an external input signal in units of macroblocks and outputting a code corresponding to the external input signal, and a fixed means for outputting a fixed code predetermined for each coding type of the unit image. Code output means, storage means for storing the fixed code output from the code or the fixed code output means output from the encoding means,
Code amount control means for controlling to store either the code or the fixed code in the storage means based on the elapsed time since the leading macroblock constituting the unit image is input. Encoding device. 12. An encoding means for encoding an external input signal in macroblock units and outputting a code corresponding to the external input signal, and a fixed means for outputting a fixed code predetermined for each coding type of the unit image. Code output means; storage means for storing the code output from the coding means or the fixed code output from the fixed code output means; and the code amount of the code coded by the coding means and Code amount control means for controlling to store either the code or the fixed code in the storage means based on the elapsed time since the leading macroblock constituting the unit image is input. Encoding device. 13. An encoding step of encoding an external input signal in units of macroblocks and outputting a code corresponding to the external input signal, and a fixed code for providing a predetermined fixed code for each coding type of the unit image. An output step, an accumulation step of accumulating the fixed code obtained from the code or the fixed code output step obtained by the encoding step, and a code amount of the code encoded by the encoding step. A code amount control step of controlling either the code or the fixed code to be stored in the storage step. 14. The code amount control step is for obtaining a maximum code amount that can be generated by encoding a unit image for each unit image, and for each macro block constituting the unit image based on the maximum code amount. 14. The encoding method according to claim 13, wherein one of a code given from the encoding step and a fixed code given from the fixed code output step is controlled to be accumulated in the accumulation step. 15. The code amount control step includes: a maximum code amount Tma.
15. The encoding method according to claim 14, wherein x is obtained based on the following equation. Tmax ≦ Bs−B where Tmax is the maximum code amount, Bs is the capacity allowed in the accumulation step, and B is the occupation amount in the accumulation step. 16. The code amount control step includes the step of:
15. The encoding method according to claim 14, wherein x is obtained based on the following equation. Tmax ≦ vbv_bits + Rp where Rp = R / F, where Tmax is the maximum code amount, Rp is the number of bits read in the storage step in the unit image, R is the bit rate of the read in the storage step, and F is the unit image to be coded. , Vbv_bits is the occupation amount of the VBV buffer in the immediately preceding unit image. 17. The code amount control step includes calculating a maximum code amount Tm based on the following expression or a value equivalent to the following expression.
The encoding method according to claim 14, wherein ax is obtained. Tmax ≦ min (vbv_bits + Rp, Bs−
B) where Rp = R / F, where Tmax is the maximum code amount, Rp is the number of bits read in the storage process in the unit image, R is the bit rate of read in the storage process, and F is the rate of the unit image to be coded. , Vbv_bits are the occupancy of the VBV buffer in the immediately preceding unit image, Bs is the capacity allowed in the accumulation step, and B is the occupancy in the accumulation step. 18. The encoding method according to claim 16, wherein the read bit rate R in the accumulation step is variable. 19. The code amount controlling step includes the step of controlling the code amount m of the code given from the encoding step for the current macroblock.
b_bit, the code amount Sc up to the macroblock immediately before the current macroblock forming the unit image, the maximum code amount Tmax of the unit image, and the current macroblock forming the unit image is coded following the current macroblock. The relationship between the number of macroblocks M to be expended, the code length L of the code output by the fixed code output step for the unit image, and the code amount α of the additional code generated in video packet units constituting the unit image is as follows. , Sc + mb_bit + M × L + α> Tmax (where α ≧ 0), control is performed such that the fixed code given from the fixed code output step is accumulated in the accumulation step for the current macroblock. The encoding method according to claim 14. 20. The code amount control step includes a code amount m of a code given from the encoding step for the current macroblock.
b_bit, the code amount Sc up to the macroblock immediately before the current macroblock constituting the unit image, the maximum code amount Tmax of the unit image, encoding following the current macroblock constituting the unit image Between the number M of macroblocks to be performed, the code length L of a fixed code given from the fixed code output step for the unit image, and the code amount α of an additional code generated in video packet units constituting the unit image When the relationship of Sc + mb_bit + M × L + α> Tmax (where α ≧ 0) is satisfied, the current macroblock is controlled so that the fixed code given from the fixed code output step is accumulated in the accumulation step. And M macroblocks to be encoded following the current macroblock constituting the unit image 15. The encoding method according to claim 14, further comprising controlling the fixed code provided from the fixed code output step to be stored in the storage step. 21. A code amount control step, when the relation of Sc + mb_bit + M × L + α> Tmax is satisfied and the coding type of the unit image is intra, configures a new video packet from the current macroblock. 3. The method according to claim 2, wherein
0. The encoding method according to 0. 22. The fixed code output step, wherein a code corresponding to a macroblock having the minimum code amount in a coding type of a unit image is output.
5. The encoding method according to 4. 23. An encoding step of encoding an external input signal in units of macroblocks and outputting a code corresponding to the external input signal, and a fixed code for providing a predetermined fixed code for each coding type of the unit image. An output step; an accumulating step of accumulating the code given from the encoding step or the fixed code given from the fixed code output step; and an elapsed time since a leading macroblock constituting the unit image is input. A code amount control step of controlling either the code or the fixed code to be stored in the storage step based on the following. 24. An encoding step of encoding an external input signal in units of macroblocks and outputting a code corresponding to the external input signal, and outputting a fixed code predetermined for each encoding type of the unit image. A code output step; an accumulation step of accumulating the code given from the encoding step or the fixed code given from the fixed code output step; a code amount of the code encoded by the encoding step and the unit And a code amount control step of controlling to store either the code or the fixed code in the storage step based on the elapsed time since the leading macroblock constituting the image is input. Encoding method to be used.