JP2001157210A - Encoder and encoding method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To transmit encoded data produced through encoding without missing. SOLUTION: The encoder is provide with an encoding section that applies compression encoding to a moving picture to calculate a generated code amount of encoded data generated from an encoded object image, an original image data storage section that stores the encoding object image, an encoding data storage section that stores the encoded data, and a generated code amount comparison section that obtains a threshold value with respect to the generated code amount for each encoded object image so as to discriminate whether or not the generated code amount exceeds the threshold value. In the case that the object image can be re-encoded because there is a time margin to re-encoded the encoding object image until the encoding of the image to be encoded next is started and when it is discriminated that the generated code amount exceeds a threshold value, the encoding section revises an encoding parameter to decrease the generated code amount and applies re-encoding to the encoding object image.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an MPEG (Moving
The present invention relates to an encoding device that performs moving image compression encoding using an algorithm such as a Picture Experts Group.

[0002]

2. Description of the Related Art MPEG is used as a moving picture compression encoding method.
It has been known. In the MPEG system, a virtual decoder model (VBV: Video Bufferin) connected to an encoder output is used.
g Verifier). This model includes a reception buffer (VBV buffer), and the restriction on the code amount of the encoded data is defined by the constraint on the occupation amount of the VBV buffer. The constraint is that overflow and underflow should not occur in the VBV buffer.

[0003] It is assumed that a VBV buffer operates under the following ideal conditions. (1) The encoder and VBV operate completely synchronously. (2) The decoding of each picture is performed instantaneously, and the encoded data for each picture is immediately extracted from the VBV buffer.

[0004] The occupancy of the VBV buffer is calculated from the cumulative value of the amount of data transmitted by the transmission path from the encoder to the decoder.
Encoded data to be decoded by the decoder, ie,
It can be said that the value indicates a value obtained by subtracting the accumulated value of the data amount of the encoded data output from the encoder.

[0005] The overflow of the VBV buffer means that the encoded data is extracted from the VBV buffer before being extracted from the VBV buffer.
This means that the occupancy of the buffer reaches the capacity of the VBV buffer. The underflow of the VBV buffer means that the occupation amount of the VBV buffer is insufficient when extracting encoded data from the VBV buffer because the data amount of one picture is large.

When an underflow occurs, the accumulated value of the data amount of the encoded data output from the encoder becomes larger than the accumulated value of the transmitted data amount, so that all of the generated encoded data is transmitted. Cannot be performed, and data necessary for decoding is lost from data received by the decoder. Therefore, when underflow is likely to occur, it is necessary to control the encoder so as to reduce the amount of codes generated for the image to be encoded, and to prevent the occurrence of underflow.

[0007] An image to be encoded is divided into small areas of 16 x 16 pixels called macroblocks. In a conventional encoding apparatus, when an underflow is likely to occur in a VBV buffer, the amount of generated code is controlled by changing a quantization step for each macroblock in an image being encoded. ing.

[0008]

However, MPEG-1
When encoding is performed on an image having a small image size (the number of pixels is small), which is often handled by the system, it is difficult to control the generated code amount because the number of macroblocks in one image is small. The number of macroblocks is, for example, 1620 for an image with an image size of 720 × 576 pixels, but 3 for an image with an image size of 352 × 288 pixels.
96. For this reason, the generated code amount per image is increased, and an underflow is often generated in the VBV buffer.

According to the present invention, the image data is re-encoded as necessary when the encoding processing capacity is sufficient so that encoded data generated by encoding can be transmitted without loss. An object of the present invention is to provide an encoding device for performing the above.

[0010]

Means for Solving the Problems In order to solve the above-mentioned problems, a means according to the invention of claim 1 is a coding apparatus that performs compression coding on a moving image and performs coding on a moving image, An encoding unit that calculates a generated code amount of the generated encoded data, an original image data storage unit that stores the encoding target image, an encoded data storage unit that stores the encoding data, and an encoding target image A threshold for the generated code amount, and a generated code amount comparison unit for determining whether or not the generated code amount exceeds the threshold value. There is a margin of time to re-encode the encoding target image until the encoding unit determines that it is possible to re-encode,
And when it is determined by the generated code amount comparison unit that the generated code amount exceeds the threshold value, the encoding unit changes an encoding parameter so as to reduce the generated code amount, and The encoding target image is re-encoded.

According to the first aspect of the present invention, when there is enough time for re-encoding, it is possible to reduce the amount of codes generated by image coding so that the amount of generated codes does not exceed a threshold value. it can.

According to a second aspect of the present invention, in the encoding apparatus according to the first aspect, the threshold value is VBV (Vi
(deo Buffering Verifier) The buffer has a code amount determined so as not to cause an underflow.

According to the second aspect of the present invention, the generated code amount is controlled so that the VBV buffer does not underflow, that is, the coded data output from the coding device satisfies the restriction on the code amount. be able to.

According to a third aspect of the present invention, in the encoding apparatus according to the first aspect, the encoding section performs variable length encoding on the moving image.

According to the third aspect of the present invention, the amount of generated codes can be easily reduced by changing the encoding parameters.

According to a fourth aspect of the present invention, in the encoding apparatus according to the first aspect, the change of the encoding parameter increases a quantization step.

According to the fourth aspect of the present invention, it is possible to surely reduce the generated code amount so as not to exceed the threshold value.

According to a fifth aspect of the present invention, in the encoding apparatus according to the first aspect, the encoding of the moving image is completed such that encoding of one image is completed within a time interval of the image interval of the moving image. It is characterized in that it is performed in real time.

According to the fifth aspect of the present invention, the amount of generated codes can be reduced as needed under the condition of performing real-time encoding.

According to a sixth aspect of the present invention, in the encoding device according to the fifth aspect, when the re-encoding is possible,
It is characterized in that the number of pixels of the image to be encoded is less than half the maximum number of pixels of the image that can be processed by the encoding unit.

According to the sixth aspect of the present invention, since the number of pixels of an image is small, re-encoding can be performed, and the amount of code generated by image encoding can be reduced.

According to a seventh aspect of the present invention, in the encoding apparatus according to the fifth aspect, the encoding apparatus further comprises an encoding number counter for counting the number of encodings of the image to be encoded by the encoding unit. In the case where encoding is possible, the number of pixels of the encoding target image is equal to or less than N (N is an integer of 3 or more) of the maximum number of pixels of the image that can be processed by the encoding unit, It is characterized in that the number of times of encoding for the encoding target image counted by the encoding number counter is less than N times.

According to the seventh aspect of the present invention, since the coding parameters can be changed little by little, the deterioration of the image quality due to the decrease in the generated code amount can be suppressed to a small level.

According to the invention of claim 8, in the encoding apparatus according to claim 7, the change of the encoding parameter is performed when the encoding target image is encoded up to the (N-1) th encoding. The quantization step is gradually increased, and in the N-th encoding of the encoding target image, the quantization step is maximized within an allowable range.

According to the eighth aspect of the invention, it is possible to suppress the deterioration of the image quality as much as possible and to ensure that the generated code amount does not exceed the threshold value.

According to a ninth aspect of the present invention, in the encoding apparatus according to the first aspect, when encoding is started, the encoding unit re-encodes an image to be encoded first. It is characterized by performing.

According to the ninth aspect of the present invention, it is possible to improve the image quality of an image to be encoded first while preventing the generated code amount from exceeding the threshold value.

According to a tenth aspect of the present invention, in the encoding apparatus according to the first aspect, when the encoding is stopped, the encoding unit re-encodes the last image to be encoded. It is characterized by performing.

According to the tenth aspect, it is possible to improve the image quality of an image to be encoded last, while keeping the generated code amount from exceeding the threshold value.

[0030] According to an eleventh aspect of the present invention, as a coding method, a step of performing compression coding on a moving image and a step of calculating a generated code amount of coded data generated for a coding target image. Determining a threshold value for the generated code amount for each image to be coded, and determining whether the generated code amount exceeds the threshold value. It is determined that there is a margin for re-encoding the image to be encoded before encoding is started, and that re-encoding is possible, and that the generated code amount exceeds the threshold. Sometimes, in the step of performing the compression encoding, the encoding parameter is changed so as to reduce the generated code amount, and the encoding target image is re-encoded.

According to the eleventh aspect of the present invention, when there is enough time for re-encoding, it is possible to reduce the amount of codes generated by image coding so that the amount of generated codes does not exceed a threshold value. it can.

[0032]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, an encoding device according to an embodiment of the present invention will be described with reference to the drawings. Hereinafter, as an example, a case where an image having a frame structure (frame picture) is encoded will be described.

FIG. 1 is a block diagram showing a configuration of an encoding device according to an embodiment of the present invention. The encoding device of FIG.
An encoding unit 11, an original image data storage unit 12, an encoded data storage unit 13, and a generated code amount comparison unit 14 are provided.

The encoding unit 11 receives an image to be encoded one picture at a time from the outside and outputs it as it is to the original image data storage unit 12. The original image data storage unit 12 stores the input image data. Also, the encoding unit 11
Reads image data from the original image data storage unit 12 and performs moving image compression encoding by a method such as MPEG.
The obtained encoded data is output to the encoded data storage unit 13, and the generated code amount is calculated and output to the generated code amount comparison unit 14. The encoding unit 11 calculates 8 × 8
DCT (discrete cosine transfor
m), and the obtained DCT coefficients are subjected to variable-length coding, thereby performing moving image compression coding.

The encoded data storage section 13 stores the inputted encoded data and outputs it to the encoding section 11 as needed. The generated code amount comparison unit 14 calculates a threshold value of the generated code amount for each picture, compares the data of the generated code amount with this threshold value, and when the generated code amount exceeds the threshold value, , And outputs a request to change the encoding parameter to the encoding unit 11. The threshold is a code amount that does not cause underflow in the VBV buffer, and is usually the maximum value of such a code amount.

When receiving the request for changing the coding parameter, the coding unit 11 changes the coding parameter so as to reduce the generated code amount, and replaces the coding target image which has already been coded with the original image data. Read again from the storage unit 12,
The image is re-encoded, and the encoded data is output to the encoded data storage unit 13. The encoded data storage unit 13 stores the input encoded data. The encoding unit 11 reads the encoded data from the encoded data storage unit 13 and outputs the encoded data to the outside as encoded data.

The encoding unit 11 shown in FIG.
It has the ability to encode an image of × 576 pixels in real time by the MPEG-2 system. This encoding unit 11
An image having an image size smaller than 720 × 576 pixels can be encoded. Also, encoding can be performed by the MPEG-1 system. Hereinafter, as an example, the encoding apparatus in FIG. 1 converts an image having an image size of 352 × 288 pixels into M
The case of encoding by the PEG-1 method will be described. In this example, the number of pixels of the image to be coded is
Is less than half the maximum number of pixels of the image that can be processed.

FIG. 2 is a timing chart for explaining the operation of the encoding apparatus shown in FIG. In FIG. 2, TA is the time at which image processing is started, and TB is the image size of 352 × 28.
The time at which the first encoding of the 8-pixel image ends, TC is the start time of re-encoding for this image, TD is the end time of re-encoding, and TE is 1 for real-time encoding.
This is the time at which the permissible processing time per picture ends.
During the period from TB to TC, the generated code amount is compared with a threshold value,
And a period required for changing the encoding parameter. FIG.
The encoding device of the image size 7
Encoding of one picture of 20 × 576 pixels can be completed. The period from TA to TE is equal to the frame interval (for example, 1/30 second) of the moving image input to the encoding unit 11.

In general, the size of an image having a maximum number of pixels that can be processed by an encoder (hereinafter, the size of this image is referred to as a maximum image size) is smaller than that of a case where encoding is performed. When encoding is performed on an image having a small number of pixels, the processing time required for encoding is short. For example, when encoding an image having an image size of 352 × 288 pixels, a simple calculation is performed based on the number of pixels.
It is possible to perform encoding in a time that is equal to or less than の of the case where an image having an image size of 720 × 576 pixels is encoded by the same encoding method. Also, when using an encoding method that can be encoded with a smaller amount of operation than the MPEG-2 method,
Encoding can be performed in a shorter time.

Therefore, when the encoding unit 11 of FIG. 1 encodes an image having an image size of 352.times.288 pixels by the MPEG-1 method, even if the encoding must be performed in real time, I have time to spare. That is, as shown in FIG. 2, encoding can be performed once in the period from TA to TB, and the period from TB to TE is a margin time, so that the same image can be re-encoded in this period. It is.

FIG. 3 is a flowchart showing the flow of processing when the encoding apparatus of FIG. 1 encodes one picture. The operation of the encoding device in FIG. 1 will be described with reference to FIGS. The encoding unit 11 knows that the number of pixels of the input image is 352 × 288 from the parameters related to the encoding input together with the image data,
Since this number of pixels is equal to or less than half the number of pixels of the maximum image size, it is assumed that it has been determined that re-encoding is possible.

In step S1 of FIG. 3, the generated code amount comparison unit 14 determines whether the underflow has occurred in the VBV buffer based on the occupation amount of the VBV buffer, the input rate of the encoded data to the VBV buffer, and the time until the start of decoding the next picture. The maximum value of the code amount of one picture that does not cause the error is calculated for each picture and stored as a threshold value of the generated code amount.

In step S2, the data of one picture to be encoded is input from the outside to the encoding unit 11, and the encoding unit 11 outputs the image data to the original image data storage unit 12 as it is. The original image data storage unit 12 stores the input image data.

In step S3, the encoding unit 11 performs
During the period from TA to TB in FIG.
, The encoding target image and the data of the image necessary for predictive encoding are read out, the first encoding is performed on the encoding target image, and the generated encoded data is output to the encoded data storage unit 13. The encoded data storage unit 13 stores the input encoded data.

In step S4, the encoding unit 11 calculates the generated code amount and outputs it to the generated code amount comparison unit 14.

In step S5, the generated code amount comparing unit 14
Compares the input generated code amount with the threshold value obtained in step S1. When the generated code amount comparison unit 14 determines that the generated code amount does not exceed the threshold, the process proceeds to step S6. If it is determined that the generated code amount exceeds the threshold, that is, if it is determined that the VBV buffer causes an underflow, step S
Processing proceeds to 7.

In step S6, the encoding unit 11 reads out the encoded data from the encoded data storage unit 13, determines the encoded data, and outputs it to the outside at time TE in FIG. This is 1
The encoding of the picture ends.

In step S7, the generated code amount comparison unit 14
Outputs an encoding parameter change request to the encoding unit 11 so as to reduce the generated code amount during the period from TB to TC in FIG. The encoding unit 11 changes the encoding parameters according to the request.

In step S8, the encoding unit 11 re-encodes the same image as the first encoding target image (periods TC to TD in FIG. 2). The encoding unit 11 outputs the encoded data generated by the re-encoding to the encoded data storage unit 13, and the encoded data storage unit 13 converts the data into encoded data generated by the first encoding. Remember in the form of overwriting. Thereafter, the process proceeds to step S6, and the encoding of one picture is completed.

A method of calculating the threshold value of the generated code amount in step S1 in FIG. 3 will be described. FIG. 4 is a graph showing a temporal change in the occupation amount of the VBV buffer by the encoded data. In FIG. 4, the vertical axis represents the VBV buffer occupancy, the horizontal axis represents time, and B represents the capacity of the VBV buffer.

The generated code amount of the previous picture coded immediately before the current picture to be coded (hereinafter referred to as the current picture) is represented by C (n), and the VBV buffer immediately before the decoding of the previous picture is performed. Is assumed to be Bp (n), the decoding of the previous picture is instantaneously performed at time t (n), and the occupancy of the VBV buffer after the encoded data for the previous picture is instantaneously extracted is , B
(N) = Bp (n) -C (n). Assuming that the interval for decoding an image is T [s] and the data input rate to the VBV buffer is a constant value R [bit / s], the VBV buffer immediately before decoding at the decoding time t (n + 1) of the current picture is used. The occupation amount is Bp (n + 1) = B (n) + RT. Here, T = t (n + 1) -t (n).

The generated code amount C (n + 1) of the current picture is
If the occupation amount of the VBV buffer is larger than Bp (n + 1), encoded data of the current picture to be decoded cannot be prepared at time t (n + 1), causing an underflow in the VBV buffer. Therefore, VBV
The occupation amount Bp (n + 1) of the buffer is set as a threshold value of the generated code amount for the current picture.

The change of the coding parameter in step S7 in FIG. 3 is, for example, a change to increase the quantization step which is one of the coding parameters. In the present embodiment, the quantization step is set to a maximum value within an allowable range in order to surely reduce the generated code amount and prevent the VBV buffer from underflowing. The quantization step for re-encoding may be any other value as long as the value is larger than that at the time of the first encoding.

By increasing the quantization step, the D
The number of CT coefficients having a value of zero increases. Since the encoding unit 11 performs variable-length encoding on the DCT coefficient, the encoding amount can be reduced.

FIG. 5 is a block diagram showing another example of the configuration of the encoding apparatus shown in FIG. The encoding device in FIG. 5 includes an encoding unit 21, an original image data storage unit 12, an encoded data storage unit 13, a generated code amount comparison unit 14, and a generated code amount calculation unit 15.

The original image data storage unit 12, encoded data storage unit 13, and generated code amount comparison unit 14 are the same as those described with reference to FIG. The encoding unit 21
It is the same as the encoding unit 11 of FIG. 1 except that the generated code amount is not calculated. The generated code amount calculation unit 15 receives the encoded data output from the encoding unit 21 as input, calculates the data amount of the encoded data, that is, the generated code amount, and outputs the calculated data amount to the generated code amount comparison unit 14. Therefore, the encoding device of FIG. 5 can perform the same operation as that of the encoding device of FIG.

The threshold of the generated code amount may be determined not by the generated code amount comparison unit 14 but by the generated code amount calculation unit 15.

As described above, according to the present embodiment, when an image smaller than the maximum image size that can be processed is input and there is a margin in the processing time, encoding is performed in real time without causing underflow in the VBV buffer. It is possible to provide a coding device that can be converted.

(Modification) The case where re-encoding is performed only once has been described above, but a modification in which re-encoding is performed a plurality of times will be described.

FIG. 6 is a block diagram showing a configuration of an encoding device according to a modification of the embodiment of the present invention. The encoding device in FIG. 6 includes an encoding unit 31, an original image data storage unit 12,
An encoded data storage unit 13, a generated code amount comparison unit 14,
An encoding frequency counter 16 is provided. The original image data storage unit 12, the encoded data storage unit 13, and the generated code amount comparison unit 14 are the same as those described with reference to FIG. 1, so the same numbers are assigned and the description is omitted.

It is assumed that the encoding unit 31 shown in FIG. 6 has an ability to encode an image having the number of pixels M in real time according to the MPEG-2 system, and the encoding apparatus shown in FIG. Will be described in accordance with the MPEG-1 method. In this case, the same image can be encoded N times in real time. Here, M and N are integers of 3 or more.

The encoding number counter 16 includes an encoding unit 31
Receives the encoding end signal generated each time encoding ends, counts the number of encodings performed on the same encoding target image, and notifies the encoding unit 31 of the count.

When the number of times of encoding has not reached N and the encoding unit 31 receives a request to change the encoding parameter from the generated code amount comparing unit 14, the encoding unit 31 changes the encoding parameter and re-encodes it. Perform the conversion. The method of changing the encoding parameter is different from that of the encoding unit 11 in FIG. Encoding unit 3
1 is to reduce the generated code amount little by little at the time of the second to (N-1) -th coding, and to reduce the generated code amount as much as possible at the time of the N-th coding. Change parameters.

FIG. 7 is a timing chart for explaining the operation when the encoding apparatus of FIG. 6 performs re-encoding of an image having the number of pixels M / N a plurality of times.

In FIG. 7, Ta is a time when image processing starts, Tb is a time when the first encoding ends, Tc is a second encoding start time, and Td is a second encoding end time. , Te is the start time of the third encoding, Tf is the start time of the Nth encoding, Tg is the end time of the Nth encoding, and Th is the allowable processing time per picture for realizing real-time encoding. Is the time to end. Tb to T
The period of c, Td to Te is the time required for comparing the generated code amount with the threshold value and changing the coding parameter.
The encoding apparatus of FIG. 6 performs the pixel number M within the period from Ta to Th.
, The encoding of one picture is completed.

FIG. 8 is a flowchart showing the flow of processing when the encoding apparatus of FIG. 6 performs re-encoding a plurality of times to encode one picture. Steps S1 to S6 correspond to FIG.
The description is omitted because it is the same as that described above. The encoding unit 31 knows that the number of pixels of the input image is M / N from the parameters related to the encoding input together with the image data, and this pixel number is 1 / N of the number of pixels of the maximum image size. Therefore, it is assumed that it has been determined that encoding is possible N times.

After calculating the generated code amount in step S4, in step S9, the encoding number counter 16 increases the count number by one, and notifies the encoding unit 31 of the encoding number for the encoding target image. I do.

In step S10, the encoding unit 31 determines whether or not encoding of this image has been performed N times based on the number of encodings. If the encoding has already been performed N times, the process proceeds to step S6, where the encoding unit 31 outputs the encoded data to the outside and ends the encoding for this image. If the encoding has not been performed N times, the process proceeds to step S5.

In step S5, the generated code amount comparing section 14
Compares the generated code amount obtained in step S4 with a threshold value. If the generated code amount does not exceed the threshold value, the process proceeds to step S6, and the coding of one picture to be coded ends. If the generated code amount exceeds the threshold, the process proceeds to step S11.

In step S11, the generated code amount comparing unit 1
Reference numeral 4 denotes an encoding unit 3 in a period such as Tb to Tc in FIG.
A request to change the encoding parameter is output so as to suppress the generated code amount for the number 1. The encoding unit 31 changes the encoding parameters according to the request. Then, step S3
To perform re-encoding.

The change of the encoding parameter in step S11 is, for example, a change to increase the quantization step. Because re-encoding can be performed multiple times,
In consideration of the number N of times that the same image can be encoded, the quantization step is gradually increased step by step at each re-encoding. In the N-th encoding, which is the last encoding opportunity, the quantization step is set to the maximum value within an allowable range in order to surely reduce the amount of generated codes and prevent underflow of the VBV buffer. . The quantization step for the N-th encoding may be another value as long as the value is larger than that at the time of the encoding up to the (N-1) -th encoding.

As described above, according to the present modification, even before the number of times of encoding for the same image reaches N,
If the generated code amount comparison unit 14 determines in step S5 that the generated code amount does not exceed the threshold,
That is, when it is determined that no underflow occurs in the VBV buffer, the subsequent re-encoding is not performed.
Therefore, it is possible to suppress the deterioration of the image quality without making the quantization step unnecessarily large.

In the above embodiment, when it is determined from the relationship between the number of pixels of the input image and the number of pixels of the maximum image size that there is enough time for re-encoding, re-encoding is performed. May be determined based on the number of pixels. That is, from the time required for encoding and the time until the encoding of the image to be encoded next starts, it is determined whether there is enough time for re-encoding, that is, re-encoding is possible. The determination may be made, for example, every time the encoding unit ends the encoding.

An example in which it is determined whether there is enough time for re-encoding and re-encoding is performed regardless of the number of pixels will be described. In the following, a case where an image having the maximum image size that can be processed is encoded by the MPEG-2 method will be described using the encoding device in FIG.

First, an example of re-encoding an I (intra coded) picture will be described. In the MPEG system, I,
P (predictive coded), B (bidirectionally predic
tivecoded). When encoding an I-picture, it is not necessary to perform processes that require an extremely large amount of calculation, such as motion detection and motion compensation.
Therefore, the processing time required for encoding an I picture is shorter than that for encoding a P picture and a B picture. Therefore, it is possible to re-encode the I picture having the maximum image size, as in the case of an image smaller than the maximum image size.

The encoding unit 11 calculates the time required for encoding,
If it is determined that there is enough time for re-encoding from the time until the encoding of the next image to be encoded is started, and that re-encoding is possible, re-encoding is performed. Encoding unit 1
1 can make such a determination every time the encoding is completed, and repeat the re-encoding.

An I-picture is often intra-coded, so that the generated code amount is often large. Since the generated code amount often exceeds the threshold value determined by the generated code amount comparison unit 14,
The effect of performing re-encoding is also significant.

Next, an example of re-encoding when starting the encoding will be described. In the MPEG system, the order of encoding is different from the order of input of images, and thus it is necessary to rearrange the input images. Therefore, re-encoding is performed using the time required for the rearrangement.

FIG. 9 is an explanatory diagram of an example of re-encoding when encoding is started. In FIG. 9, one rectangle represents one image. As shown in FIG.
As shown in (a), images I0, B1, B2, P3, B4
Images are input in the order of. Here, I, B, and P represent the picture type of each image.

FIG. 9B shows the contents stored in the original image data storage unit 12. The original image data storage unit 12 has a memory for rearranging images. Here, as an example,
It is assumed that the original image data storage unit 12 includes four frames of memory. A state in which there is no data to be stored and the frame memory is empty is represented by E.

Since the B picture is coded after the coding of the rear reference picture, the coding order of the input picture in the order of FIG. 9A is as shown in FIG. 9C. I0, P
3, B1, B2,... In such a case, the image I0 to be encoded first is normally encoded only once in the period TSc immediately before the image P3 is encoded.

However, the image I0 already has the period TSc of 3
Since the original image data is input to the original image data storage unit 12 before the frame, encoding of the image I0 can be started in the period TSa two frames before the period TSc. The encoding of the image P3 to be encoded next may be started after the period TSc. Therefore, the encoding unit 11 performs the period TS
At the end of a and TSb, it is determined that there is enough time for re-encoding. In this case, as shown in FIG.
In b and TSc, the encoding unit 11 performs re-encoding on the image I0 twice.

In general, the quantization step is determined on the basis of a history of encoding of a past image. However, such a history cannot be used at the start of encoding, so that the quantization step cannot be set to an appropriate value. Therefore, in the first encoding of the image I0 in the period TSa, the quantization step is set to, for example, 1, which is the minimum value that can be set.

If the generated code amount at this time does not cause an underflow in the VBV buffer, the obtained encoded data is output. When the generated code amount is large and an underflow occurs in the VBV buffer, re-encoding is performed by increasing the quantization step, and the generated code amount can be reduced.

At the start of encoding, it is difficult to obtain appropriate image quality because information on past encoding cannot be used. However, according to this example, it is possible to re-encode an image that is encoded first at the start of encoding, even for an image having the maximum image size. Therefore, it is possible to improve the image quality of the image to be encoded first within a limited range of the generated code amount without causing an underflow in the VBV buffer.

Next, an example of re-encoding when the encoding is stopped will be described.

FIG. 10 is an explanatory diagram of an example of re-encoding when encoding is stopped. FIG. 10 shows a state after the image P6 in FIG. 9 is input. Here, as shown in FIG. 10 (a), the encoding is stopped for the input image P9, and after the input of the image is stopped for three frames, the image after the input image B10 is encoded. A case where coding is restarted for the target will be described.

FIG. 10B shows the contents stored in the original image data storage unit 12, as in FIG. 9B. FIG.
The order in which the images input in the order of FIG.
.., P9, B7, B8 as in 0 (c). In such a case, the last image to be encoded B8 is usually
Encoding is performed only once in the period TEa.

However, the encoding of the image I12 to be encoded next after the encoding is resumed may be started after the period TEd. Therefore, the encoding unit 11 determines that the periods TEa, T
At the end of Eb and TEc, it is determined that there is enough time for re-encoding. In this case, as shown in FIG. 10D, in the periods TEb, TEc, and TEd, the encoding unit 11
Performs re-encoding on the image B8 three times.

Also in this case, at the time of the first encoding of the image B8 in the period TEa, the quantization step is set to, for example, 1, which is the minimum value that can be set. If the generated code amount at this time does not cause an underflow in the VBV buffer, the obtained encoded data is output. When the generated code amount is large and an underflow occurs in the VBV buffer, re-encoding is performed by increasing the quantization step, and the generated code amount can be reduced.

When the coding is stopped, the last image to be coded does not need to be coded in real time. Therefore, as in this example, the last image to be coded when the coding is stopped must be re-coded. However, it is also possible for an image having the maximum image size. Therefore, it is possible to improve the image quality of the last encoded image within a limited range of the generated code amount without causing the VBV buffer to underflow.

In the above embodiment, the case where the image having the frame structure is encoded has been described. However, the same applies to the case where the image having the field structure is encoded.

In the above embodiment, the MPE
A case has been described in which an image is encoded by the MPEG-1 or MPEG-2 system using an encoding unit capable of performing compression encoding by the G-2 system. The encoding may be performed by the moving image encoding method using an encoding unit capable of performing the encoding.

Further, the quantization step is not limited to the one described in the above embodiment, and the quantization step used for re-encoding an image is the one used before the re-encoding of the image. Should be larger than

[0095]

As described above, according to the present invention, the amount of calculation required for encoding can be performed in real time by the encoding device, such as an image smaller than the largest image that can be processed by the encoding device. In the case where moving image encoding is performed on an image having a smaller amount than the maximum calculation amount, underflow does not occur in the VBV buffer. Therefore, since the generated coded data can be transmitted without being lost, an encoding device capable of encoding in real time is provided so that a frame is not lost when decoding is performed by a decoder. be able to.

[Brief description of the drawings]

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

FIG. 2 is a timing chart illustrating the operation of the encoding device in FIG. 1;

FIG. 3 is a flowchart showing the flow of processing when the encoding device in FIG. 1 encodes one picture.

FIG. 4 is a graph showing a temporal change of an occupation amount of a VBV buffer by encoded data.

FIG. 5 is a block diagram illustrating another example of the configuration of the encoding device in FIG. 1;

FIG. 6 is a block diagram illustrating a configuration of an encoding device according to a modification of the embodiment of the present invention.

7 is a timing chart illustrating an operation when the encoding device in FIG. 6 performs re-encoding of an image having the number of pixels M / N a plurality of times.

8 is a flowchart showing a flow of processing when the encoding device in FIG. 6 performs re-encoding a plurality of times to encode one picture.

FIG. 9 is an explanatory diagram of an example of re-encoding when encoding is started.

FIG. 10 is an explanatory diagram of an example of re-encoding when encoding is stopped.

[Explanation of symbols]

 11, 21, 31 Encoding unit 12 Original image data storage unit 13 Encoded data storage unit 14 Generated code amount comparison unit 15 Generated code amount calculation unit 16 Encoding count counter C (n + 1) Generated code amount of image to be encoded Bp (N + 1) Threshold for generated code amount

Claims (11)

[Claims] An encoding unit configured to perform compression encoding on a moving image and calculate a generated code amount of encoded data generated for the encoding target image; and an original image storing the encoding target image. A data storage unit, an encoded data storage unit for storing the encoded data, and a threshold value for the generated code amount is determined for each image to be encoded, and whether or not the generated code amount exceeds the threshold value And a generated code amount comparing unit that determines whether the image to be coded has sufficient time to re-code the image to be coded before coding of the image to be coded next starts. And the encoding unit determines, and when the generated code amount comparing unit determines that the generated code amount exceeds the threshold value, the encoding unit determines the generated code amount Change coding parameters to reduce And an encoding device that re-encodes the encoding target image. 2. The encoding apparatus according to claim 1, wherein the threshold value is a code amount determined so that a VBV (Video Buffering Verifier) buffer does not cause underflow. apparatus. 3. The encoding apparatus according to claim 1, wherein the encoding unit performs variable-length encoding on the moving image. 4. The encoding device according to claim 1, wherein the change of the encoding parameter increases a quantization step. 5. The encoding apparatus according to claim 1, wherein encoding of the moving image is performed in real time such that encoding of one image is completed within a time interval of an image interval of the moving image. An encoding device characterized by the following. 6. The encoding device according to claim 5, wherein when the re-encoding is possible, the number of pixels of the image to be encoded is equal to the maximum number of pixels of the image that can be processed by the encoding unit. An encoding apparatus characterized in that the number is less than half. 7. The encoding device according to claim 5, further comprising an encoding number counter that counts the number of encodings of the encoding target image by the encoding unit. In the case where the number of pixels of the image to be encoded is equal to or less than N (N is an integer of 3 or more) of the maximum number of pixels of the image that can be processed by the encoding unit, An encoding apparatus wherein the number of times of encoding of the encoding target image is less than N times. 8. The encoding apparatus according to claim 7, wherein the change of the encoding parameter is such that a quantization step is gradually increased at the time of encoding up to the (N−1) -th encoding of the encoding target image. An encoding device for maximizing a quantization step within an allowable range at the time of the N-th encoding of the encoding target image. 9. The encoding apparatus according to claim 1, wherein the encoding unit re-encodes an image to be encoded first when encoding is started. Device. 10. The encoding apparatus according to claim 1, wherein the encoding unit re-encodes an image to be encoded last when encoding is stopped. Device. 11. A step of performing compression encoding on a moving image, a step of calculating a generated code amount of coded data generated for an encoding target image, and a step of calculating the generated code amount for each encoding target image. And a step of determining whether or not the generated code amount exceeds the threshold value. The image to be coded is started before the coding of the next image to be coded starts. If there is enough time to re-encode, if it is possible to re-encode, and if it is determined that the generated code amount exceeds the threshold, the step of performing the compression-encoding comprises: An encoding method in which encoding parameters are changed so as to reduce the generated code amount, and the encoding target image is re-encoded.