JP5094760B2 - Video encoding device - Google Patents

Description

  The present invention relates to a moving picture coding apparatus that divides each picture of a digital video signal into predetermined unit areas and performs compression coding for each unit area.

For example, MPEG (Moving Picture Experts Group) and ITU-T H.264. In an international standard video encoding scheme such as 26x, block data (16 × 16 pixels of luminance signal and color difference signals corresponding to 8 × 8 pixels corresponding to the luminance signal of 16 × 16 pixels in each frame of the video signal ( Hereinafter, a method of performing compression coding (compression coding based on motion search / compensation technology and orthogonal transform / transform coefficient quantization technology) is adopted using “macroblock” as a unit.
Even when decoding a bitstream that is compressed moving image data output from a moving image encoding apparatus that employs the above international standard video encoding method, decoding processing is performed in units of macroblocks, and finally all the images are Are decoded and output as a decoded image.

International standard video encoding system ITU-T H.264. In H.264, in order to improve the image quality of the decoded image, a method is adopted in which the quantization width of the luminance signal and the color difference signal can be set to different values.
That is, the international standard video encoding method ITU-T H.264. H.264 standardizes a method of encoding the quantization width of the color difference signal with a difference value from the quantization width of the luminance signal.

For example, Patent Document 1 below discloses a moving image encoding apparatus that determines the quantization width of a luminance signal and a color difference signal.
That is, in this moving image coding apparatus, the variance of the luminance signal and the color difference signal and the sum of absolute differences from the average value are calculated in units of pixels, so that the variation amount of the luminance signal is small and the variation amount of the color difference signal is large. In this case, the quantization width of the color difference signal is reduced.
Therefore, when encoding with an emphasis on color difference, the quantization width of the color difference signal can be reduced, so that the image quality can be improved.

JP 2007-116558 A (paragraph number [0014])

Since the conventional moving image encoding apparatus is configured as described above, when encoding with an emphasis on color difference, the quantization width of the color difference signal can be reduced to improve the image quality. However, when encoding is performed with emphasis on luminance, there is no means for improving the image quality, and there is a problem that the image quality cannot be improved.
In addition, pre-processing such as calculating the sum of absolute differences from the variance and average value of luminance signals and chrominance signals in units of pixels is necessary, so there is a problem that the amount of processing is large and unsuitable for real-time processing. It was.

  The present invention has been made in order to solve the above-described problems, and is not limited to encoding with emphasis on color difference, but is applicable to real-time processing not only when encoding with emphasis on luminance. It is an object of the present invention to obtain a moving image coding apparatus capable of improving the image quality with a processing amount.

The moving image encoding apparatus according to the present invention determines an offset value according to information indicating which of image quality and motion, which is encoding mode information included in encoding settings , is important. Determining means, and when the compression encoding means performs compression encoding for each unit area, the code amount to be assigned to the luminance signal and the color difference signal is determined according to the offset value determined by the offset value determination means. It is something that is set.

According to this invention, there is provided an offset value determining means for determining an offset value according to information indicating which of image quality and motion, which is encoding mode information included in the setting content of encoding, is important , When the compression encoding unit performs compression encoding for each unit region, the code amount allocated to the luminance signal and the color difference signal is set according to the offset value determined by the offset value determination unit. Therefore, not only when encoding with emphasis on color difference, but also when encoding with emphasis on luminance, the effect of improving image quality with a small amount of processing applicable to real-time processing There is.

It is a block diagram which shows the system which consists of a moving image encoding device and a moving image decoding device by Embodiment 1 of this invention. It is explanatory drawing which shows an example of the encoding setting information given to the moving image encoder 1. It is a block diagram which shows the moving image encoder by Embodiment 1 of this invention. It is a block diagram which shows the chroma QP offset determination part 11 of the moving image encoder by Embodiment 1 of this invention. It is explanatory drawing which shows an example of the threshold value process in the chroma QP offset calculation part 11b. It is a block diagram which shows the moving image encoder by Embodiment 2 of this invention. It is a block diagram which shows the chroma QP offset determination part 13 of the moving image encoder by Embodiment 2 of this invention. It is a block diagram which shows the moving image encoder by Embodiment 3 of this invention. It is a block diagram which shows the chroma QP offset determination part 14 of the moving image encoder by Embodiment 3 of this invention. It is a block diagram which shows the moving image encoder by Embodiment 4 of this invention. It is a block diagram which shows the chroma QP offset determination part 16 of the moving image encoder by Embodiment 4 of this invention. It is a block diagram which shows the moving image encoder by Embodiment 5 of this invention. It is a block diagram which shows the chroma QP offset determination part 17 of the moving image encoder by Embodiment 5 of this invention. It is a block diagram which shows the moving image encoder by Embodiment 6 of this invention. It is a block diagram which shows the chroma QP offset determination part 18 of the moving image encoder by Embodiment 6 of this invention. It is explanatory drawing which shows an example of the threshold value process of the macroblock average code amount in the chroma QP offset calculation part 18a. It is explanatory drawing which shows an example of the calculation process of the picture type dependence fluctuation range in the chroma QP offset calculation part 18a.

Embodiment 1 FIG.
FIG. 1 is a block diagram showing a system comprising a moving picture coding apparatus and a moving picture decoding apparatus according to Embodiment 1 of the present invention.
In FIG. 1, the moving image encoding apparatus 1 is, for example, H.264. H.264 / AVC encoding scheme is adopted, and when moving image data that is a digital video signal and encoding setting information indicating the setting content of encoding are input, each picture of the moving image data is converted into macroblock units. Dividing into (predetermined unit areas), performing compression encoding for each macroblock, and executing processing to output moving image compressed data that satisfies the encoding settings indicated by the encoding setting information as a bit stream To do.
The video decoding device 2 decodes the bit stream output from the video encoding device 1 in units of macroblocks, and finally decodes all macroblocks of one image, and then outputs display video data as a decoded image. Perform the process.

FIG. 2 is an explanatory diagram showing an example of the coding setting information given to the moving picture coding apparatus 1.
In the example of FIG. 2, the bit rate, frame rate, image resolution, coded picture type (I / P / B picture), coding mode information (whether emphasizing image quality or emphasizing motion are set as coding settings. ), GOP configuration, and picture type code amount ratio.
Such encoding setting information is generally well-known information.

FIG. 3 is a block diagram showing a moving picture coding apparatus 1 according to Embodiment 1 of the present invention.
In FIG. 3, the chroma QP offset determination unit 11 determines an offset value of the quantization width of the chrominance signal with respect to the quantization width of the luminance signal in the moving image data in accordance with the encoding setting content indicated by the encoding setting information. To implement.
Hereinafter, the offset value of the quantization width of the color difference signal with respect to the quantization width of the luminance signal is referred to as “chroma QP offset”.
“Chroma QP offset” is information indicating a color difference QP as a difference value from a luminance QP. It is “chroma_qp_index_offset” and “second_chroma_qp_index_offset” described in the H.264 / AVC coding standard. This value has a value in the range of “−12” to “+12”. It is defined in the H.264 / AVC encoding standard.
Here, “chroma” means a color difference, and “QP” means a quantization width.
The chroma QP offset determining unit 11 constitutes an offset value determining unit.

H. The H.264 / AVC encoding core unit 12 is an H.264 / AVC encoding core unit. H.264 / AVC encoding method is adopted, each picture of moving image data is divided into macroblock units, compression encoding is performed for each macroblock, and the encoding setting indicated by the encoding setting information is performed. A process of outputting a bit stream satisfying the contents is executed.
H. The H.264 / AVC encoding core unit 12 is a general H.264 standard. Similar to the H.264 / AVC encoder, a macroblock is used as a unit, and compression coding is performed based on intra pixel prediction / motion search / compensation technology and orthogonal transform / transform coefficient quantization / variable length coding technology. It is.
However, ITU-TH. H.264 uses a method that can set the quantization width of the luminance signal and the color difference signal to different values. The H.264 / AVC encoder encodes the chroma QP offset as “0”.
On the other hand, H.H. In the H.264 / AVC encoding core unit 12, when compression encoding is performed for each macroblock, codes assigned to luminance signals and color difference signals in moving image data according to the chroma QP offset determined by the chroma QP offset determination unit 11 Set the amount of each macroblock. H.264 / AVC encoding.
H. The H.264 / AVC encoding core unit 12 constitutes compression encoding means.

FIG. 4 is a block diagram showing the chroma QP offset determining unit 11 of the moving picture encoding apparatus 1 according to Embodiment 1 of the present invention.
In FIG. 4, the macroblock average code amount calculation unit 11a calculates the macroblock average code amount (the code amount per unit area) from the bit rate, the frame rate, and the image resolution included in the encoding setting content indicated by the encoding setting information. ) To calculate the macroblock average code amount avg_code_per_MB.

The chroma QP offset calculation unit 11b performs a process of calculating a chroma QP offset from the macroblock average code amount avg_code_per_MB calculated by the macroblock average code amount calculation unit 11a.
That is, the chroma QP offset calculation unit 11b performs threshold processing on the macroblock average code amount avg_code_per_MB, and when the macroblock average code amount avg_code_per_MB is small, calculates a large value of the chroma QP offset, When the amount avg_code_per_MB is large, a chroma QP offset having a small value is calculated.
FIG. 5 is an explanatory diagram showing an example of threshold processing in the chroma QP offset calculation unit 11b.

Next, the operation will be described.
When the encoding setting information is given from the outside, the chroma QP offset determining unit 11 determines the chroma QP offset according to the encoding setting content indicated by the encoding setting information.
That is, the macroblock average code amount calculation unit 11a of the chroma QP offset determination unit 11 performs the bit rate and frame included in the encoding setting content indicated by the encoding setting information as shown in the following equation (1). The macroblock average code amount avg_code_per_MB is calculated as an expected value of the macroblock average code amount from the rate and the image resolution.

avg_code_per_MB
= Bit_rate
/ ((Width / 16) × (height / 16) × frame_rate)
(1)
Bit_rate: Bit rate (bits / sec)
frame_rate: Frame rate (frames / sec)
width, height: image resolution (pixels)

Generally, only the encoded picture type has a value that changes in units of pictures among the encoding setting contents indicated by the encoding setting information.
For this reason, the calculation of equation (1) H.264 / AVC encoding core unit 12 is H.264. Before the start of H.264 / AVC encoding, when encoding setting information is given from the outside, it is sufficient to perform only once, and the amount of processing in the chroma QP offset calculation unit 11b is extremely small. It becomes.

  When the macroblock average code amount calculation unit 11a calculates the macroblock average code amount avg_code_per_MB, the chroma QP offset calculation unit 11b performs the threshold processing on the macroblock average code amount avg_code_per_MB, thereby obtaining the macroblock average code amount avg_code_per_MB. When the value is small, the chroma QP offset having a large value is calculated. When the macroblock average code amount avg_code_per_MB is large, the chroma QP offset having a small value is calculated.

For example, when five levels of threshold processing are performed, four threshold values Th1 to Th4 (Th <Th2 <Th3 <Th4) are set in advance, and as shown in FIG. 5, the macroblock average code amount avg_code_per_MB is a threshold value. If it is less than Th1, the chroma QP offset is determined to be “+12”.
If the macroblock average code amount avg_code_per_MB is equal to or greater than the threshold Th1, but less than the threshold Th2, the chroma QP offset is determined to be “+6”.
Similarly, if the macroblock average code amount avg_code_per_MB is equal to or greater than the threshold Th2, but if it is less than the threshold Th3, the chroma QP offset is determined to be “0” and the macroblock average code amount avg_code_per_MB is equal to or greater than the threshold Th3. If it is less than Th4, the chroma QP offset is determined to be “−6”.
If the macroblock average code amount avg_code_per_MB is greater than or equal to the threshold Th4, the chroma QP offset is determined to be “−12”.
In this example, the chroma QP offset calculation unit 11b performs the five-step threshold processing. However, this is only an example, and for example, more detailed 25-step threshold processing may be performed.

H. When the moving image data and the encoding setting information are input, the H.264 / AVC encoding core unit 12 divides each picture of the moving image data into macroblock units, performs compression encoding for each macroblock, A bit stream that satisfies the encoding setting content indicated by the encoding setting information is output.
However, H. The H.264 / AVC encoding core unit 12 is a code assigned to a luminance signal and a color difference signal in moving image data according to the chroma QP offset determined by the chroma QP offset determination unit 11 when performing compression encoding for each macroblock. Set the amount of each macroblock. H.264 / AVC encoding.

That is, H.C. When the chroma QP offset value determined by the chroma QP offset determination unit 11 is large (when the macroblock average code amount avg_code_per_MB is small), the H.264 / AVC encoding core unit 12 focuses on the small code amount given to luminance. Try to allocate. As a result, when the macroblock average code amount avg_code_per_MB is small, the luminance is emphasized, and as a result, the advantage of improving the subjective image quality is obtained.
On the other hand, when the value of the chroma QP offset determined by the chroma QP offset determination unit 11 is small (when the macroblock average code amount avg_code_per_MB is large), the given large code amount is sufficiently allocated to the color difference. As a result, when the macroblock average code amount avg_code_per_MB is large, a sufficient code amount is also assigned to the color difference signal, and as a result, an advantage of improving the subjective image quality is obtained.

  As apparent from the above, according to the first embodiment, the chroma QP offset determining unit 11 that determines the chroma QP offset according to the setting content of the encoding indicated by the encoding setting information is provided. When the H.264 / AVC encoding core unit 12 performs compression encoding for each macroblock, the code amount allocated to the luminance signal and the color difference signal in the moving image data according to the chroma QP offset determined by the chroma QP offset determination unit 11 Therefore, not only when encoding with emphasis on color difference but also when encoding with emphasis on luminance, image quality can be improved with a small amount of processing applicable to real-time processing. The effect which can aim at is produced.

  That is, according to the first embodiment, the macroblock average code amount calculation unit 11a calculates the macroblock average code amount avg_code_per_MB from the bit rate, the frame rate, and the image resolution, and the chroma QP offset calculation unit 11b calculates the macroblock average code amount. Since the chroma QP offset is calculated from the macroblock average code amount avg_code_per_MB calculated by the amount calculation unit 11a, Before starting the H.264 / AVC encoding, when the encoding setting information is given from the outside, the chroma QP offset needs to be calculated only once, and the luminance and color difference are balanced with a small amount of processing. There is an effect that can be.

Embodiment 2. FIG.
6 is a block diagram showing a moving picture encoding apparatus 1 according to Embodiment 2 of the present invention. In the figure, the same reference numerals as those in FIG.
Similar to the chroma QP offset determination unit 11 in FIG. 3, the chroma QP offset determination unit 13 performs a process of determining a chroma QP offset according to the encoding setting content indicated by the encoding setting information.
However, the chroma QP offset determining method is different from the chroma QP offset determining unit 11 of FIG.
The chroma QP offset determining unit 13 constitutes an offset value determining unit.

FIG. 7 is a block diagram showing the chroma QP offset determination unit 13 of the moving picture coding apparatus 1 according to Embodiment 2 of the present invention.
In FIG. 7, a chroma QP offset calculation unit 13a performs a process of calculating a chroma QP offset according to the picture type included in the setting content of encoding indicated by the encoding setting information.

Next, the operation will be described.
However, as compared with the first embodiment, since the chroma QP offset determining unit 13 in FIG. 3 is replaced by a chroma QP offset determining unit 13, the chroma QP offset determining unit 13 has the same configuration. Only the operation will be described.

When the coding setting information is given from the outside, the chroma QP offset calculating unit 13a of the chroma QP offset determining unit 13 responds to the coded picture type included in the coding setting content indicated by the coding setting information. A chroma QP offset is calculated.
Specifically, an example in which the chroma QP offset is determined as follows can be considered.
For example, when the encoded picture type is an I picture, the chroma QP offset is determined to be “0”.
When the encoded picture type is a P picture or a reference B picture, the chroma QP offset is determined to be “+6”, and when the encoded picture type is a non-reference B picture, the chroma QP offset is set to “+12”. To decide.

General H. In H.264 / AVC encoding, it is known that all encoded images can reach all I pictures if the reference source is traced.
In other words, the chroma QP offset is set to “0” in the I picture and the color difference signal is handled in the same manner as the luminance signal, and the chroma QP offset is set to “+6” or “+12” in the P picture or B picture. Even if the color difference signal is neglected, it can be taken over indirectly while referring to the color difference signal encoded by the I picture that is the reference source. As a result, the effect of improving the overall subjective image quality can be obtained. .

  For the same reason, the chroma QP offset is set to about “+6” for P pictures and reference B pictures that may be referred to, and the chroma QP offset is set to “+12” for non-reference B pictures that are not referenced. Thus, with respect to an image that can be referred to, the effect of improving the overall subjective image quality can be obtained by placing a little emphasis on the handling of the color difference signal.

  As is apparent from the above, according to the second embodiment, the chroma QP offset calculation unit 13a is configured to calculate the chroma QP offset according to the picture type. Thus, it is possible to determine the chroma QP offset, and as a result, an effect of improving the overall subjective image quality can be obtained.

Embodiment 3 FIG.
FIG. 8 is a block diagram showing a moving picture encoding apparatus 1 according to Embodiment 3 of the present invention. In the figure, the same reference numerals as those in FIG.
Similarly to the chroma QP offset determination unit 11 of FIG. 3, the chroma QP offset determination unit 14 performs a process of determining the chroma QP offset according to the encoding setting content indicated by the encoding setting information.
However, the chroma QP offset determining method is different from the chroma QP offset determining unit 11 of FIG.
The chroma QP offset determining unit 14 constitutes an offset value determining unit.

FIG. 9 is a block diagram showing the chroma QP offset determining unit 14 of the moving picture coding apparatus 1 according to Embodiment 3 of the present invention.
In FIG. 9, the chroma QP offset calculation unit 14a responds to encoding mode information (information indicating whether image quality is important or motion is important) included in the setting content of encoding indicated by the encoding setting information. A process of calculating a chroma QP offset is performed.

Next, the operation will be described.
However, as compared with the first embodiment, since the chroma QP offset determining unit 14 in FIG. 3 is replaced by a chroma QP offset determining unit 14, the chroma QP offset determining unit 14 has the same configuration. Only the operation will be described.

When the coding setting information is given from the outside, the chroma QP offset calculating unit 14a of the chroma QP offset determining unit 14 responds to the coding mode information included in the coding setting content indicated by the coding setting information. A chroma QP offset is calculated.
Specifically, an example in which the chroma QP offset is determined as follows can be considered.
For example, if the encoding mode information indicates that image quality is important, the chroma QP offset is determined to be “0”, and if the encoding mode information indicates that motion is important, the chroma QP offset is determined. Determine “+12”.

General H. In H.264 / AVC encoding, particularly when the bit rate value included in the encoding setting content indicated by the encoding setting information is small, it is difficult to perform encoding so as to satisfy the bit rate. There is often.
In such a case, when maintaining a small bit rate by performing frame dropping while maintaining the image quality, the coding setting information including the coding mode information indicating that “image quality is emphasized” is the moving image code. On the other hand, when the frame rate is maintained even when the image quality is lowered, the encoding setting information including the encoding mode information indicating “motion-oriented” is given to the moving image encoding device. Generally, it is given to 1.

  Therefore, the chroma QP offset calculation unit 14a maintains the image quality as much as possible by making the handling of the luminance and the color difference equal when the image quality is important. In addition, in the case of emphasizing motion, even if the color difference is neglected to the utmost, frames are prevented from dropping as much as possible, and as a result, encoding is performed in a manner that complies with the desired encoding setting information as much as possible. As a result, the effect of improving the overall subjective image quality can be obtained.

  In the third embodiment, the encoding mode information is information indicating whether the image quality is important or the motion is important. However, the encoding mode information is not two stages of “image quality importance” and “motion importance”. For example, the chroma QP offset may be determined finely by inputting coding mode information divided into a number of stages, including “slightly emphasizing image quality” and “slightly emphasizing motion”.

  As apparent from the above, according to the third embodiment, the chroma QP offset calculation unit 14a sets the chroma QP offset according to the coding mode information (information indicating whether the image quality is important or the motion is important). Since it is configured to calculate, it is possible to determine the chroma QP offset so that the color difference is neglected as the mode in which the movement is emphasized, and as a result, the mode that follows the desired encoding setting information as much as possible. Therefore, the effect of improving the overall subjective image quality can be obtained.

Embodiment 4 FIG.
10 is a block diagram showing a moving picture encoding apparatus 1 according to Embodiment 4 of the present invention. In the figure, the same reference numerals as those in FIG.
The scene change detection unit 15 monitors the moving image data, detects a scene change, and executes a process of outputting scene change information indicating the presence or absence of the scene change.
The chroma QP offset determining unit 16 performs a process of determining a chroma QP offset according to the encoding setting content indicated by the encoding setting information and the scene change information output from the scene change detecting unit 15.
The chroma QP offset determining unit 16 constitutes an offset value determining unit.

FIG. 11 is a block diagram showing the chroma QP offset determining unit 16 of the moving picture coding apparatus 1 according to Embodiment 4 of the present invention.
In FIG. 11, the chroma QP offset calculation unit 16 a performs chroma QP according to the picture type included in the setting content of encoding indicated by the encoding setting information and the scene change information output from the scene change detection unit 15. A process for calculating the offset is performed.

Next, the operation will be described.
However, as compared with the first embodiment, a chroma QP offset determining unit 16 is provided instead of the chroma QP offset determining unit 11 in FIG. 3, and a scene change detecting unit 15 is provided outside the moving image encoding device 1. It differs in that it is provided.

The scene change detection unit 15 monitors the moving image data provided to the moving image encoding device 1 to detect a scene change, and outputs scene change information indicating the presence or absence of a scene change to the chroma QP offset determination unit 16.
However, since the scene change detection process is a known technique, a detailed description thereof is omitted here.

When the chroma QP offset calculation unit 16a of the chroma QP offset determination unit 16 receives the scene change information from the scene change detection unit 15, the chroma QP offset calculation unit 16a calculates a chroma QP offset according to the scene change information and the encoded picture type.
Specifically, an example in which the chroma QP offset is determined as follows can be considered.
For example, if the scene change information output from the scene change detection unit 15 indicates “with scene change” and the encoded picture type is “other than I picture”, the chroma QP offset is determined to be “+12”, If the scene change information indicates “no scene change” or the encoded picture type is “I picture”, the chroma QP offset is determined to be “0”.

General H. In H.264 / AVC coding, an I picture has no reference image and is an important image that is indirectly referenced from all images even though the compression efficiency is low. Usually, a considerably larger amount of code is assigned than B picture.
In general, switching between an I picture, a P picture, and a B picture is often determined only by the order of images, for example, IBBPBBPBBPBBPBBIBBBPBBPBBPBBPBBIBBPBB.
However, since it is generally unknown where the scene change is, the possibility that the scene change happens to be an I picture is extremely low.
When a scene change occurs in a P picture or a B picture, a code amount that is suddenly larger than an assumed assigned code amount is generally generated, so that frame dropping for maintaining a bit rate is likely to occur thereafter.
In order to prevent abrupt code generation due to such a scene change, the chroma QP offset may be output as “+12” when “there is a scene change” and “the picture type is other than I picture”. .

  Therefore, as described above, the chroma QP offset calculation unit 16a of the chroma QP offset determination unit 16 indicates that the scene change information indicates “with scene change” and the encoded picture type is “other than I picture”. By determining the chroma QP offset to “+12”, it is possible to suppress a sudden large amount of code generation when a scene change occurs in a picture type with a small allocated code amount such as a P picture or B picture. , I try to prevent dropped frames.

  As is apparent from the above, according to the fourth embodiment, the chroma QP offset calculation unit 16a outputs the picture type included in the encoding setting content indicated by the encoding setting information and the scene change detection unit 15. Since the chroma QP offset is calculated in accordance with the scene change information to be changed, the sudden code amount when a scene change occurs in a picture type with a small assigned code amount such as a P picture or B picture. It is possible to suppress a large number of occurrences and prevent frame dropping, and as a result, encoding is performed in a manner that follows the desired encoding setting information as much as possible. An effect that leads to an improvement in image quality can be obtained.

Embodiment 5 FIG.
FIG. 12 is a block diagram showing a moving picture encoding apparatus 1 according to Embodiment 5 of the present invention. In the figure, the same reference numerals as those in FIG.
The chroma QP offset determination unit 17 calculates the expected value of the code amount of the picture type included in the setting content from the bit rate and the frame rate included in the setting content of the encoding indicated by the encoding setting information. , The expected value of the code amount and H.264. A process for determining a chroma QP offset is performed according to the code amount of the bit stream (moving image compressed data of the previous picture) output from the H.264 / AVC encoding core unit 12.
The chroma QP offset determining unit 17 constitutes an offset value determining unit.

FIG. 13 is a block diagram showing the chroma QP offset determining unit 17 of the moving picture coding apparatus 1 according to Embodiment 5 of the present invention.
In FIG. 13, the picture average code amount calculation unit 17a determines the code amount of the picture type included in the setting content from the bit rate and the frame rate included in the setting content of encoding indicated by the encoding setting information. A process of calculating the average picture code amount as the expected value is performed.
The chroma QP offset calculation unit 17b is configured to calculate the picture average code amount calculated by the picture average code amount calculation unit 17a, H.264 / AVC encoding core unit 12 is compared with the code amount of the bit stream (moving image compressed data of the previous picture) to calculate a chroma QP offset.

Next, the operation will be described.
However, as compared with the first embodiment, since the chroma QP offset determining unit 17 is the same as the chroma QP offset determining unit 17 in FIG. 3 except that a chroma QP offset determining unit 17 is provided, the chroma QP offset determining unit 17 Only the operation will be described.

  When the picture average code amount calculation unit 17a of the chroma QP offset determination unit 17 is provided with encoding setting information from the outside, as shown in the following equation (2), the setting contents of encoding indicated by the encoding setting information As the expected value of the code amount of the picture type (any one of I, P, and B) included in the setting content from the bit rate and frame rate included in the picture, the picture average code amount avg_code_per_PIC_X ( X = I, P, or B) is calculated.

Avg_code_per_PIC_B
= ((1 + np + nb) × bit_rate)
/ ((Wi + np · wp + nb) × frame_rate)
Avg_code_per_PIC_P
= Avg_code_per_PIC_B × wp
Avg_code_per_PIC_I
= Avg_code_per_PIC_B × wi
(2)
Bit_rate: Bit rate (bits / sec)
frame_rate: Frame rate (frames / sec)
np: Number of P pictures included in 1 GOP
nb: Number of B pictures included in 1 GOP
wi, wp: Expected picture code amount ratio of each picture of I / P / B

  For this reason, since the calculation of Expression (2) is sufficient for one picture encoding, the processing amount of calculation in the picture average code amount calculation unit 17a is extremely small.

When the picture average code amount calculation unit 17a calculates the picture average code amount avg_code_per_PIC_X, the chroma QP offset calculation unit 17b inputs the picture average code amount avg_code_per_PIC_X and the bitstream every time one picture is encoded, and the picture average code The chroma QP offset is calculated by comparing the amount avg_code_per_PIC_X with the bitstream code amount CV.
For example, when the bitstream code amount CV is smaller than the picture average code amount avg_code_per_PIC_X by a certain value α1 or more, the chroma QP offset value is determined to be smaller than the chroma QP offset of the current picture.
CV + α1 <avg_code_per_PIC_X
→ Decrease the value of chroma QP offset

Conversely, if the bitstream code amount CV is larger than the picture average code amount avg_code_per_PIC_X by a certain value α2 or more, the chroma QP offset value is determined to be larger than the chroma QP offset of the current picture.
CV> avg_code_per_PIC_X + α2
→ Increase the value of the chroma QP offset Note that the bitstream code amount CV is not smaller than the picture average code amount avg_code_per_PIC_X by a constant value α1 or more, and the bitstream code amount CV is larger than the picture average code amount avg_code_per_PIC_X by a constant value α2 or more. If not, the value of the chroma QP offset is determined to be the same as the chroma QP offset of the current picture.
However, for the chroma QP offset of the first encoded picture, for example, “0” is set as an initial value, and the chroma QP offset of the next encoded picture is determined by comparison as described above.

General H. In H.264 / AVC encoding, it is normal that the chroma QP offset is uniformly “0” or the chroma QP offset is constant regardless of the actually generated code amount.
On the other hand, in the fifth embodiment, for example, when the actual generated code amount is small, the chroma QP offset value is set to be smaller than the chroma QP offset of the current picture in view of the margin of the code amount. Therefore, the color difference code amount of the next picture is increased. As a result, the expected value of the generated code amount of the next picture comes closer to the picture average code amount.
On the contrary, when the actual generated code amount is large, the chroma QP offset value is made larger than the chroma QP offset of the current picture in view of the fact that there is no margin in the code amount. Will be reduced. As a result, the expected value of the generated code amount of the next picture comes closer to the picture average code amount.

  As is apparent from the above, according to the fifth embodiment, the picture type included in the setting content is determined from the bit rate and the frame rate included in the setting content of the encoding indicated by the encoding setting information. Picture average code amount avg_code_per_PIC_X is calculated, and the picture average code amount avg_code_per_PIC_X and the H.H. Since the chroma QP offset is determined according to the code amount CV of the bitstream (moving image compressed data of the previous picture) output from the H.264 / AVC encoding core unit 12, the next picture The expected value of the generated code amount approaches the picture average code amount, and as a result, encoding is performed with the code amount as close as possible to the desired encoding setting information, so that the overall subjective image quality is improved. The effect that leads to is obtained.

Embodiment 6 FIG.
FIG. 14 is a block diagram showing a moving picture encoding apparatus 1 according to Embodiment 6 of the present invention. In the figure, the same reference numerals as those in FIGS.
The chroma QP offset determining unit 18 calculates the macroblock average code amount avg_code_per_MB from the bit rate, the frame rate, and the image resolution included in the encoding setting content indicated by the encoding setting information, and is included in the setting content. The picture average code amount avg_code_per_PIC_X of the picture type included in the setting content is calculated from the bit rate, the frame rate, etc., the macro block average code amount avg_code_per_MB, the picture average code amount avg_code_per_PIC_X, included in the setting content Encoding mode information (information indicating whether the image quality is important or the motion is important), the scene change information output from the scene change detection unit 15, and the H.264 encoding information. A process for determining a chroma QP offset is performed according to the code amount of the bit stream (moving image compressed data of the previous picture) output from the H.264 / AVC encoding core unit 12.
The chroma QP offset determining unit 18 constitutes an offset value determining unit.

FIG. 15 is a block diagram showing the chroma QP offset determining unit 18 of the moving picture coding apparatus 1 according to Embodiment 6 of the present invention. In the figure, the same reference numerals as those in FIGS. 4 and 13 indicate the same or corresponding parts. Description is omitted.
The chroma QP offset calculation unit 18a includes a picture average code amount avg_code_per_MB calculated by the macroblock average code amount calculation unit 11a, a picture average code amount avg_code_per_PIC_X calculated by the picture average code amount calculation unit 17a, and coding mode information (in consideration of image quality). Information indicating whether or not there is an emphasis on motion), the scene change information output from the scene change detection unit 15, and H.264. A process of calculating a chroma QP offset is performed according to the code amount of the bit stream (moving image compressed data of the previous picture) output from the H.264 / AVC encoding core unit 12.

In the sixth embodiment, the general H.264 standard is used. As in the case of H.264 / AVC encoding, description will be made assuming that only the encoded picture type has a value that changes in units of pictures in the encoding setting information.
The processing of the chroma QP offset calculation unit 18a in the chroma QP offset determination unit 18 includes the following steps 1 to 3.
Step 1: Calculation processing of base value CQPOFF1 of chroma QP offset Step 2: Calculation of picture type-dependent variation width CQPOFF2 of chroma QP offset
Processing Step 3: Calculation processing of picture type update width CQPOFF3 of chroma QP offset

Hereinafter, the processing content of steps 1-3 will be described in detail.
[Step 1]
The process of step 1 is executed only once when the encoding setting information is given from the outside.
First, the chroma QP offset calculation unit 18a initializes the picture type update width CQPOFF3 of the chroma QP offset to “0”.
Next, the chroma QP offset calculation unit 18a performs threshold processing on the picture average code amount avg_code_per_MB calculated by the macroblock average code amount calculation unit 11a, so that the macroblock average code amount avg_code_per_MB is large. The base value CQPOFF1 of the chroma QP offset of the value is calculated, and when the macroblock average code amount avg_code_per_MB is large, the base value CQPOFF1 of the chroma QP offset of the small value is calculated.
However, the chroma QP offset calculation unit 18a calculates the base value CQPOFF1 of the chroma QP offset in consideration of the coding mode information (image quality emphasis / motion emphasis) in order not to neglect the color difference when the image quality is important. To do.

FIG. 16 is an explanatory diagram showing an example of threshold processing for the average macroblock code amount in the chroma QP offset calculation unit 18a.
In the example of FIG. 16, when the image quality is emphasized, if the macroblock average code amount avg_code_per_MB is less than the threshold Th3, the chroma QP offset base value CQPOFF1 is determined to be “0”.
Further, if the macroblock average code amount avg_code_per_MB is equal to or greater than the threshold Th3, but less than the threshold Th4, the base value CQPOFF1 of the chroma QP offset is determined to be “−6”, and the macroblock average code amount avg_code_per_MB is equal to or greater than the threshold Th4. If there is, the base value CQPOFF1 of the chroma QP offset is determined to be “−12”.

In the case of emphasizing motion, if the macroblock average code amount avg_code_per_MB is less than the threshold Th1, the base value CQPOFF1 of the chroma QP offset is determined to be “+12”.
If the macroblock average code amount avg_code_per_MB is equal to or greater than the threshold value Th1, but less than the threshold value Th2, the base value CQPOFF1 of the chroma QP offset is determined to be “+6”.
Similarly, if the macroblock average code amount avg_code_per_MB is equal to or greater than the threshold Th2, but less than the threshold Th3, the base value CQPOFF1 of the chroma QP offset is determined to be “0”, and the macroblock average code amount avg_code_per_MB is equal to or greater than the threshold Th3. If it is less than the threshold Th4, the base value CQPOFF1 of the chroma QP offset is determined to be “−6”.
If the macroblock average code amount avg_code_per_MB is equal to or greater than the threshold Th4, the base value CQPOFF1 of the chroma QP offset is determined to be “−12”.

In this example, the chroma QP offset calculation unit 18a performs the five-step threshold processing. However, this is only an example, and for example, more detailed 25-step threshold processing may be performed.
In addition, although the encoding mode information is information indicating whether the image quality is important or the motion is important, the encoding mode information is not two stages of “image quality importance” and “motion importance”, but, for example, “slight image quality” The base value CQPOFF1 of the chroma QP offset may be determined finely by inputting encoding mode information divided into a number of stages including “important” and “slightly important”.

[Step 2]
The process of step 2 is executed once after the process of step 1 and before the encoding of each picture.
The chroma QP offset calculation unit 18a determines the picture type of the chroma QP offset according to the scene change information output from the scene change detection unit 15, the encoded picture type, and the encoding mode information (image quality priority / motion priority). The dependence fluctuation range CQPOFF2 is calculated.
FIG. 17 is an explanatory diagram showing an example of a calculation process of the picture type dependent variation width CQPOFF2 in the chroma QP offset calculation unit 18a.

In the example of FIG. 17, the picture type-dependent variation width CQPOFF2 is calculated by dividing the table into the case where there is no scene change and the case where there is a scene change.
In the case of no scene change, if motion is emphasized, the concept is the same as in the second embodiment. If the encoded picture type is an I picture, the picture type-dependent variation width CQPOFF2 is “0”, and the encoded picture type is If is a P picture or a reference B picture, the picture type dependent variation width CQPOFF2 is determined to be “+6”, and if the encoded picture type is a non-reference B picture, the picture type dependent variation width CQPOFF2 is determined to be “+12”.
If the image quality is important, the picture type dependent variation width CQPOFF2 is uniformly determined to be “0” regardless of the encoded picture type so as not to neglect the color difference.

When there is a scene change, if motion is important, the concept is the same as in the fourth embodiment. If the encoded picture type is an I picture, the picture type-dependent variation width CQPOFF2 is “0”, and the encoded picture type is If it is a P picture, a reference B picture or a non-reference B picture, the picture type-dependent variation width CQPOFF2 is determined to be “+12”.
If the image quality is important, even if it is a scene change, in order to minimize the variation of the chroma QP offset depending on the picture type and to avoid neglecting the color difference, if it is an I picture, P picture or reference B picture, the picture type If the dependent variation width CQPOFF2 is “0” and the encoded picture type is a non-reference B picture, the picture type dependent variation width CQPOFF2 is determined to be “+6”.

  Here, the encoding mode information is information indicating whether the image quality is important or the motion is important. However, the encoding mode information is not in two stages of “image quality importance” and “motion importance”. The picture type dependent variation width CQPOFF2 may be determined in detail by inputting coding mode information divided into a number of stages including “emphasis on image quality” and “a little emphasis on motion”.

  When the chroma QP offset calculation unit 18a calculates the picture type-dependent variation width CQPOFF2 as described above, the value of the base value CQPOFF1 of the chroma QP offset + the picture type-dependent variation width CQPOFF2 + the picture type update width CQPOFF3 is determined from “−12”. Clip to a value between “+12”, and use the clipped value as a chroma QP offset. H.264 / AVC encoding core unit 12 for output.

[Step 3]
The process of step 3 is executed once after the process of step 2 (after encoding of each picture), and then the process returns to step 2.
The chroma QP offset calculation unit 18a includes a picture average code amount avg_code_per_PIC_X calculated by the picture average code amount calculation unit 17a, The picture type update width CQPOFF3 of the chroma QP offset is determined according to the code amount of the bit stream (moving image compressed data of the previous picture) output from the H.264 / AVC encoding core unit 12.

A specific example of a method for determining the picture type update width CQPOFF3 of the chroma QP offset from the picture average code amount avg_code_per_PIC_X and the bitstream code amount is as follows.
For example, when the bitstream code amount CV is smaller than the picture average code amount avg_code_per_PIC_X by a certain value α3 or more, the picture type update width CQPOFF3 is determined to be smaller than the picture type update width CQPOFF3 of the current picture.
CV + α3 <avg_code_per_PIC_X
→ CQPOFF3 = CQPOFF3-1

Conversely, when the bitstream code amount CV is larger than the picture average code amount avg_code_per_PIC_X by a certain value α4 or more, the picture type update width CQPOFF3 is determined to be larger than the picture type update width CQPOFF3 of the current picture.
CV> avg_code_per_PIC_X + α4
→ CQPOFF3 = CQPOFF3 + 1
If the bitstream code amount CV is not smaller than the picture average code amount avg_code_per_PIC_X by a certain value α3 or more, and the bitstream code amount CV is not larger than the picture average code amount avg_code_per_PIC_X by a certain value α4 or more, the picture type update width CQPOFF3 is determined to be the same value as the picture type update width CQPOFF3 of the current picture.

  As is apparent from the above, according to the sixth embodiment, the macro QP offset determination unit 18 determines the macroblock based on the bit rate, frame rate, and image resolution included in the encoding setting content indicated by the encoding setting information. The average code amount avg_code_per_MB is calculated, the picture average code amount avg_code_per_PIC_X of the picture type included in the setting content is calculated from the bit rate and the frame rate included in the setting content, and the macroblock average code The amount avg_code_per_MB, the picture average code amount avg_code_per_PIC_X, the coding mode information included in the setting contents (information indicating whether the image quality is important or the motion is important), output from the scene change detection unit 15 Scene change information and H. Since the chroma QP offset is determined according to the code amount of the bit stream output from the H.264 / AVC encoding core unit 12, the same effects as those obtained in the first to fifth embodiments can be obtained. As a result, the effect of improving the overall subjective image quality can be obtained.

  In the first to sixth embodiments, the video encoding method is H.264. H.264 / AVC has been shown. An encoding method similar to H.264 / AVC (for example, MPEG-2, MPEG-4 Visual, SMPTE VC-1, etc.) may be used, and similar effects can be obtained.

  DESCRIPTION OF SYMBOLS 1 moving image encoder, 2 moving image decoder, 11 chroma QP offset determination part (offset value determination means), 11a macroblock average code amount calculation part, 11b chroma QP offset calculation part, 12H. H.264 / AVC encoding core unit (compression encoding unit), 13 chroma QP offset determination unit (offset value determination unit), 13a chroma QP offset calculation unit, 14 chroma QP offset determination unit (offset value determination unit), 14a chroma QP offset Calculation unit, 15 Scene change detection unit, 16 Chroma QP offset determination unit (offset value determination unit), 16a Chroma QP offset calculation unit, 17 Chroma QP offset determination unit (offset value determination unit), 17a Picture average code amount calculation unit, 17b Chroma QP offset calculator, 18 Chroma QP offset determiner (offset value determiner), 18a Chroma QP offset calculator.

Claims (3)

Compression coding means that divides each picture of a digital video signal into predetermined unit areas, performs compression coding for each unit area, and outputs a bitstream that satisfies the setting contents of coding given from the outside In the moving image encoding apparatus, an offset value determining means for determining an offset value according to information indicating which of image quality and motion, which is encoding mode information included in the setting content of the encoding, is important When the compression encoding unit performs compression encoding for each unit area, the code amount allocated to the luminance signal and the color difference signal is set according to the offset value determined by the offset value determination unit. A moving picture coding apparatus characterized by the above. Compression coding means that divides each picture of a digital video signal into predetermined unit areas, performs compression coding for each unit area, and outputs a bitstream that satisfies the setting contents of coding given from the outside In the moving image encoding apparatus, an offset value determining means for determining an offset value according to a picture type included in the setting content of the encoding and scene change information indicating presence / absence of a scene change given from outside And when the compression coding means performs compression coding for each unit area, setting a code amount to be assigned to the luminance signal and the color difference signal according to the offset value determined by the offset value determination means A video encoding apparatus characterized by the above. Compression coding means that divides each picture of a digital video signal into predetermined unit areas, performs compression coding for each unit area, and outputs a bitstream that satisfies the setting contents of coding given from the outside In the moving image encoding apparatus, the expected value of the code amount per unit area is calculated from the bit rate, the frame rate, and the resolution included in the setting content of the encoding, and the bit included in the setting content The expected value of the code amount of the picture type included in the setting content is calculated from the rate and the frame rate, the expected value of the code amount per unit area, the expected value of the code amount of the picture type, and the setting content Information indicating whether image quality or motion, which is encoding mode information included in the image, is important, and whether there is a scene change given from the outside Nchenji information and in accordance with the code amount of the output bit stream from the compression encoding means, the provided offset value determining means for determining the offset value, when the compression coding means for carrying out compression coding for each unit area, A moving picture encoding apparatus, wherein a code amount to be assigned to the luminance signal and the color difference signal is set according to the offset value determined by the offset value determining means.

Images