JPH10164579A - Encoding control method and encoding controller using the method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an encoding control method and its device for improving an encoding speed and reducing an encoding quantity. SOLUTION: An encoding controller 20 is placed on the prestage of an encoder 2. This controller is provided with a judging pre-processing part 22, extracting the high-frequency component of an image by a high-pass filter, a judging part 30 finding the image of fewer high-frequency components and an instruction part 32 instructing a quantizing part 6 to reduce the found image. For the image with low importance, the number of quantizing levels are reduced for reducing a coded quantity.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an encoding control method and apparatus, and more particularly, to a method of controlling a code amount when encoding a picture and an apparatus using the method. This encoding control device can be used, for example, as a pre-processing device of an MPEG encoding device or incorporated in an MPEG encoding device.

[0002]

2. Description of the Related Art As represented by the international encoding standard MPEG, a technique of compressing and encoding data such as a moving image and storing and providing the data in various storage media has been developed.
Today, various manufacturers are focusing on the development of multimedia-related devices, and are trying to introduce MPEG-compliant products to the market.

FIG. 8 is a schematic block diagram of a conventional general MPEG encoding apparatus. As shown in the figure, the encoding device 2 performs discrete cosine transform (Discreet Cosine) on a video input.
DCT unit 4 that performs Transformation (hereinafter referred to as DCT)
And a quantizing unit 6 for quantizing the data after DCT, and receives the quantized data (hereinafter, referred to as quantized data), and performs variable-length coding on the data using a probability method with a code amount as small as possible. An encoding unit 8 and a FIFO for storing encoded data (hereinafter referred to as encoded data)
It has a transmission buffer 10 having a structure. Although not shown, there are other components related to motion compensation from the output of the quantization unit 6 to the input of the DCT unit 4.

In this configuration, first, a video input is decomposed into discrete spatial frequency components by a DCT unit 4 on a picture (frame) basis. Here, high-frequency components are dropped and quantized by the quantization unit 6 at a predetermined number of levels. The quantized data is encoded by the encoding unit 8 to generate a stream of encoded data specific to MPEG. This data stream is stored in the transmission buffer 10 and sequentially sent out as an encoded output to the outside.

[0005]

In MPEG coding, it can be considered that two processes of data compression and coding are roughly performed. The former is based on the DCT unit 4 and the quantization unit 6, and the latter is based on the encoding unit 8. However, the encoding unit 8 also has a data compression effect. That is, MPEG
By employing a method called entropy coding based on a probability method, the code amount of a P picture (inter-frame predictive coded picture) and a B picture (bidirectional predictive coded picture) described in the form of difference data Is considered to be as low as possible. Therefore, the code amount reduction effect of the entire encoding device is large.

However, recently, the development and commercialization of a device capable of encoding a moving image in real time have been attempted, and there is a strong demand for improving the encoding efficiency and further reducing the code amount. The present invention has been made in view of such a problem, and an object of the present invention is to provide an encoding control method and apparatus for increasing the encoding speed and reducing the code amount.

[0007]

An encoding control method according to the present invention is a method for controlling the encoding of a plurality of pictures, in which the importance of a picture is determined, and the amount of code for a picture having a low importance is determined. Is controlled so as to be smaller than the code amount to be achieved when not taking into account. Here, “importance” refers to the degree of importance from some viewpoint regardless of subjectivity or objectivity. "Picture" means MPEG
Is an arbitrary unit in video processing or display, such as a frame or a field, in addition to a picture in a narrow sense. The importance may be determined for each picture, a plurality of pictures, or a predetermined period.

For example, an MPEG encoding apparatus does not have the concept of the importance of a picture. Therefore, the same processing is performed regardless of the degree of importance, and a large code amount may be given to a less important picture. In the present invention, since the importance of a picture is determined in advance, the code amount for a picture with a low importance can be limited. Specific methods for reducing the code amount include a reduction in the number of quantization levels and a reduction in picture quality.

According to the present invention, the entire code amount can be reduced, and the speed of the entire encoding process can be increased by reducing the code amount. Note that the criteria for determining the importance include a frequency component of a picture, an audio-related signal accompanying the picture, a difference between pictures, and the like.

On the other hand, the encoding control device of the present invention is used as a pre-processing device of an encoding device such as an MPEG device, for example, as a device integrated with the encoding device. Here, the encoding device has a configuration including, for example, a transform unit that performs orthogonal transform such as DCT or other transform, a quantization unit, and an encoding unit, or a motion compensation unit in addition to the transform unit. A coding control device according to the present invention includes a judging means for judging importance of an input picture (30, 62, 86).
Instruction means (32, 42, 66, 8) for instructing a reduction in the code amount of a picture of low importance based on the determination result.
8, 104). In this configuration, when a picture is input, the judging means judges its importance. Subsequently, the instructing unit instructs to reduce the code amount of the low importance picture using the determination result. According to this device, the same effect as the above-described encoding control method can be realized as a device.

The instruction means may issue an instruction in the form of reducing the number of quantization levels of a picture. Further, the apparatus may include means (44) for changing the picture size according to the instruction of the instruction means. "Image" is a resolution determined by the number of pixels in the vertical and horizontal directions. In this case, the code amount is naturally reduced by lowering the picture quality of the picture of low importance.
The apparatus may further include a configuration for notifying the functional block that finally generates the encoded data stream (the encoding unit 8 in FIG. 8) that the image has been changed. In this case, it is easy to maintain the consistency of the entire apparatus.

[0012]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will be described below with reference to the drawings.

Embodiment 1 FIG. 1 shows a configuration of an encoding control device 20 of the present embodiment. As shown in the figure, the encoding control device 20 is placed before the MPEG encoding device 2, and determines a degree of importance of each picture based on a spatial frequency component of the picture. And an instruction unit 32 that outputs an instruction signal 34 to the quantization unit 6 of the encoding device 2 according to In addition, a pre-determination processing unit 22 is provided for preparing the determination by the determination unit 30. Note that the encoding device 2 itself is equivalent to that of FIG.

The pre-judgment processing unit 22 includes a horizontal high-pass filter (hereinafter abbreviated as HPF) 24 acting on horizontal pixels of a picture given in the form of a video input, and a vertical high-pass filter 24. It includes a vertical HPF 26 that operates and a horizontal and vertical HPF 28 that operates on pixels in both horizontal and vertical directions. These three H
The output of the PF is provided to the determination unit 30. In the present embodiment, high-frequency components of a picture are extracted by these HPFs, and a picture in which the component is dominant has a high importance.
If the opposite is the case, it is determined that the importance is low. In general, pictures with high frequency components have fine patterns,
This is because it is better to reduce the code amount too much. Here, three HPFs are provided, but the determination unit 30 may pay attention to only one of the outputs or all of them. Since it is considered that which HPF is referred to and the best result is obtained depends on the content of the picture, reference or non-reference may be set for each HPF.

The operation of the above configuration will be described. First, when there is a video input, the video input is subjected to a filtering process by the three HPFs of the pre-determination processing unit 22 for each picture. Since the output of the HPF is the high frequency component of the picture,
When there are many high-frequency components, that is, for example, when the absolute value average is large, the determination unit 30 regards the picture as having high importance. On the other hand, when the high-frequency component is small, the determination unit 30 regards the picture as having low importance, and notifies the instruction unit 32 that a picture with low importance has been detected.

The instructing unit 32 that has received the notification operates the encoding device 2
, Outputs an instruction signal 34 for instructing the quantization unit 6 to reduce the number of quantization levels. Upon receiving this instruction, the quantization unit 6 takes measures such as halving the number of quantization levels. As a result, the code amount of a picture of low importance is finally reduced.

The above is the configuration and operation of the encoding control device according to the present embodiment. This device not only has the effect of reducing the total code amount, but also can reduce the deterioration of the image quality of the entire moving image because pictures whose code amount is reduced are considered to be originally of low importance. Further, according to this device, the amount of quantized data in the encoding device 2 is reduced, so that the processing load on the encoding unit 8 of the encoding device 2 can be reduced, and as a result, the entire encoding process Speed up. Further, when the encoding device 2 complies with the MPEG, there is a configuration for changing the number of quantization levels in order to avoid overflow of the normal transmission buffer 10.
What should I send? Therefore, realization is extremely easy.

Although the encoding control device according to the present embodiment is placed before the encoding device, these two devices may be integrated as a matter of course. In the present embodiment, the importance is determined for each picture. However, the importance may be determined for a plurality of pictures, or may be determined for each predetermined period. For example, MPE
In the case of G, since there is a picture called an I picture to be closed and coded in the own picture for each GOP (described later), the importance determined for this I picture is calculated for all pictures included in the GOP. May be regarded as importance. On the other hand, conversely, the importance may be determined in units of a partial area of the picture. That is, the importance of the entire picture may be determined by looking at a partial area,
Fine code amount control may be performed, for example, by judging the importance for each area and changing the number of quantization levels for each area.

Embodiment 2 FIG . In the first embodiment, the entire code amount is reduced by changing the number of quantization levels. In the present embodiment, a similar effect is obtained by changing the image. FIG. 2 shows the configuration of the encoding control device 40 of the present embodiment. In this figure, the same members as those in FIG. 1 are denoted by the same reference numerals, and the description will focus on differences from FIG.

The new configuration in FIG. 2 includes an instruction unit 42 for issuing more instructions than the instruction unit 32 in FIG. 1 and a case change unit 44 for changing the image in accordance with the instruction from the instruction unit 42. is there. The picture changing unit 44 performs sub-sampling to reduce the resolution of the picture. The instruction unit 42 outputs a first instruction signal 46 for notifying that the image should be changed to the image change unit 44 and outputs a second instruction signal 48 to the encoding unit 8 of the encoding device 2. .

The operation of the above configuration will be described. First, when there is a video input, this is the horizontal HP
F24, the vertical HPF 26, and the horizontal / vertical HPF 28 undergo filtering processing. The determination unit 30 determines the number of high-frequency components from the outputs from these three HPFs, for example, by comparing with a predetermined threshold value. Here also, the importance is high when there are many high frequency components, and the importance is low when the opposite is the case.In the present embodiment, the importance of the picture is horizontally, vertically and horizontally and vertically independently in three directions. judge. The output of the determination unit 30 is specifically the following three types.

1. 1. The importance is low in the horizontal direction. 2. Low importance in vertical direction When the determination results are 1 to 3, the instruction unit 42 having a low degree of importance in the horizontal and vertical directions transmits the following instructions to the image quality changing unit 44 by the first instruction signal 46 corresponding to each of them.

(1) Decreasing the resolution in the horizontal direction (2) Decreasing the resolution in the vertical direction (3) Decreasing the resolution in the horizontal and vertical directions The image quality changing unit 44 performs sub-sampling in accordance with the instructions (1) to (3). That is, the picture quality is lowered by sampling pixels in a discrete manner. FIG. 3 shows how the picture size is changed. FIG. 3A shows that the picture size of the original picture is 720 horizontal pixels × 480 vertical pixels.
On the other hand, FIG. 6B shows a case where the horizontal pixels are reduced to half of the original 360 according to the above (1). Similarly, FIG. 11C shows a case where a vertical pixel is dropped to half of the original pixel, and FIG. 10D shows a case where the pixel is dropped to half of the original pixel in both directions. In the same figure, a pixel of ○ indicates a sample to be sampled, and a pixel of × indicates a sample not sampled. This completes the change of the image.

On the other hand, the instruction unit 42 sends a second instruction signal 48 to the encoding unit 8 of the encoding device 2 when the picture is to be changed. The second instruction signal 48 stores image information after the change, for example, data in the form of “360 × 480” in the case of FIG. 3B. The encoding unit 8 embeds the changed image information in the encoded data stream.

FIG. 4 shows the structure of an encoded data stream in MPEG. As shown in the figure, the data stream is composed of a sequence header (S
H) and a set of encoded data sets of a plurality of pictures called a group of pictures (GOP). In the sequence header, there are data areas called HSV and VSV, which store the resolution (number of pixels) in the horizontal direction and the vertical direction, respectively. Therefore, the encoding unit 8 extracts the picture information from the second instruction signal 48 and embeds the picture information in the HSV and VSV. As a result, the change of the model in the model changing unit 44 can be reflected in the data stream without contradiction.

As described above, according to the present embodiment, the total amount of code can be greatly reduced, and the processing load on the coding device 2 can be naturally reduced. The picture whose picture is changed is considered to be originally low in importance, and therefore has little effect on image quality.

Embodiment 3 Embodiments 1 and 2 both attempt to reduce the code amount by utilizing the properties of the picture itself. In the present embodiment, the code amount is reduced by focusing on the audio input accompanying the video input. The encoding control device according to the present embodiment is applied to television images.

FIG. 5 is a configuration diagram of the encoding control device 60 of the present embodiment. The apparatus includes a determination unit 62 that determines the importance of a picture based on a television sound-related signal, and an instruction unit that sends an instruction according to the determination result to the quantization unit 6 of the encoding device 2 as an instruction signal 68. It has 66. Judgment unit 62
Further has a control channel signal detection unit 64 for detecting a control channel signal from the audio input. The control channel signal is a signal that is continuously activated in the case of stereo broadcasting and multiplex audio broadcasting. By referring to this signal, it is possible to distinguish between normal audio broadcasting, stereo broadcasting, and multiplex audio broadcasting. Can be.

The operation of the above configuration will be described. When there is a video input to the encoding device 2, an audio input synchronized with the video input is given to the encoding control device 60. For example, if the television broadcast program is a normal audio broadcast and the commercial video is a stereo broadcast, the control channel signal detection unit 64 does not detect the control channel signal while the program video is being broadcast. , While the commercial video is being broadcast. Therefore, in the present embodiment, while the control channel signal is not detected, the importance of the picture is determined to be high, and while the control channel signal is detected, the importance is determined to be low, and the determination result is sent to the instruction unit 66.

The instruction unit 66 outputs an instruction signal 68 to reduce the number of quantization levels in the quantization unit 6 when it is determined that the importance of the picture is low. Subsequent processing is the same as in the first embodiment.

As described above, according to the present embodiment, the code amount can be reduced based on information other than video input.
In addition to the effect of 2, the valuation that can be realized by the encoding control device can be increased. An advantage of this embodiment is that the size of the device is relatively small.

Note that, as shown in FIG. 2 by adding the case change unit 44 to FIG. 1, the present embodiment may be modified to add a case change unit to FIG. 5 to provide a new embodiment.

Embodiment 4 FIG . In this embodiment, the code amount is reduced by focusing on the difference between pictures. FIG. 6 is a configuration diagram of the encoding control device 80 of the present embodiment. This device is
A frame memory 82 for temporarily storing a video input in units of pictures, a subtractor 84 for taking a difference between a picture currently being received and a previous picture stored in the frame memory 82, and an output of the differentiator 84. Has a determining unit 86 for determining the importance of a picture, and an instruction unit 88 for outputting an instruction signal 90 to the quantization unit 6 of the encoding device 2 according to the determination result.

In this configuration, when there is a video input, it is stored in the frame memory 82 for each picture on a picture-by-picture basis. Therefore, the previous picture always remains in the frame memory 82. For example, if the n-th picture has been input, the contents of the frame memory 82 are the (n-1) -th picture. Therefore, the difference unit 84 subtracts the (n−1) th picture data from the nth picture data, and inputs the difference to the determination unit 86.

The judging section 86 regards the importance of the n-th picture as being high when the difference is large, and as having low importance when the opposite is true. Therefore, when the sum of the differences is smaller than a certain threshold value, the instruction unit 90 outputs an instruction signal 90 to reduce the number of quantization levels. Subsequent processing is the same as in the first embodiment.

In this embodiment, the same effects as those of the other embodiments such as reduction of the code amount and reduction of the processing load can be obtained. The encoding control device according to the present embodiment is suitable for, for example, video processing of a surveillance camera. This is because it is only necessary to increase the code amount and obtain a fine image only when there is a change in the picture. In this embodiment, a new embodiment can also be provided by adding a scale changing unit.

Embodiment 5 A modified example based on Embodiment 2 will be described. In the present embodiment, a noise removal function is added to the encoding control device of the second embodiment.

FIG. 7 shows an encoding control device 100 according to this embodiment.
FIG. The new configuration for FIG.
From the three outputs of F24, vertical HPF 26 and horizontal / vertical HPF 28 and the total of four data of “0”, the indicating unit 10
The selection unit 102 selects and outputs one piece of data in accordance with the third instruction signal 106 output from the fourth unit 4. The difference unit 108 subtracts the output of the selection unit 102 from the currently input picture data. Note that the first instruction signal 46 and the second instruction signal 48 output by the instruction unit 104 are the same as in the second embodiment.

In the above configuration, if there is a video input,
The first instruction signal 46 and the second instruction signal 48 of the pre-determination processing unit 22, the determination unit 30, and the instruction unit 104 perform the same operation as in the second embodiment. If only the high frequency component in the horizontal direction is small, the first instruction signal 46 indicates “reduce the resolution in the horizontal direction”. At this time, the instruction unit 104
Indicates that the output of the horizontal HPF 24 is to be selectively output by the third instruction signal 106. The selection unit 102 follows this instruction.

Since the differentiator 108 subtracts the picture that has passed through the horizontal HPF 24 from the current picture, the output of the differentiator 108 has a form in which aliasing noise existing as a high-frequency component in the horizontal direction is reduced. According to the sampling theorem, a signal band that can be sampled at a certain frequency fs is limited to fs / 2 or less, so that aliasing noise occurs on the high frequency side from fs / 2. When sub-sampling video data containing aliasing noise at different frequencies,
The aliasing noise is superimposed on the originally required frequency component, and the image quality deteriorates. In the present embodiment, by removing unnecessary high-frequency components before sub-sampling, the adverse effect of aliasing noise can be significantly reduced.

It is to be noted that one input of the selection unit 102 is given "0" when the image is not changed, that is, when there are many high-frequency components in both the horizontal and vertical directions. This is because the picture data is not modified. Specifically, when the first instruction signal 46 does not instruct the change of the image, “0” is selectively output by the third instruction signal 106.

The above is the configuration and operation of this embodiment. According to the present embodiment, the image quality when changing the image quality can be improved as compared with the second embodiment. It is also effective in that the total data amount is reduced.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of an encoding control device according to a first embodiment.

FIG. 2 is a configuration diagram of an encoding control device according to a second embodiment.

FIGS. 3A to 3D are diagrams illustrating a state of sub-sampling performed by a resolution changing unit according to the second embodiment.

FIG. 4 is a diagram showing an MPEG encoded data stream.

FIG. 5 is a configuration diagram of an encoding control device according to a third embodiment.

FIG. 6 is a configuration diagram of an encoding control device according to a fourth embodiment.

FIG. 7 is a configuration diagram of an encoding control device according to a fifth embodiment.

FIG. 8 is a schematic configuration diagram of a conventional general MPEG encoding device.

[Explanation of symbols]

2 coding device, 4 DCT unit, 6 quantization unit, 8 coding unit, 10 transmission buffer, 20, 40, 60, 8
0,100 coding control device, 22 pre-determination processing unit, 2
4 Horizontal HPF, 26 Vertical HPF, 28 Horizontal / vertical H
PF, 30, 62, 86 judgment unit, 32, 42, 66,
88, 104 indicating unit, 34, 68, 90 indicating signal, 4
4 image change unit, 46 first instruction signal, 48 second instruction signal, 64 control channel signal detection unit, 82 frame memory, 84, 108 difference unit, 102 selection unit, 10
6 Third instruction signal.

Claims (7)

[Claims] 1. A method for controlling encoding of a plurality of pictures, wherein importance of a picture is determined, and a code amount for a picture having a low importance is smaller than a code amount to be achieved when the importance is not considered. An encoding control method, characterized in that the encoding is controlled so as to be reduced. 2. The encoding control method according to claim 1, wherein the importance of the picture is determined based on a frequency component of the picture. 3. The encoding control method according to claim 1, wherein the importance of the picture is determined based on an audio-related signal accompanying the picture. 4. The encoding control method according to claim 1, wherein importance of the picture is determined based on a difference between the pictures. 5. A determining means for determining importance of an input picture, and instructing means for instructing a reduction in a code amount of a picture with low importance based on the determination result. Encoding control device. 6. The encoding control device according to claim 5, wherein said instruction means instructs, when quantizing a picture, a reduction in the number of quantization levels. 7. The encoding control apparatus according to claim 5, further comprising: means for changing a picture size according to an instruction from said instruction means.