JP2004015351A - Encoding apparatus and method, program, and recording medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To encode a signal in a bit amount adjusted in response to a scene change. <P>SOLUTION: An encoder is provided with: a scene change position detection section 12 for detecting the position of a scene change from a moving picture signal; a degree of encoding difficulty calculation section 10 for calculating the degree of encoding difficulty d for each unit time of the moving picture signal; a bit amount calculation section 11 for calculating a bit amount b required for transfer of the moving picture signal within the unit time on the basis of the degree of encoding difficulty d; a bit amount adjustment section for adjusting the bit amount assigned before and after a prescribed position decided on the basis of the scene change position when the bit amount b is lower than a prescribed threshold value; a delay section 13 for delaying the moving picture signal; and an encoding section 14 for revising part of image structure into a prescribed image structure on the basis of the scene change position and encoding the moving picture signal outputted from the delay section on the basis of the adjustment performed by the bit amount adjustment section according to the image structure after the revision. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an encoding apparatus and method, a program, and a recording medium in high-efficiency encoding for efficiently encoding digital signals, and more particularly to a one-pass system for encoding moving picture signals. The present invention relates to an encoding apparatus and method, a program, and a recording medium that perform encoding by controlling the allocation amount of a variable bit rate.
[0002]
2. Description of the Related Art Since a digital video signal has a very large amount of data, a high-efficiency encoding means for efficiently encoding a video signal at a high compression rate when it is desired to record the digital video signal on a small recording medium having a small storage capacity for a long time. Becomes indispensable. In order to meet such demands, high-efficiency coding methods using correlation of video signals have been proposed, one of which is the MPEG method.
[0003]
The MPEG (Moving Picture Image Coding Experts Group) is discussed in ISO-IEC / JTC1 / SC29 / WG11 and proposed as a standard, and includes motion compensation predictive coding and discrete cosine transform (DCT: Discrete). This is a hybrid system that combines Cosine Transform (encoding) and coding. In this MPEG system, the redundancy in the time axis direction is first reduced by taking the difference between frames of the video signal, and then the redundancy in the spatial axis direction is reduced by using the discrete cosine transform. Encode efficiently. In addition, in the MPEG system, in order to perform the above-described motion compensation prediction encoding, a GOP including three elements of an I picture (Intra encoded image), a P picture (Predictive encoded image), and a B picture (Bidirectionally Predictive encoded image). (Group of Picture) structure.
[0004]
Generally, a video signal is not stationary, and the information amount of each picture changes with time. Therefore, it is known that higher image quality can be obtained by using variable bit rate coding than by constant bit rate coding with the same code amount.
[0005]
For example, for a video signal recorded on a so-called DVD-video, two-pass variable bit rate encoding is generally used. This two-pass method performs two encodings: an encoding process for obtaining the code amount, and an encoding process performed while variably controlling the bit rate based on the obtained code amount. Although there is an advantage that the total amount of possible coded bits can be used effectively, there is a disadvantage that the processing time is required to be about twice the time length of the moving image sequence, so that it is not suitable for real-time processing.
[0006]
One-pass variable bit rate coding schemes aimed at shortening the processing time are disclosed in the specification and drawings of Japanese Patent Application Nos. 7-31418 and 9-113141, for example. I have.
[0007]
Here, FIG. 8 shows a configuration example of a moving picture encoding apparatus to which a conventional one-pass type variable bit rate encoding method is applied, and FIG. 9 shows a flowchart of a one-pass type variable bit rate encoding process. Show.
[0008]
The encoding device 2 includes an encoding difficulty calculator 20, an allocated bit amount calculator 21, a delay unit 22, and a video encoder 23, as shown in FIG. In the encoding device 2, the input video signal supplied to the input terminal A is sent to the encoding difficulty calculator 20 and the delay unit 22. An output from the encoding difficulty calculator 20 is sent to an assigned bit amount calculator 21 for calculating an assigned bit amount per unit time, and an output from the assigned bit amount calculator 21 is sent to a video encoder 23. Can be The moving picture encoder 23 encodes the output signal from the delay unit 22 according to the assigned bit amount from the assigned bit amount calculator 21 and outputs the encoded signal from the terminal 205 as an encoded bit stream.
[0009]
Here, the operation of the encoding device 2 will be described with reference to the flowchart of FIG.
[0010]
In step ST10, the encoding device 2 inputs the video signal supplied to the input terminal A to the encoding difficulty calculator 20, and calculates the encoding difficulty d of the input image per unit time. The unit time is, for example, about 0.5 seconds. The calculation of the encoding difficulty is performed, for example, by fixing the quantization step, encoding the input moving image signal, and calculating the generated code amount for each predetermined time.
[0011]
In step ST <b> 11, the encoding device 2 calculates the assigned bit amount b for the encoding difficulty d obtained from the encoding difficulty calculator 20 by the assigned bit amount calculator 21. In this case, the degree of encoding difficulty d per unit time and the allocated bit amount b when the reference moving image sequence is subjected to variable bit rate encoding at a predetermined average bit rate are associated in advance. Here, the total sum of the allocated bit amounts per unit time with respect to the reference moving image sequence is set to be equal to or less than the storage capacity of the target recording medium. FIG. 4 shows an example of the relationship between the encoding difficulty d and the allocated bit amount b.
[0012]
In FIG. 4, the horizontal axis indicates the encoding difficulty d, and the vertical axis indicates the probability h (d) of appearance of the encoding difficulty d in the reference moving image sequence. Then, an allocated bit amount for an arbitrary encoding difficulty is calculated based on the function b (d). This relationship can be determined empirically through trial and error, for example, by conducting an experiment to encode a moving image sequence such as a movie at a predetermined average bit rate, evaluating the image quality, and applying it to most sequences in the world. A general relationship that is adaptable. The method of obtaining the information is disclosed in the specification and the drawings of Japanese Patent Application No. 7-31418, for example. The allocated bit amount calculator 21 calculates the allocated bit amount b for the unit time encoding difficulty d of the image input from the input terminal A based on the relationship of FIG.
[0013]
The delay unit 22 in the encoding device 2 of the one-pass method uses the encoding difficulty calculator 20 and the allocated bit amount calculator 21 for the input image having the unit time length, and the processing is completed within the unit time. , For delaying the input of the image signal to the moving image encoder 23 by a unit time.
[0014]
Also, in step ST12, the encoding device 2 sets the moving image encoder 23 such that the input moving image for each unit time has the assigned bit amount given from the assigned bit amount calculator 21 corresponding thereto. Encode. That is, the moving image encoder 23 encodes the input moving image for each unit time using a quantization step size based on the allocated code amount.
[0015]
In such a one-pass system, variable bit rate encoding can be performed in almost real time with an optimal allocated bit amount according to the signal encoding difficulty in accordance with the input of an image signal.
[0016]
[Problems to be solved by the invention]
By the way, the relationship in FIG. 4 can be applied to most moving image sequences, but cannot be applied to some special sequences. For example, it is a joint portion where different input image signals are temporally continuous. Note that such a joint of the input image signals is called a scene change.
[0017]
In the scene change part, the difference is taken out from a continuous image and coded, and the original image is reconstructed. In order to obtain the image quality equivalent to that of a general input image signal, so-called motion compensation prediction coding cannot be applied. It is necessary to allocate more bits to the scene change part or to change the MPEG GOP structure at the scene change part. For example, by changing a P picture to an I picture, it is possible to prevent the image quality from deteriorating at the time of a scene change.
[0018]
However, when the GOP structure is changed, generally, the coding efficiency of the GOP deteriorates. Therefore, simply changing the GOP structure makes it difficult to obtain image quality equivalent to that of a general input image signal. In particular, when only a low-rate bit amount is allocated to one GOP, the bit amount required when changing from a P picture to an I picture becomes insufficient, and the image quality of a scene in which the GOP structure is changed deteriorates.
[0019]
Therefore, an object of the present invention is to propose an encoding apparatus and method, a program, and a recording medium that can detect a scene change and improve the image quality of a scene change portion even when the GOP structure is changed.
[0020]
[Means for Solving the Problems]
An encoding apparatus according to the present invention is an encoding apparatus that encodes a moving image signal based on a predetermined image structure in order to solve the above-described problem. Encoding difficulty calculating means for calculating the degree d, bit amount calculating means for calculating the bit amount b for transferring the video signal in a unit time based on the coding difficulty d, A scene change position detecting means for detecting a change position; an image structure changing means for changing the predetermined image structure based on the scene change position; and the image structure changing means when the bit amount b is lower than a predetermined threshold value. Bit amount adjusting means for adjusting the amount of bits allocated before the position where the image structure is changed by the means, and allocating the surplus bit amount obtained thereby to the position after the position; A delay unit that delays and outputs a signal; and an encoding unit that encodes a moving image signal output from the delay unit based on the image structure changed by the image structure changing unit according to the adjustment of the bit amount adjustment unit. Means.
[0021]
When the bit amount b calculated by the bit amount calculating unit is lower than a predetermined threshold, such an encoding device reduces the bit amount to be allocated before the position where the image structure is changed by the bit amount adjusting unit. The surplus bit amount obtained by the above is adjusted after the above position, and the moving image signal is encoded based on the image structure changed by the encoding unit according to the adjustment of the bit amount adjustment unit.
[0022]
An encoding method according to the present invention is an encoding method for encoding a moving image signal based on a predetermined image structure in order to solve the above-described problem. Calculate the degree d, calculate the bit amount b for transferring the video signal within a unit time based on the encoding difficulty d, detect the scene change position from the video signal, and The predetermined image structure is changed, and when the bit amount b is lower than a predetermined threshold, the amount of bits allocated before the position where the image structure is changed is reduced, and the surplus bit amount obtained by the An adjustment to be assigned to the position and thereafter is performed, the moving image signal is output with a delay, and based on the changed image structure, the moving image signal output with the delay is encoded according to the adjustment.
[0023]
Such an encoding method reduces a bit amount to be allocated before a position of an image structure changed based on a scene change position, when a bit amount b for transferring a moving image signal within a unit time is lower than a predetermined threshold, Adjustment for allocating the surplus bit amount thus obtained to the position after the position is performed, and the moving image signal is encoded based on the changed image structure according to the adjustment.
[0024]
A program to be executed by the computer according to the present invention includes a step of calculating an encoding difficulty d per unit time of a moving image signal, based on the encoding difficulty d, Calculating a bit amount b for transferring the moving image signal within a unit time, detecting a scene change position from the moving image signal, and changing the predetermined image structure based on the scene change position. Adjusting the amount of bits allocated before the position where the image structure is changed, and allocating the surplus bit amount obtained from the position after the position, when the bit amount b is lower than a predetermined threshold value; A step of delaying and outputting a moving image signal, and a step of encoding the delayed and output moving image signal based on the changed image structure in accordance with the adjustment. .
[0025]
When such a program is executed by a computer, if the bit amount b for transferring a moving image signal within a unit time is lower than a predetermined threshold, the program is allocated before the position of the image structure changed based on the scene change position. The bit amount is reduced, and the surplus bit amount thus obtained is adjusted to be allocated to the position after the position, and the moving image signal is encoded based on the changed image structure according to the adjustment.
[0026]
The recording medium according to the present invention, in order to solve the above-described problems, a step of calculating the encoding difficulty d per unit time of the moving image signal, and the step of calculating the moving image signal based on the encoding difficulty d. Calculating a bit amount b to be transferred within a unit time; detecting a scene change position from a moving image signal; changing the predetermined image structure based on the scene change position; Is lower than a predetermined threshold, a step of reducing the amount of bits to be allocated before the position where the image structure is changed, and performing an adjustment of allocating the surplus bit amount obtained by the position after the position; and A computer that records a program for executing a step of outputting with delay and a step of encoding a video signal output with delay based on the changed image structure in accordance with the adjustment. Over data is readable media.
[0027]
When such a recording medium is executed by a computer capable of reading, when the bit amount b for transferring the moving image signal in a unit time is lower than a predetermined threshold, the position before the position of the image structure changed based on the scene change position is determined. Is adjusted to allocate the surplus bit amount obtained after the above to the above position, and a moving image signal is encoded based on the changed image structure according to the above adjustment.
[0028]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
[0029]
First, the background of the present invention will be described. For example, a single-sided, single-layer DVD (Digital Versatile Disc), which is one of the optical disk media, is a large-capacity medium in which the amount of data that can be recorded is about 4.7 GB, which is about seven times that of a CD. However, if an attempt is made to digitize and record a video signal without performing any processing, long-time recording cannot be performed. For example, when an NTSC video signal is digitized, the data amount per second becomes 20 MB or more. Here, 20 MB is a value calculated assuming that 29.97 screens of 720 × 480 pixels are provided per second with 8 bits for luminance and 8 bits for color per pixel.
[0030]
Therefore, when trying to record an NTSC video signal on the DVD, only about 4 minutes of video can be recorded. Therefore, the DVD adopts MPEG (Moving Picture Image Coding Experts Group) 2, which compresses an NTSC video signal so that about 133 minutes of video can be recorded.
[0031]
A moving image signal compression method based on MPEG2 uses discrete cosine transform (DCT), which is compression using correlation within a screen, motion compensation using correlation between screens, and correlation between code strings. It is a combination of three Huffman codings, which is compression.
[0032]
Here, the motion compensation will be described. The video signal of the NTSC system is composed of 30 frames per second, and is scanned by an interlace (interlaced scanning) system in order to eliminate flicker. Therefore, the display device displays the screen at 60 fields per second. Note that 30 frames are 30 pictures formed from 525 scanning lines.
[0033]
Looking at each of the 30 pictures per second, successive images often have a large area of the screen occupied by the same element, and do not change much in a short amount of time.
[0034]
Therefore, for example, when there is a portion that changes in 30 pictures, only the change (difference) is extracted, and the other portions are recorded only once. Then, if the differences are combined at the time of reproduction, the amount of information necessary for recording / reproduction can be reduced. Such a technique of extracting a difference from a continuous image, encoding the original image, and reconstructing the original image is called predictive encoding because the current image is predicted from the immediately preceding image.
[0035]
In addition, the changing part of the image, that is, the moving part, includes a part whose shape does not change and its position on the screen simply changes with time, and a part whose shape changes with time. In the former case, the shape data can be used as it is. For an element whose position on the screen changes over time without changing its shape, the amount of the change, that is, the amount of motion shift is called a motion vector. By encoding this motion vector and transmitting the encoded signal, the original image can be reconstructed with a smaller amount of data. Note that a method of using this motion vector in predictive coding is called motion compensation.
[0036]
Next, the above-described prediction coding and motion compensation processing will be described with reference to FIG. In the predictive coding process, for example, when there is an image A and an image B to be reproduced according to the flow of time, since the contents of the images A and B have many common elements, the images A and B are left as they are. Instead of encoding, an image C which is the difference between the images A and B is generated, and the generated image C is encoded together with the image A. The image A and the image C thus encoded can be combined with the image A and the image C to reconstruct the image B at the time of decoding. Note that the image C is called a prediction coded image.
[0037]
In the motion compensation processing, an image is divided into blocks, and a motion vector indicating a direction and how much the image has moved is extracted from a portion where the shape has not changed and only the position has changed, and coding is performed. Combining the above-described motion compensation and predictive coding enables efficient data compression.
[0038]
In addition, in MPEG2, in order to perform predictive coding using motion compensation, as shown in FIG. 2, an I picture (Intra coded picture), a P picture (Predictive coded picture) and a B picture (Bidirectionally Predictive coded picture). A GOP (Group of Picture) structure using three elements, that is, an image) is used. An I picture is a picture created by intra-frame encoding, and does not perform predictive encoding from a previous image. The P picture is an inter-frame forward prediction coded image created by performing predictive encoding from the immediately preceding image, and is a picture created based on an I picture. A B picture is a bidirectionally predicted coded image, and is a picture created by performing prediction from two preceding and succeeding P pictures. In general, one GOP is formed by a group of 15 pictures. Note that one GOP is composed of one I picture, four P pictures, and ten B pictures. The bit amount required to generate an I picture is 5 to 6 times the bit amount required to generate a B picture, and the bit amount required to generate a P picture is It is 2-3 times the amount of bits needed to generate.
[0039]
In general, a video signal is not stationary, and the information amount of each picture changes with time. Therefore, it is known that higher image quality can be obtained by using variable bit rate coding than by fixed bit rate coding that performs coding with the same code amount.
[0040]
For a video signal recorded on a DVD, variable bit rate encoding called a two-pass method is generally used. In the two-pass method, two encoding processes are performed: an encoding process for obtaining a code amount, and an encoding process for performing encoding while variably controlling a bit rate based on the obtained code amount. There is an advantage that the total amount of available coded bits can be used effectively. However, there is a disadvantage that the processing time is required to be about twice the time length of the moving image sequence, so that it is not suitable for real time processing.
[0041]
In the real-time processing, it is necessary to perform coding once by calculating the code amount and variably controlling the bit rate based on the code amount. Such processing is called a one-pass method. An encoding apparatus 1 to which the present invention is applied employs the one-pass method, detects a scene change position from an input moving image signal, and determines a predetermined bit amount to be assigned per unit time of the moving image signal. If the threshold value is lower than the threshold value, the bit amount originally allocated before the predetermined position determined based on the scene change position is reduced, and the surplus bit amount obtained thereby is allocated to the predetermined position and thereafter. And encodes the moving image signal in accordance with the adjustment. Hereinafter, the configuration and operation of the encoding device 1 will be described in detail.
[0042]
As shown in FIG. 3, the encoding device 1 includes an encoding difficulty calculator 10, an allocated bit amount calculator 11, a scene change position detector 12, a delay unit 13, a moving image encoder 14, Is provided. The encoding difficulty calculator 10 calculates the encoding difficulty d of the video signal supplied from the input terminal for each unit time, and outputs the calculated encoding difficulty d to the allocated bit amount calculator 11. The unit time is an elapsed time of 0.5 seconds or 1 GOP. Further, the encoding difficulty calculator 10 encodes the input moving image signal with the quantization step fixed, and calculates the encoding difficulty d by calculating the generated code amount every predetermined time. . Note that the encoding difficulty calculator 10 calculates the encoding difficulty d higher for a scene with a lot of motion, and calculates the encoding difficulty d lower for a scene with a little motion.
[0043]
The assigned bit amount calculator 11 calculates an assigned bit amount b for the inputted encoding difficulty d. In this case, the degree of encoding difficulty d per unit time and the allocated bit amount b when the reference moving image sequence is subjected to variable bit rate encoding at a predetermined average bit rate are associated in advance. Here, the total sum of the allocated bit amounts per unit time with respect to the reference moving image sequence is set to be equal to or less than the storage capacity of the target recording medium. FIG. 4 shows an example of the relationship between the encoding difficulty d and the allocated bit amount b.
[0044]
In FIG. 4, the horizontal axis represents the encoding difficulty d, and the vertical axis represents the appearance probability h (d) of the encoding difficulty d in the reference moving image sequence. Then, an allocated bit amount for an arbitrary encoding difficulty is calculated based on the function b (d). This relationship can be determined empirically through trials and errors by performing experiments on encoding a large number of moving image sequences (for example, movies) at a predetermined average bit rate, and evaluating the image quality. A general relationship that is adaptable. The allocated bit amount calculator 11 calculates the allocated bit amount b for the encoding difficulty d per unit time of the moving image signal based on the relationship in FIG. When the encoding difficulty d is high, the allocated bit amount calculator 11 increases the encoding rate to reduce noise due to encoding. When the encoding difficulty d is low, the noise due to the encoding does not occur much, and the encoding rate is calculated to be lower. By doing so, the image quality of the entire moving image signal becomes uniform, and the average bit amount per unit time decreases.
[0045]
However, in the conventional encoding device, since the algorithm controls the bit amount per unit time to be low within 30 minutes or 1 hour, a large amount of bits is allocated without any particular control for the first few minutes. A process of allocating a small bit amount in the latter half may be performed. With such a control, for example, if the playback operation of a piece is repeatedly performed in which a moving image signal is recorded for one minute, stopped, and then recorded again for one minute, a large amount of bits is allocated, and a large amount of bits per unit time is allocated. There is a problem that the average bit amount becomes high. The encoding device 1 according to the present invention can control the average bit amount per short time to be low, particularly when the fragment playback operation is repeated. The allocated bit amount calculator 11 allocates a bit amount b of about 2.8 Mbps to 15 Mbps, for example.
[0046]
When the calculated bit amount b is lower than a predetermined threshold value and a signal indicating that there is a scene change is supplied from a scene change position detector 12 described later, the allocated bit amount calculator 11 calculates the bit amount b. The adjustment is performed before and after the scene change position, and the adjustment result is output to the video encoder 14. The adjustment by the allocated bit amount calculator 11 will be described later.
[0047]
The scene change position detector 12 detects a scene change position from the moving image signal supplied from the input terminal, generates a signal indicating the detection, and outputs the signal to the assigned bit amount calculator 11 and the moving image encoder 14. .
[0048]
Here, a specific operation of the scene change position detector 12 will be described below. The scene change position detector 12 detects a scene change position by detecting a large change in a parameter characterizing each frame of the moving image signal. Therefore, the scene change position detector 12 only needs to be able to detect a change in the signal such as the luminance level and the chroma level of the screen. The scene change position detector 12 detects a scene change position using, for example, a change in an average luminance level, a difference amount between frames, a level fluctuation amount between frames, and the like.
[0049]
Next, detection of a scene change position by the above-described scene change position detector 12 will be described with reference to a flowchart shown in FIG.
[0050]
In step ST1, the scene change position detector 12 determines whether a scene change position has been detected from the input moving image signal. If a scene change position has been detected, the process proceeds to step ST2.
[0051]
In step ST2, the scene change position detector 12 generates a signal indicating that the scene change position has been detected, and outputs the signal to the video encoder 14. The moving picture encoder 14 changes the GOP structure according to the scene change position.
[0052]
In step ST3, the scene change position detector 12 outputs a signal to the effect that the scene change position has been detected to the allocated bit amount calculator 11. The allocated bit amount calculator 11 adjusts the bit amount allocated before and after the scene change position according to the scene change position.
[0053]
The delay unit 13 temporarily stores the moving image signal for at least the time required for the encoding difficulty calculator 10 and the allocated bit amount calculator 11 to calculate the bit amount b per unit time of the moving image signal. The delay unit 13 supplies the moving image signal to the moving image encoder 14 after temporarily storing the moving image signal.
[0054]
The moving image encoder 14 encodes the supplied moving image signal according to the bit amount b supplied from the allocated bit amount calculator 11 per unit time. Further, the moving picture encoder 14 changes the GOP structure when a signal indicating that there is a scene change is supplied from the scene change position detector 12. Specifically, as shown in FIG. 6, a P picture closest to the scene change position is changed to an I picture in the time direction. The video encoder 14 changes the P picture (P5) to an I picture (I5) when a scene change occurs at a position A, and changes the P picture (P8) to an I picture (I8) when a scene change occurs at a position B. The picture is changed to picture (I8), and when a scene change occurs at position C, the P picture (P11) is changed to I picture (I11).
[0055]
As described above, when the picture is changed from the P picture to the I picture, a bit amount twice or three times is required. There is no problem when the bit amount allocated to the unit time is higher than the predetermined threshold, but when the bit amount is lower than the predetermined threshold, the bit amount becomes insufficient due to the change of the GOP structure, and Deteriorates. Therefore, when the bit amount b to be allocated to the unit time of the moving image signal is lower than the predetermined threshold, the allocated bit amount calculator 11 determines the predetermined bit amount determined based on the scene change position according to the detection result of the scene change position. The bit amount allocated before the position and the bit amount allocated after the predetermined position are adjusted.
[0056]
Further, as shown in FIG. 7, it is possible to respond to a scene change by moving a picture within a 2GOP without changing from a P picture to an I picture. In this case, although the position of the I picture (I2) is only moved from the P picture (P11) within the 2GOP, there is no large change in the bit amount required in the 2GOP, but it is difficult to encode the scene change itself. Since a larger bit amount is required, it is necessary to allocate a larger bit amount than usual after the sea change.
[0057]
Here, the adjustment operation by the allocated bit amount calculator 11 will be described below. When the calculated bit amount b is lower than a predetermined threshold, the allocation bit amount calculator 11 performs the following adjustment. When a scene change as described above occurs, an image encoded by prediction cannot be used, and the image quality after the scene change may deteriorate. Therefore, when a signal indicating that there is a scene change is supplied from the scene change position detector 12, the moving image encoder 14 described later performs a process of changing the GOP structure based on the scene change. This can prevent the image quality from deteriorating at the time of a scene change.
[0058]
However, when the GOP structure is changed, generally the coding efficiency of the GOP becomes poor, and it is difficult to obtain image quality equivalent to that of a general moving image signal by simply changing the GOP structure. Therefore, when a signal indicating that there is a scene change is supplied from the scene change position detector 12, the allocated bit amount calculator 11 calculates the bits allocated before a predetermined position determined based on the scene change position. The amount is reduced, and an adjustment is made to allocate the surplus bit amount thus obtained to the predetermined position and beyond, and the adjustment result is supplied to the moving picture encoder 14. The moving picture encoder 14 codes the moving picture signal output from the delay unit 13 based on the supplied adjustment result.
[0059]
The assigned bit amount calculator 11, for example, assuming that the bit amount originally scheduled to be assigned to the moving image signal before the predetermined position determined based on the scene change position is R_pre and the decreasing bit rate rate is SC_rate, The bit amount R_pre ′ to be allocated to the video signal before the position of
R_pre '= R_pre × SC_rate
When the bit amount originally allocated to the moving image signal after the predetermined position determined based on the scene change position is R_post, the bit amount R_post ′ to be allocated to the moving image signal after the predetermined position is
R_post '= R_post + (R_pre-R_pre')
Adjust according to.
[0060]
Also, the decreasing bit rate SC_rate changes according to the scene change position. For example, when a scene change occurs at the position A shown in FIG. 6, SC_rate = 0.6, and the scene change occurs at the position B. If a scene change has occurred, SC_rate = 0.7, and if a scene change has occurred at position C, SC_rate = 0.8.
[0061]
The video encoder 14 encodes the video signal according to the adjustment result supplied from the allocation bit detector.
[0062]
When the encoding device 1 configured as described above calculates the bit amount b lower than the predetermined threshold by the calculation of the allocated bit amount calculator 11, based on the scene change position detected by the scene change position detector 12, The bit amount at the scene change position is adjusted, the image structure at the scene change position is changed, and the moving image signal is encoded based on the adjustment according to the changed image structure. Encoding with high encoding efficiency can be performed, and deterioration of image quality due to a scene change occurring within a unit time with a small allocated bit amount b can be prevented.
[0063]
The embodiment of the present invention is not limited to the above example, and may be a computer-executable program or a computer-readable recording medium on which the program is recorded.
[0064]
【The invention's effect】
As described in detail above, when the bit amount calculating unit calculates the bit amount b lower than the predetermined threshold based on the encoding difficulty d, the encoding device according to the present invention uses the bit amount adjusting unit The amount of bits allocated before the predetermined position determined based on the change position is reduced, and the surplus bit amount obtained thereby is adjusted to be allocated after the predetermined position, and the encoding means performs scene change based on the adjustment. Since the image structure is changed based on the position and the moving image signal is coded according to the changed image structure, coding with high coding efficiency can be performed even if the image structure at the scene change position is changed. It is possible to prevent the image quality from deteriorating due to a scene change occurring within a unit time with a small bit amount b.
[0065]
Further, the encoding method according to the present invention, when a bit amount b lower than a predetermined threshold is calculated based on the calculated encoding difficulty d, a bit to be allocated before a predetermined position determined based on the scene change position. The amount is reduced and the surplus bit amount thus obtained is adjusted to be allocated after the predetermined position. Based on the adjustment, the image structure is changed based on the scene change position, and the moving image signal is changed according to the changed image structure. Encoding, it is possible to perform encoding with high encoding efficiency even if the image structure at the scene change position is changed, and the image quality is degraded due to a scene change that occurs within a unit time when the allocated bit amount b is small. Can be prevented.
[0066]
Further, the program according to the present invention, when the computer calculates a bit amount b lower than a predetermined threshold based on the calculated encoding difficulty d, allocates the bit amount before a predetermined position determined based on the scene change position. The bit amount is reduced, and the surplus bit amount obtained thereby is adjusted to be allocated after the predetermined position. Based on the adjustment, the image structure is changed based on the scene change position, and the moving image is changed according to the changed image structure. Since the signal is encoded, a step of performing encoding with high coding efficiency is performed even if the image structure at the scene change position is changed. Therefore, a code with high coding efficiency even if the image structure at the scene change position is changed. Image quality can be prevented from being deteriorated due to a scene change occurring within a unit time in which the allocated bit amount b is small. .
[0067]
Further, the recording medium according to the present invention, when calculating a bit amount b lower than a predetermined threshold based on the calculated encoding difficulty d, assigning a bit amount before a predetermined position determined based on the scene change position Is adjusted, and the surplus bit amount obtained thereby is adjusted after the predetermined position.Based on the adjustment, the image structure is changed based on the scene change position, and the moving image signal is changed according to the changed image structure. Since the encoding is performed, a computer-readable medium storing a program for executing a step of performing encoding with high encoding efficiency even if the image structure at the scene change position is changed. Can be encoded with high coding efficiency even if the It is possible to prevent the deterioration of image quality due to change.
[Brief description of the drawings]
FIG. 1 is a diagram showing how a predicted coded image is generated from a non-predicted coded image by a predictive coding process.
FIG. 2 is a structural diagram showing a GOP structure.
FIG. 3 is a block diagram illustrating a configuration example of an encoding device to which the present invention has been applied.
FIG. 4 is a distribution diagram showing a macroblock appearance probability h (d) with respect to an encoding difficulty d and an assigned code amount b.
FIG. 5 is a flowchart showing a scene change detection operation by a scene change position detector included in the encoding device to which the present invention is applied.
FIG. 6 is a diagram illustrating a state where a GOP structure is changed based on detection of a scene change position.
FIG. 7 is a diagram illustrating a case where a picture is moved within a 2GOP structure.
FIG. 8 is a block diagram illustrating a configuration example of a conventional encoding device.
FIG. 9 is a flowchart illustrating a conventional one-pass variable bit rate encoding process.
[Explanation of symbols]
REFERENCE SIGNS LIST 1 encoder, 10 encoding difficulty calculator, 11 assigned bit amount calculator, 12 scene change position detector, 13 delay unit, 14 video encoder

Claims (7)

An encoding device that encodes a moving image signal based on a predetermined image structure,
Encoding difficulty calculating means for calculating the encoding difficulty d per unit time of the moving image signal;
A bit amount calculating means for calculating a bit amount b for transferring the moving image signal in a unit time based on the encoding difficulty d;
A scene change position detecting means for detecting a scene change position from a moving image signal;
Image structure changing means for changing the predetermined image structure based on the scene change position;
When the bit amount b is lower than a predetermined threshold, the bit amount to be allocated before the position where the image structure is changed by the image structure changing means is reduced, and the surplus bit amount obtained by this is allocated to the position after the position. Bit amount adjusting means for adjusting;
Delay means for delaying and outputting a moving image signal;
Encoding means for encoding a moving image signal output from the delay means based on the image structure changed by the image structure changing means in accordance with the adjustment of the bit amount adjusting means. apparatus. The image structure is a GOP (Group of Picture) structure formed by a predetermined number of I pictures (Intra coded images), P pictures (Predictively coded images), and B pictures (Bidirectionally Predictive coded images). The encoding device according to claim 1, wherein: 3. The encoding apparatus according to claim 2, wherein said image structure changing means changes the first P picture after the scene change position detected by said scene change position detecting means to an I picture. When the bit amount b is lower than a predetermined threshold value, the bit amount adjusting unit sets the bit amount originally allocated to the moving image signal before the position where the image structure is changed to R_pre, and decreases the bit rate rate. Is SC_rate, the bit amount R_pre ′ to be allocated to the video signal before the position is
R_pre '= R_pre × SC_rate
And the bit amount originally allocated to the moving image signal after the position is R_post, the bit amount R_post ′ to be allocated to the moving image signal after the position is
R_post '= R_post + (R_pre-R_pre')
2. The encoding apparatus according to claim 1, wherein the encoding is performed according to the following. An encoding method for encoding a moving image signal based on a predetermined image structure,
Calculate the encoding difficulty d per unit time of the video signal,
Based on the encoding difficulty d, a bit amount b for transferring the video signal in a unit time is calculated,
Detect scene change position from video signal,
Changing the predetermined image structure based on the scene change position,
When the bit amount b is lower than a predetermined threshold, the amount of bits to be allocated before the position where the image structure is changed is reduced, and the surplus bit amount obtained by the adjustment after the position is adjusted.
Delays and outputs the moving image signal,
A coding method, characterized by coding a moving image signal output with a delay based on the changed image structure in accordance with the adjustment. On the computer,
Calculating an encoding difficulty d per unit time of the moving image signal;
Calculating a bit amount b for transferring the video signal within a unit time based on the encoding difficulty d;
Detecting a scene change position from the moving image signal;
Changing the predetermined image structure based on the scene change position;
When the bit amount b is lower than a predetermined threshold, a step of reducing the bit amount to be allocated before the position where the image structure is changed, and performing an adjustment to allocate the surplus bit amount obtained from the position and thereafter,
A step of delaying and outputting the moving image signal;
A program for executing, based on the changed image structure, a step of encoding a delayed moving image signal according to the adjustment. Calculating an encoding difficulty d per unit time of the moving image signal;
Calculating a bit amount b for transferring the video signal within a unit time based on the encoding difficulty d;
Detecting a scene change position from the moving image signal;
Changing the predetermined image structure based on the scene change position;
When the bit amount b is lower than a predetermined threshold, a step of reducing the bit amount to be allocated before the position where the image structure is changed, and performing an adjustment to allocate the surplus bit amount obtained from the position and thereafter,
A step of delaying and outputting the moving image signal;
A computer-readable storage medium storing a program for executing a step of encoding a moving image signal output with a delay based on a changed image structure in accordance with the adjustment.

Images