JP2001359096A - Image coding device - Google Patents

Abstract

(57) [Summary] [Problem] To obtain a high quality video with a smaller code amount by performing encoding at a resolution corresponding to a high frequency component included in an input video frame. An input video frame is input to a frequency band analyzer 101, and is converted from a spatial domain to a frequency domain. The resolution determiner 102 analyzes the output of the frequency band analyzer 101 and thereby determines the resolution of the video frame. Based on the resolution determined by the resolution determiner 102, the resolution converter 103 performs resolution conversion on the video frame. And then, the blocking unit 104, DC
T converter 105, quantizer 106, variable length encoder 10
7, and is output as a code string.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image encoding apparatus for compressing and encoding a video signal.

[0002]

2. Description of the Related Art In recent years, MPEG2 (Moving Picture Experts Groove) has been used as one of high-efficiency compression coding methods for video signals.
up 2) The method is widely known. In the MPEG2 system,
An I (intra) frame for performing intra-frame encoding of a video frame (field) without using a reference image, and a P (predictive) or B (bidirect) for performing inter-frame predictive encoding
and encoding is performed. Each frame is divided into units called macroblocks having a size of 16 × 16 pixels, and further divided into 8 × 16 pixels.
Encoding is performed by performing discrete cosine transform (DCT) on a block having a size of 8 pixels.

[0003]

Problems to be Solved by the Invention
According to the related art described in Japanese Patent Application Laid-Open No. 6-64, the product of the quantization scale (quantization step size) and the generated code amount is increased in order to always keep the generated code amount within a predetermined code amount. When the buffer occupancy increases,
That is, when a complex video that is difficult to encode is input, the resolution of the subsequent input video signal is reduced. Therefore, for example, when the generated code amount is about to exceed the predetermined code amount, the resolution is lowered even if the input video signal contains many high-frequency components, so that the image quality is deteriorated.

An object of the present invention is to provide an image coding apparatus capable of greatly reducing the amount of generated codes while minimizing image quality deterioration.

[0005]

In order to achieve the above object, the present invention is to change the resolution according to the characteristics of an input video signal. For example, according to an embodiment, the resolution of an input video signal is increased when a complex video that is difficult to encode is input, and the resolution of the input video signal is reduced when a monotonous video is input. The details will be described later.

[0006]

Embodiments of the present invention will be described below with reference to the drawings.

(First Embodiment) A first embodiment of the present invention will be described with reference to FIG. FIG. 1 is a block diagram of an image encoding device including a delay unit 100, a frequency band analyzer 101, a resolution determiner 102, a resolution converter 103, and a compression encoder 110. The compression encoder 110
Blocker 104, DCT transformer 105, quantizer 1
06, the variable length encoder 107. Here, a block diagram in the case of intra-frame encoding is shown for simplification of description.

An input video signal, which is a digital signal, is first input to a frequency band analyzer 101. The frequency band analyzer 101 converts a video frame from a spatial domain to a frequency domain. The conversion to the frequency domain can be performed, for example, by performing a Fourier transform on the video signal. This conversion may be performed for the entire frame, or may be performed, for example, by dividing into blocks or macroblocks and for each block or macroblock. Then, the frequency band analyzer 101 outputs a result obtained by converting the video signal into a frequency domain.

A resolution determiner 102 analyzes a video signal converted into a frequency domain, which is an output of the frequency band analyzer 101, and how to change the resolution (number of pixels × number of lines) of a video frame accordingly. To determine. This determination method will be described.

FIGS. 2A to 2D are schematic diagrams showing examples of resolution. Here, it is assumed that a resolution is selected from three types of resolutions A, B, and C shown in FIGS. 2B to 2D. 2A shows the resolution of the input video frame, FIG. 2B shows the resolution A obtained by halving the resolution of the input video frame in the horizontal direction, and FIG. 2C shows the resolution B of the input video frame. The resolution obtained by reducing the frame resolution to 水平 in the horizontal direction, and the resolution C in FIG. 2D is the same as the resolution of the input video frame.

FIG. 3 is a flowchart for determining the resolution, and FIG.
FIGS. 4A and 4B are graphs showing examples of frequency components. The resolution determiner 102 determines the resolution according to the flowchart shown in FIG. First, the highest frequency component fmax of the video frame converted into the frequency domain is obtained.
fmax can be obtained as the maximum value of the frequency component when the frequency conversion in the frequency band analyzer 101 is performed on the entire frame. If the frequency conversion in the frequency band analyzer 101 is performed for each block or macroblock, fmax is set as the maximum value of the highest frequency component of each block or macroblock.
Can be obtained. Then, a comparison is made as to whether fmax is smaller than a predetermined threshold value TH1. If fmax is smaller than TH1, the resolution is determined to be resolution A. If fmax is equal to or greater than the threshold value TH1, a comparison is made as to whether fmax is smaller than a predetermined threshold value TH2. Here, TH1 <TH2. If fmax is smaller than TH2, the resolution is determined to be resolution B. If fmax is equal to or larger than the threshold value TH2, the resolution is determined to be the resolution C. For example, when the frequency components of the video signal output from the frequency band analyzer 101 and the threshold values TH1 and TH2 are values as shown in FIG. 4A, the resolution B is selected according to the flowchart of FIG. . FIG. 4B shows a case where the maximum frequency at which the frequency component is equal to or higher than a predetermined threshold value fth is fmax. In this case, the resolution A is selected according to the flowchart of FIG.

An input video signal is supplied to a resolution converter 103 via a delay unit 100. The delay unit 100 delays the input video signal by the time required for the operation of the frequency band analyzer 101 and the resolution determiner 102. On the other hand, the resolution determined by the resolution determiner 102 is output to the resolution converter 103. Thereby, the resolution converter 10
Reference numeral 3 applies a folding noise removal filter to the video frame provided via the delay unit 100, and then converts the resolution by performing interpolation or thinning. For example, in the case of a video frame having frequency components shown in FIG. 4A, the resolution B, that is, the horizontal direction is reduced to ３ of the original video.
When the resolution C is selected, the resolution converter 1
03 outputs the input video frame as it is.

The video frame output from the resolution converter 103 is input to a blocker 104 and is divided into blocks. The block has a size of, for example, about 8 × 8 pixels. Then, each block is input to the DCT converter 105. The DCT converter 105 applies two-dimensional DCT to each block to convert the blocks into DCT coefficients. DCT converter 1
The DCT coefficient which is the output of 05 is input to the quantizer 106, and the coefficient is rounded by a quantization process. The output of the quantizer 106 is input to the variable-length encoder 107, subjected to variable-length encoding, and then output as a code string.
Blocker 104, DCT transformer 105, quantizer 1
06, the resolution determiner 1 in the variable length encoder 107
The encoding is controlled based on the resolution determined in 02.

As described above, the image coding apparatus according to the first embodiment analyzes the frequency components included in the input video signal,
The resolution is determined based on which frequency band the highest frequency component is in. In this case, the resolution is determined such that the higher the frequency component, the higher the resolution. Then, the input video is converted to the determined resolution and encoded. Therefore, by using the image encoding device of the first embodiment, it is possible to encode an input video with a resolution corresponding to the frequency component of the input video. That is, since the number of blocks to be coded can be reduced according to the frequency components of the input video, the amount of generated codes can be greatly reduced with little image quality deterioration.

In this embodiment, the frequency is 2
The case where three threshold values TH1 and TH2 are set and one resolution is selected from three types of resolutions (A, B, and C) has been described. It may be a different value from the embodiment.

FIG. 5 shows a modification of FIG. FIG. 5 differs from FIG. 1 in that the input video signal is an analog signal. The resolution converter 103 in FIG. 1 is replaced by a sampler 103a. According to FIG. 5, the resolution determined by the resolution determiner 102 is output to the sampler 103a. Thus, the sampler 103a samples the input video frame given via the delay unit 100 at a sampling frequency according to the determined resolution. Further, at the time of sampling, an aliasing distortion removal filter corresponding to the sampling frequency is applied. For example, CCIR REC. 601
According to the standard, in the case of NTSC, sampling of the luminance signal in the horizontal direction is performed at a sampling frequency of 13.5 MHz. Therefore, in the sampler 103a, the sampling frequency is 13.5 MHz when the resolution C is selected, the sampling frequency is 10.125 MHz when the resolution B is selected, and the sampling frequency is set when the resolution A is selected. Sampling is performed at a frequency of 6.75 MHz, and a digital signal is output. The operation after the block generator 104 is the same as that in FIG.

(Second Embodiment) A second embodiment of the present invention will be described with reference to FIG. FIG. 6 shows a delay unit 100, filters 201 and 202, output value comparators 204 and 205,
FIG. 2 is a block diagram of an image encoding device including a resolution determiner 207, a resolution converter 103, and a compression encoder 110.

An input video signal, which is a digital signal, is first input to filters 201 and 202. Filter 20
Reference numeral 1202 denotes low-pass filters having different cutoff frequencies. FIG. 7 shows an example of the frequency characteristics of the filters 201 and 202. As shown in FIG. 7, here, the cut-off frequency fc2 of the filter 202 is higher than the cut-off frequency fc1 of the filter 201.
Shall be higher. Outputs of the filters 201 and 202 are input to output value comparators 204 and 205, respectively.

The output value comparators 204 and 205 receive the filter output and calculate the energy of the filter output signal. Then, the energy value is compared with a predetermined threshold value. The comparison result is output to the resolution determiner 207. Here, the output value comparator 2
04 and 205 output “large” when the filter output is larger than the threshold, and output “small” when the filter output is smaller than the threshold.

The resolution determiner 207 is provided for the output value comparator 20
The resolution of the input video frame is determined based on the comparison result of 4,205. Here, similarly to the first embodiment, three types of resolutions A, B, and B shown in FIGS. 2B to 2D are used.
C shall be determined. FIG. 8 shows the determination method.

As shown in FIG. 8, the resolution determiner 207
When the comparison result of the output value comparator 204 is “small”, the resolution A that is the lowest resolution is selected. Also,
When the comparison result of the output value comparator 204 is “large” and the comparison result of the output value comparator 205 is “small”,
Select resolution B. When the comparison result of the output value comparator 205 is “large”, the resolution C is selected. That is, when there is one of the output values of the filters 201 and 202 larger than a predetermined threshold value, the resolution is determined so that the higher the cutoff frequency of the filter, the higher the resolution.

The resolution determined by the resolution determiner 207 is output to the resolution converter 103. Thereby, the resolution converter 103 applies the aliasing removal filter to the video frame given through the delay unit 100,
Convert the resolution by performing interpolation or thinning. The video frame output from the resolution converter 103 is output to the compression encoder 1
It is converted into a code string by 10.

As described above, the image coding apparatus according to the second embodiment applies a low-pass filter having a different cutoff frequency to an input video signal, and compares the filter output with a predetermined threshold value. Compare. Then, the resolution is determined so that the higher the cutoff frequency of the filter that outputs an output larger than the threshold value, the higher the resolution. Then, the input video is converted to the determined resolution and encoded. Therefore, by using the image encoding device of the second embodiment, it is possible to encode the input video with a resolution corresponding to the frequency component of the input video. That is, since the number of blocks to be coded can be reduced according to the frequency components of the input video, the amount of generated codes can be greatly reduced with little image quality deterioration.

In this embodiment, a case has been described in which one of three resolutions (A, B, C) is selected by using the filters 201 and 202 having two different cutoff frequencies. The number of filters and the number of resolution types may be different values from the present embodiment.

FIG. 9 shows a modification of FIG. FIG. 9 differs from FIG. 6 in that the input video signal is an analog signal. The resolution converter 103 in FIG. 6 is replaced by a sampler 103a. According to FIG. 9, the resolution determined by the resolution determiner 207 is output to the sampler 103a. Thus, the sampler 103a samples the input video frame given via the delay unit 100 at a sampling frequency corresponding to the determined resolution, and supplies a digital video signal to the compression encoder 110.

(Third Embodiment) A third embodiment of the present invention will be described with reference to FIG. FIG.
0, activity measuring device 301, resolution determining device 30
FIG. 2 is a block diagram of an image encoding device including a resolution converter 103 and a compression encoder 110.

An input video signal, which is a digital signal, is first input to the activity measuring device 301. The activity measuring device 301 calculates the activity of a video frame. Here, the activity is a variance of pixel values in the frame. In a frame composed of N pixels × N lines, when i is a pixel position in the horizontal direction, j is a pixel position in the vertical direction, and Pij is a pixel value, the average value Pm and the variance value var are Pm = (1 / N ^2). ) {Pij var = (1 / N ² )} (Pij−Pm) ² Where ΣΣ is from 0 to N for i
It means an operation of summing up to −1 and summing up j from 0 to N−1. Then, the activity measuring device 301 outputs the calculated activity. The activity may be a variance of pixel values for each macroblock or block.

The resolution determining unit 302 uses the activity value output from the activity measuring unit 301 to
Determine how to change the resolution of the video frame. This determination method will be described.

FIG. 11 is a flowchart for determining the resolution. Now, as in the first embodiment, it is assumed that a resolution is selected from three types of resolutions A, B, and C shown in FIGS. 2B to 2D.

The resolution determiner 302 determines the resolution according to the flowchart shown in FIG. First, the activity act is determined. For example, when the activity is obtained as the variance of the pixel values in the frame by the activity measuring device 301, the value is act. When the activity measuring device 301 calculates an activity as a variance value for each macroblock or block, the average value of those activities in one frame may be set as act, or the maximum value in the frame may be set as act. Is also good. FIG. 12 shows the latter detailed procedure. Then, as shown in FIG. 11, a comparison is made as to whether act is smaller than a predetermined threshold value TH3. If act is smaller than TH3, the resolution is determined to be resolution A. If act is equal to or greater than threshold TH3, act is a predetermined threshold T
A comparison is made to see if it is smaller than H4. Where TH3 <TH
4. If act is smaller than TH4, the resolution is determined to be resolution B. If act is equal to or greater than threshold value TH4, the resolution is determined to be resolution C.

The resolution determined by the resolution determiner 302 is output to the resolution converter 103. Thereby, the resolution converter 103 applies the aliasing removal filter to the video frame given through the delay unit 100,
Convert the resolution by performing interpolation or thinning. The video frame output from the resolution converter 103 is output to the compression encoder 1
It is converted into a code string by 10.

As described above, the image coding apparatus according to the third embodiment calculates the activity of an input video, and determines the resolution according to the value of the activity. In this case, the resolution is determined so that the higher the activity, the higher the resolution. Then, the input video is converted to the determined resolution and encoded. Therefore, by using the image encoding device of the third embodiment, it is possible to encode an input video at a resolution corresponding to the activity of the input video. That is, since the number of blocks to be coded can be reduced according to the activity of the input video, the amount of generated codes can be greatly reduced with little image quality deterioration.

In this embodiment, the case where two threshold values TH3 and TH4 are set for the activity and one resolution is selected from three types of resolutions (A, B and C) has been described. However, the number of thresholds and the number of types of resolutions for activities may be different from those in the present embodiment.

FIG. 13 shows a modification of FIG. FIG. 13 differs from FIG. 10 in that the input video signal is an analog signal. The resolution converter 103 in FIG. 10 is replaced by a sampler 103a. According to FIG. 13, the resolution determined by the resolution determiner 302 is output to the sampler 103a. Thus, the sampler 103a uses the sampling frequency corresponding to the determined resolution to set the delay
The sampling of the input video frame given via 0 is performed, and the digital video signal is supplied to the compression encoder 110.

(Fourth Embodiment) A fourth embodiment of the present invention will be described with reference to FIG. FIG. 14 shows a DCT coefficient analyzer 401, a resolution determiner 402, and a resolution converter 10
FIG. 3 is a block diagram of an image encoding device including a compression encoder 110. The compression encoder 110 includes a block converter 104, a DCT transformer 105, a quantizer 106, and a variable length encoder 107.

An input video signal, which is a digital signal, is first input to the resolution converter 103. Then, the resolution converter 103 converts the resolution of the video frame based on the resolution determined by a method described later. Resolution converter 10
3 is output to the blocker 104
And is divided into blocks. The block has a size of, for example, about 8 × 8 pixels. Then, each block is input to the DCT converter 105. DCT converter 105
Then, two-dimensional DCT is applied to each block to convert it into DCT coefficients. The DCT coefficients output from the DCT transformer 105 are input to a quantizer 106, where the coefficients are rounded by a quantization process. The output of the quantizer 106 is input to the variable-length encoder 107, subjected to variable-length encoding, and then output as a code string.

On the other hand, D which is the output of DCT converter 105
The CT coefficients are input to a DCT coefficient analyzer 401. D
In the CT coefficient analyzer 401, the input DCT coefficient is
Analyze over the frame period. DCT coefficient analyzer 40
The analysis in step 1 is performed to check how high-frequency components the video frame has. Various methods are conceivable for this, examples of which will be described with reference to FIG.

FIG. 15 is a schematic diagram showing a DCT coefficient block. As shown in FIG. 15, the DCT coefficient block is divided into zones 1 to 4 by horizontal and vertical frequency components. The DCT coefficient analyzer 401 counts which zone the highest frequency component belongs to for each DCT coefficient block input from the DCT converter 105 over one frame, and sends the analysis result to the resolution determiner 402. Output. For example, when the analysis method is the above method, the count number of each zone is output. The resolution determiner 402
Based on the analysis result of the DCT coefficient analyzer 401, how to change the resolution of the next video frame is determined.

For example, in the case where the analysis method in the DCT coefficient analyzer 401 is the above method, a method of determining the resolution will be described with reference to FIG. In this case, the count number is compared with a predetermined threshold value in the order of zones 4, 3, 2, and 1, and the resolution is determined according to the zone where the count number is equal to or greater than the threshold value. Here, the resolution corresponding to each zone increases as the zones become 1, 2, 3, and 4.

The resolution determined by the resolution determiner 402 is output to the resolution converter 103. As a result, the resolution converter 103 performs resolution conversion of the next video frame.

As described above, the image coding apparatus according to the fourth embodiment analyzes a DCT coefficient block of an input video, and determines a resolution based on a distribution of a high frequency component of the DCT coefficient. In this case, the resolution of the input video is determined so that the higher the number of blocks having high-frequency components, the higher the resolution. Then, the next frame of the input video is converted to the determined resolution and encoded. Therefore, by using the image encoding device of the fourth embodiment, it is possible to encode an input video with a resolution corresponding to the frequency component of the input video. That is, since the number of blocks to be coded can be reduced according to the frequency components of the input video, the amount of generated codes can be greatly reduced with little image quality deterioration.

In the present embodiment, the case where the analysis is performed by dividing the DCT coefficient block into four zones has been described. However, the number of zones may be another value.

In this embodiment, the case where the DCT coefficient block is analyzed over one frame period and the result is used to determine the resolution of the next frame has been described. However, the analysis period of the DCT coefficient block is different. It may be a period.

FIG. 17 shows a modification of FIG. FIG. 17 differs from FIG. 14 in that the input video signal is an analog signal. The resolution converter 103 in FIG. 14 is replaced by a sampler 103a. According to FIG. 17, the resolution determined by the resolution determiner 402 is output to the sampler 103a. As a result, the sampler 103a samples the input video frame at a sampling frequency corresponding to the determined resolution, and supplies the digital video signal to the compression encoder 110.

(Fifth Embodiment) A fifth embodiment of the present invention will be described with reference to FIG. FIG. 18 shows an average quantization step calculator 501, a generated code amount counter 502, a resolution determiner 503, a resolution converter 103, and a compression encoder 11.
FIG. 2 is a block diagram of an image encoding device composed of 0.
The compression encoder 110 includes a block converter 104, a DCT transformer 105, a quantizer 106, and a variable length encoder 107.

An input video signal, which is a digital signal, is first input to a resolution converter 103. Then, the resolution converter 103 converts the resolution of the video frame based on the resolution determined by a method described later. Resolution converter 10
3 is output to the blocker 104
And is divided into blocks. Then, each block is input to the DCT converter 105. DCT converter 10
In step 5, two-dimensional DCT is performed on each block to convert the blocks into DCT coefficients. The DCT coefficients output from the DCT transformer 105 are input to a quantizer 106, where the coefficients are rounded by a quantization process. The output of the quantizer 106 is input to a variable-length encoder 107 and subjected to variable-length encoding.
Output as a code string.

On the other hand, the generated code amount of the code string generated by the variable length encoder 107 is input to the generated code amount counter 502. The generated code amount counter 502 calculates the cumulative sum of the generated code amounts for a predetermined period (for example, one frame). The cumulative sum of the generated code amounts for a predetermined period is output to the resolution determiner 503.

The quantization step used by the quantizer 106 for quantizing the DCT coefficient is input to the average quantization step calculator 501. The average quantization step calculator 501 calculates an average value of quantization steps in a predetermined period (for example, one frame). The obtained average value of the quantization step is output to the resolution determiner 503.

The resolution determiner 503 calculates the cumulative sum of the generated code amount for a predetermined period obtained by the generated code amount counter 502,
The product of the average quantization step calculator 501 and the average value of the quantization steps for a predetermined period is calculated. The value of this product is defined as complexity X. Complexity X is a value indicating the degree of difficulty in encoding a video, and the larger it is, the more difficult it is to encode (when encoded in the same quantization step,
(The generated code amount increases).

The resolution determiner 503 determines the resolution according to the flowchart shown in FIG. Here, FIG.
It is assumed that the resolution is determined to one of the three resolutions A, B, and C shown in FIGS. First, a complexity X for a predetermined period is obtained. Then, X is compared with a predetermined threshold value TH5. If X is TH5
If smaller, the resolution is determined to be resolution A. X
Is greater than or equal to the threshold value TH5, X is compared with a predetermined threshold value TH6. Where TH5 <
TH6. If X is smaller than TH6, the resolution is determined to be resolution B. If X is greater than or equal to threshold value TH6, the resolution is determined to be resolution C.

The resolution determined by the resolution determiner 503 is output to the resolution converter 103. As a result, the resolution converter 103 performs resolution conversion of the next video frame.

As described above, the image coding apparatus according to the fifth embodiment calculates the complexity from the generated code amount and the quantization step in the past predetermined period, and determines the resolution according to the magnitude of the complexity. I do. In this case,
The resolution of the input video is determined so that the higher the complexity, that is, the higher the video encoding difficulty, the higher the resolution. Then, the next frame of the input video is converted to the determined resolution and encoded. Therefore,
By using the image encoding device of the fifth embodiment,
The input video can be encoded at a resolution corresponding to the encoding difficulty of the input video. That is, since the number of blocks to be encoded can be reduced in accordance with the degree of difficulty of encoding the input video, the amount of generated codes can be greatly reduced with little deterioration in image quality.

In this embodiment, the case where the predetermined period for obtaining the accumulated value of the generated code amount and the average value of the quantization step is one frame is described, but this may be a different period.

In this embodiment, two threshold values TH5 and TH6 are set for the complexity, and one of the three resolutions (A, B, C) is selected. As described above, the number of thresholds and the number of types of resolutions may be different from those in the present embodiment.

FIG. 20 shows a modification of FIG. FIG. 20 differs from FIG. 18 in that the input video signal is an analog signal. The resolution converter 103 in FIG. 18 is replaced by a sampler 103a. According to FIG. 20, the resolution determined by the resolution determiner 503 is output to the sampler 103a. As a result, the sampler 103a samples the input video frame at a sampling frequency corresponding to the determined resolution, and supplies the digital video signal to the compression encoder 110.

In the first to fifth embodiments, the case of intra-frame coding has been described. However, it goes without saying that resolution conversion can be similarly performed in the case of inter-frame predictive coding.

In the first to fifth embodiments, the case where the resolution is lowered in the horizontal direction has been described. However, even if the resolution is lowered in the vertical direction, the resolution is lowered in both the horizontal direction and the vertical direction. No problem.

Further, in the first to fifth embodiments, the description has been made on the assumption that MPEG2 is used as the encoding method. However, this may be another encoding method.

The following sixth to ninth embodiments are particularly directed to a DVD-RAM which is an optical disk capable of rewriting an AV stream obtained by compression-encoding a video signal and an audio signal.
It is suitable for storing in a storage medium such as a hard disk drive.

(Sixth Embodiment) A sixth embodiment of the present invention will be described with reference to FIG. FIG. 21 shows a compression encoder 610, a code amount controller 620, and a resolution changing unit 6.
FIG. 2 is a block diagram of an image encoding device constituted by 30.

The compression encoder 610 is composed of a block generator, a DCT transformer, a quantizer, and a variable length encoder (not shown), and has a variable bit rate corresponding to a variable allocated bit amount Bt. It has a function of compression-encoding a video signal so as to output a bit stream. The allocated bit amount Bt defines a quantization step in the compression encoder 610.

The code amount controller 620 is a means for controlling the generated code amount, and includes a generated bit amount detector 621 for detecting the generated bit amount Bg of the coded bit stream every predetermined period, A cumulative error calculator 622 for sequentially accumulating a difference between the generated generated bit amount Bg and a given average target code amount Ba and calculating a cumulative error D, and a bit amount Bt allocated to the compression encoder 610. And a coding difficulty detector for detecting a coding difficulty (complexity) X indicating the complexity of the scene represented by the video signal based on the generated bit amount Bg. 624 and a difficulty change detector 625 for detecting a scene change based on a change in the encoding difficulty X. Encoding difficulty X
Is calculated based on the generated bit amount Bg and the quantization step at that time. ΔX is a difficulty level change detection signal supplied to the bit allocator 623 when a scene is switched. Di represents the initial value of the accumulated error, and Dmax represents the maximum value of the accumulated error. Here, the difference between the accumulated error D and its maximum value Dmax is defined as a margin M. The accumulated error D and the encoding difficulty X are, for example, 1 GOP (group
of pictures) calculated for each period.

The resolution changing unit 630 changes the resolution of the input video signal, which is a digital signal, and supplies it to the compression encoder 610.
Cumulative difficulty calculator 632 for accumulating encoding difficulty X and calculating cumulative difficulty AX over ＯＰ40 GOP period
And a resolution determiner 633 for determining a resolution according to at least one of the magnitude of the accumulated error D and the magnitude of the accumulated difficulty AX. The resolution converter 631 determines that the original resolution of the input video signal is horizontal 70, for example.
When 4 pixels × vertical 480 pixels, 704
It has a high-resolution mode for outputting at × 480 and a low-resolution mode for outputting at 352 × 480 by halving the horizontal resolution. The resolution changing unit 630 outputs 3 when the cumulative error D exceeds a predetermined threshold value Dth_U.
A video signal is output in a low resolution mode of 52 × 480, and thereafter, a threshold Dt in which the accumulated error D is smaller than Dth_U
h_L, the resolution of the video signal is reduced to the original high resolution (7
04 × 480). When the cumulative difficulty AX exceeds a predetermined threshold AXth_U and increases, the resolution changing unit 630 determines that the cumulative difficulty AX becomes 352 ×
A video signal is output in the low resolution mode of 480, and then the threshold value AX at which the cumulative difficulty level AX is smaller than AXth_U
It also has a function of returning the resolution of the video signal to the original high resolution (704 × 480) when it falls below th_L.

The bit allocator 623 newly determines the allocated bit amount Bt according to the encoding difficulty X immediately after the scene change, and performs feedback so as to sequentially update the immediately preceding allocated bit amount Bt within the same scene. Perform control. That is, in the frame immediately after the scene switching, the bit allocator 623 temporarily allocates a bit amount exceeding the average target amount Ba to a complex scene having the encoding difficulty X exceeding the average difficulty Xa. If the bit amount is determined and the accumulated error D exceeds the predetermined amount at this point, the provisionally assigned bit amount is set according to the magnitude of the accumulated error D so that the temporarily assigned bit amount is reduced to the maximum average target amount Ba. To determine the allocated bit amount Bt. However, a bit amount Bt smaller than the average target amount Ba is assigned to a monotonous scene indicating the encoding difficulty X lower than the average difficulty Xa. For frames in the same scene, the bit allocator 623
The immediately preceding allocated bit amount Bt is sequentially updated so that the accumulated error D does not exceed a predetermined maximum amount Dmax.
At this time, the bit allocator 623 determines the updated allocated bit amount Bt such that the allocated bit amount Bt approaches the average target amount Ba as the cumulative error D approaches the maximum amount Dmax.

FIGS. 22 (a) to 22 (c) show an operation example in the case where the resolution is not changed in the image coding apparatus of FIG. 21, and it is assumed that scenes with high coding difficulty continue. I assume. FIG. 22A shows the encoding difficulty X.
FIG. 22B shows the allocation bit amount Bt of FIG.
(C) shows the time transition of each of the accumulated errors D. Here, it is assumed that the initial value Di = 0 of the accumulated error, and therefore the margin M in the initial state is set to Dmax.

According to FIGS. 22 (b) and 22 (c),
As the margin M decreases, the allocated bit amount Bt gradually decreases. As it is, the accumulated error D becomes Dma
When x reaches x, that is, when the margin M becomes zero, only the average target amount Ba is given even in a scene where the encoding difficulty X is high. Therefore, since the total code amount does not exceed the predetermined amount, the recording time within the predetermined amount is guaranteed, and the image quality equivalent to the fixed bit rate can be guaranteed even when the margin amount M disappears.

However, since the bit rate becomes equal to the fixed bit rate control after the margin M disappears, it is not possible to perform the variable bit rate control according to the encoding difficulty of each scene of the input video signal. Therefore, if a video signal having a high degree of coding difficulty is input in such a state, a sufficient coding amount cannot be provided, and coding noise peculiar to DCT coding processing such as block noise and mosquito noise is generated. The image quality is significantly reduced.

Therefore, in the present embodiment, the resolution of the input video signal is reduced, for example, when the margin M decreases and it becomes difficult to continue the variable bit rate control. In this way, the generated code amount is suppressed and slightly changed, so that the variable bit rate control can be continuously performed.

For example, the cumulative error D increases and approaches the maximum cumulative error Dmax, and the margin M (= Dmax−D)
Becomes smaller, the resolution of the input video signal is reduced and the encoding difficulty X is reduced, so that the generated bit amount Bg can be suppressed lower than when encoding is performed at the original resolution. It is possible to suppress the amount M from further decreasing. Further, since the margin M has not completely disappeared and the variable bit rate control can be continuously performed according to the encoding difficulty of the input video signal, DCT encoding processing such as block noise and mosquito noise is performed. This is effective in suppressing specific coding noise.

FIGS. 23 (a) to 23 (c) show the effect of changing the resolution based on the accumulated error D in the image coding apparatus of FIG. Here, Dth_U = Dmax
× 3/4, and Dth_L = Dmax / 2. The model of the input video signal is shown in FIG.
Same as (a). However, between times t1 and t2 when the conversion is performed to the low resolution and the encoding is performed, the encoding difficulty X detected from the generated bit amount Bg is suppressed to about 70% as compared with the original resolution. It was taken.

According to FIGS. 23 (a) and 23 (c),
Between time 0 and time t1, the input video signal continues a video scene with a high encoding difficulty X, and the accumulated error D monotonically increases, and exceeds the threshold value Dth_U near time t1.
At this time t1, the resolution of the input video signal is switched from the original 704 × 480 to the low-resolution mode 352 × 480 by the resolution changing unit 630.

The encoding difficulty X of the input video signal in the low resolution mode is relatively lower than that in the high resolution mode, and shows a transition as shown in FIG. . During this period, the accumulated error D generally changes in a decreasing direction, and falls below the threshold value Dth_L near time t2. At this time, the resolution changing unit 630 returns from the low-resolution mode of 352 × 480 to the original 704 × 480.

FIG. 23 (c) also shows the transition of the accumulated error D when the resolution is not changed.
In this embodiment, the accumulated error D in this embodiment has a sufficient margin M as compared with FIG. 22C. In the bit allocation after this time, the accumulated error D depends on the encoding difficulty X of the input video signal. It can be seen from the transition of the allocated bit amount Bt in FIG. 23B that the optimum allocated bit Bt is obtained.

FIGS. 24A to 24C show the effect of changing the resolution based on the cumulative difficulty level AX in the image coding apparatus of FIG. It can be seen that the same effect as in the case of changing the resolution based on the accumulated error D can be obtained.

As described above, in the image coding apparatus according to the sixth embodiment, the resolution of the input video signal is adaptively changed according to the magnitude of the cumulative error D or the cumulative difficulty AX, and the coding of the input video signal is performed. This makes it possible to continuously perform variable bit rate control according to the degree of difficulty, and is effective in suppressing coding noise. Moreover, by providing two types of thresholds, that is, a threshold value for lowering the resolution and a threshold value for restoring the lowered resolution, and having a hysteresis characteristic, it is possible to suppress frequent switching of the resolution. In addition, in a scene where movement is intense, a decrease in resolution is difficult to be recognized by human eyes, so that an adverse effect due to the decrease in resolution is minimized.

Although the resolution converter 631 of this embodiment switches between two resolutions, it may switch between three or more resolutions.
The determination of switching the resolution may be made by appropriately combining the magnitude of the accumulated error D and the magnitude of the accumulated difficulty AX. For example, switching to a lower resolution requires a cumulative error D
Is determined based on the magnitude of the cumulative difficulty AX.
In addition, there is no problem even if both of the two determination conditions are satisfied or one of the two conditions is satisfied. It is also possible to replace the resolution converter 631 with an analog input sampler.

(Seventh Embodiment) A seventh embodiment of the present invention will be described with reference to FIG. The present embodiment provides an image coding apparatus having a resolution switching unit that makes it difficult for the switching point of the resolution to be recognized by human eyes. FIG.
In, 640 represents a still scene detector, and 641 represents a scene change detector.

The still scene detector 640 has a function of detecting a still scene from an input video signal, and notifies the resolution changing unit 630 when a still scene is detected. The scene change detector 641 has a function of detecting a scene change, which is a scene change of an input video signal, and notifies the resolution change unit 630 when a scene change is detected. Detection of a still scene and detection of a scene change include a method using a feature amount such as a pixel difference between frames, a change in luminance and a color difference level, and any method may be used.

The resolution change unit 630 uses the information provided from the code amount controller 620 as a determination condition, and when it is necessary to switch the resolution, does not immediately switch the resolution, but uses a still scene detector. Still scene information or scene change detector 64 notified from 640
Wait for the scene change information notified from 1 and switch in synchronization with them.

As described above, in the image encoding apparatus according to the seventh embodiment, when the resolution needs to be switched, the switching is performed in synchronization with a still scene or a scene change. It is possible to make it difficult for the eyes to discriminate.

In this embodiment, both the still scene and the scene change are used, but there is no problem if only one of them is used.

(Eighth Embodiment) An eighth embodiment of the present invention will be described with reference to FIG. In the present embodiment, the generated AV stream is stored in a DVD-RAM together with its management information.
It is an object of the present invention to provide an image encoding apparatus that can easily match the AV stream with the management information when writing to the storage medium 654 such as the storage medium 654 and does not make the size of the management information too large. FIG. 26 is a block diagram of an image encoding device including a compression encoder 610, a code amount controller 620, a resolution change unit 630, and a parameter management unit 653. The compression encoder 610 includes a GOP structure determiner 650, a VOB switch 651, and an MPEG encoder 652.

In the MPEG2 standard, encoded bit streams of video signals having different resolutions are distinguished from each other, and are not allowed to be treated as the same sequence. That is, the encoded bit stream must be generated as a separate sequence before and after the resolution change point.

<Division of GOP and Division of Sequence> In FIG. 26, a GOP structure determiner 650 has a function of determining the GOP structure in MPEG encoding in the MPEG encoder 652, and the resolution changing unit 63
0 in synchronization with the resolution change being implemented,
It has a function to disconnect P and start a new GOP. FIG.
(A) to FIG. 27 (c) show the operation at this time. FIG.
FIG. 27 (b) and FIG. 27 (b) show a state in which a normal operation is performed at M = 3, N = 15 (M is a period in which an I or P frame appears, and N is the number of frames of a GOP). 27 (c)
Is G when the resolution is changed before frame 10.
Each of the OP structure changes is shown.

In FIG. 27B, a GOP newly started after the resolution change is set to a closed GOP that does not refer to the previous GOP.
OP. More specifically, two frames (frames 10 and 11) preceding the first I frame in the display order of a GOP newly started after the resolution change are replaced with P frames (frame 9) belonging to the previous GOP at the time of predictive encoding. Is encoded so as not to refer to the G before and after the resolution change.
There is no correlation between OPs.

In FIG. 27C, the GOP newly started after the resolution change is started from the I frame in the display order, so that there is no correlation with the immediately preceding GOP.

The MPEG2 stream generated by the image encoding device including the GOP structure determiner 650 includes:
GOP can be divided at the resolution change point,
In addition, the GOPs before and after the resolution change point have no correlation with each other. Therefore, a sequence end code (Sequenc) is placed at the end of the bit stream of the GOP immediately before the resolution change.
e_End_Code) and insert a new GOP
At the beginning of the sequence start code (Sequence_S
start_Code) can be inserted, and it can be easily divided into two independent coded stream sequences before and after the resolution change point. It should be noted that there is no particular problem if the sequence start code is inserted at the beginning of all GOPs.

<Division of VOB> An AV in which a coded bit stream of a series of video and audio is multiplexed
When a stream is stored in a storage medium such as a DVD-RAM, a VO
Consider B (video object). The VOB has parameters such as the frame frequency, resolution, and aspect ratio of video data as its management information and the coding mode, number of channels, and coding rate of audio data. Are referred to by the decoder when decoding is performed.

When the resolution of the video signal is changed by the resolution changing unit 630, the parameters of the video data change, and VOB division is required.

Referring to FIG. 26, VOB switch 651
Is a function for giving the resolution change permission to the resolution change unit 630 and, when the resolution is actually changed, a notification of the VOB division execution together with the changed video parameters to the parameter management unit 653. It has both functions of performing The parameter management unit 653 has a function of writing each parameter corresponding to the created AV stream to the storage medium 654.

The MPEG2 stream generated by the image coding device including the VOB switch 651 is:
The VOB can be divided at the resolution change point. Therefore, it is possible to generate an encoded bit stream having independent GOPs and different VOBs before and after the resolution change point. Further, the end of the VOB can be clearly indicated by the sequence end code described in the sequence division.

<Restriction on VOB division> Since the management information is steadily increased by dividing the VOB, for example, the maximum number of VOBs (VOB_Max) that can be recorded on one DVD-RAM disk must be determined in advance. is there.

The VOB switch 651 outputs the number of VOBs (V
OB_Num), and when a predetermined maximum number of VOBs VOB_Max is reached, switching of VOBs thereafter is prohibited. This makes it possible to suppress the management information size to a certain size or less.

Further, from the total capacity (Volume) of the recording medium in the storage medium 654 and the target average rate (Rs) of the coded bit stream, a total recording time (Ts) is obtained by Ts = Volume / Rs, and this is calculated as VOB. Divide by the maximum number VOB_Max, and find the minimum VOB time (Tmin) by Tmin = Ts / VOB_Max. After a new VOB is started, the VOB switch 651
During the min period, switching of the VOB again is prohibited. By doing so, the total number of VOBs generated in the present image encoding device is the maximum number of VOBs VOB_Ma.
x can be limited so as not to exceed x, and the management information size can be suppressed to a certain size or less.

For example, a target average rate Rs = 5 in a DVD-RAM having a volume of 4.7 gigabytes.
When recording video / audio at Mbps, Ts = 7520
Seconds. At this time, the maximum number of VOBs VOB_Max is set to 9
If it is 99, Tmin = 7.52 seconds. That is,
VOB switching is not performed again for 7.52 seconds after VOB switching.

When a plurality of VOBs having different target average rates are recorded in one recording medium, the total recording time Ts is obtained at the lowest allowable target average rate, and the minimum VOB time Tmin is calculated based on the total recording time Ts. Is obtained, it is possible to sufficiently guarantee the limitation of the total number of VOBs in the entire recording medium.

When an AV stream is to be recorded in addition to a recording medium on which an AV stream has already been recorded, the free space of the recording medium (Volume_res
t) and the target average rate (R
s), the remaining recording time (Ts_rest) is obtained from Ts_rest = Volume_rest / Rs, and
VO already recorded from the maximum number of OBs VOB_Max
The minimum VOB time Tmin is obtained from Tmin = Ts_rest / VOB_avail by dividing by the remainder (VOB_avail) after subtracting the B number. After a new VOB is started, the VOB switch 651
During the period of n, switching of the VOB again is prohibited.

For example, a target average rate Rs = 5 in a DVD-RAM having a volume of 4.7 gigabytes.
It is assumed that video and audio are recorded at Mbps. Assuming that video and audio for one hour have already been recorded in 100 VOBs, Volume_rest = 2.45 gigabytes, Ts_rest = 3920 seconds, VOB_ava
il = 899, and Tmin = 4.36 seconds.
In other words, for 4.36 seconds after VOB switching,
Switch is not performed.

VOB_avail may be a value obtained by subtracting the number of VOBs already used from VOB_Max as described above, or a value obtained by multiplying VOB_Max by the ratio of this value and the remaining recording time to the total recording time ( In the above example, 999 × 3920/7520 = 520) may be compared, and the smaller value may be used (520).

The counting of the number of VOBs is performed not only by the VOB division by changing the resolution but also by other factors, such as the case where the VOB is divided by an external command.

(Ninth Embodiment) A ninth embodiment of the present invention will be described with reference to FIG. In FIG. 28, a selector 660 has a function of returning a resolution to be used in response to a resolution change request from the resolution change unit 630. The selector 660 has a table of resolutions that can be used for each target average rate with the target average rate of the coded bit stream as an input. For example, a resolution without hatching in the table shown in FIG. 28 can be selected. By limiting the range of resolutions selected in this way, the resolution that matches the target average rate will be selected, and if the resolution is set too high and the code amount will be too large, the resolution will be set too low and the image quality will be set too low. Can be prevented.

FIG. 29 shows a modification of FIG. The resolution used here depends on the designation from the operator. The input device 661 of FIG.
To the available resolutions. For example, if there is an input by the operator that only the resolution of 704 × 480 can be used, the selector 660 does not select a resolution other than 704 × 480 even if there is a resolution change request from the resolution change unit 630. No resolution change occurs.
If two resolutions of 704 × 480 and 480 × 480 are permitted by the operator, the selector 660
Then, the resolution change unit 630 is instructed to switch the resolution between the two resolutions. By specifying the resolution that can be used according to the input from the operator in this way, it is possible to select the resolution according to the preference of the operator. The input from the operator is
Instead of specifying a specific resolution, it is also possible to specify only whether or not to permit the resolution change.

As described above, according to the first embodiment of the present invention, the frequency components included in the input video signal are analyzed,
The resolution is determined based on which frequency band the highest frequency component is in. In this case, the resolution of the input video signal is determined so that the higher the frequency component, the higher the resolution. Then, after converting the input video signal to the determined resolution or performing sampling at a sampling frequency corresponding to the resolution, compression encoding is performed.

Further, according to the second embodiment of the present invention,
A low-pass filter having a different cutoff frequency is applied to the input video signal, and the output of the filter is compared with a predetermined threshold. Then, the resolution of the input video signal is determined so that the higher the cutoff frequency of the filter that outputs an output larger than the threshold, the higher the resolution. Then, after converting the input video signal to the determined resolution or performing sampling at a sampling frequency corresponding to the resolution, compression encoding is performed.

Further, according to the third embodiment of the present invention,
The activity of the input video signal is calculated, and the resolution is determined according to the value of the activity. In this case,
The resolution of the input video signal is determined so that the higher the activity, the higher the resolution. Then, after converting the input video signal to the determined resolution or performing sampling at a sampling frequency corresponding to the resolution, compression encoding is performed.

Further, according to the fourth embodiment of the present invention,
The DCT coefficient block of the input video signal is analyzed and its DC
The resolution is determined by the distribution of the high frequency component of the T coefficient. In this case, the resolution of the input video signal is determined so that the higher the number of blocks having high frequency components, the higher the resolution. Then, after the next frame of the input video signal is converted to the determined resolution or sampled at a sampling frequency according to the resolution, compression encoding is performed.

Further, according to the fifth embodiment of the present invention,
The complexity is calculated from the generated code amount and the quantization step in the past predetermined period, and the resolution is determined according to the magnitude of the complexity. In this case, the resolution of the input video signal is determined so that the higher the complexity, that is, the higher the encoding difficulty, the higher the resolution. Then, after the next frame of the input video signal is converted to the determined resolution or sampled at a sampling frequency according to the resolution, compression encoding is performed.

Therefore, by using the image encoding apparatuses according to the first to fifth embodiments of the present invention, it is possible to encode an input video with a resolution according to characteristics of the input video such as frequency components. That is, since the number of blocks to be coded can be reduced according to the characteristics of the input video signal, the amount of generated codes can be greatly reduced with little image quality deterioration. Further, the input video signal can be encoded at a resolution corresponding to the encoding difficulty of the input video. Further, since the resolution conversion is performed according to the characteristics of the input video signal, visual degradation hardly occurs when the resolution is switched.

In particular, according to the first to third embodiments of the present invention, since the resolution is determined in accordance with the characteristics of the encoding target frame itself, for example, when a scene change occurs in the current frame, Even if there is, image quality does not deteriorate.

According to the sixth embodiment of the present invention,
When generating a variable rate coded bit stream while guaranteeing the recording time in real time, if it becomes difficult to assign variable bits according to the encoding difficulty of the input video signal, the resolution of the input video signal is reduced. It is possible to continue variable bit allocation according to the degree of difficulty of encoding, to reduce encoding noise peculiar to DCT encoding processing such as block noise and mosquito noise, and to improve image quality.

According to the seventh embodiment of the present invention,
The switching of the resolution is synchronized with the scene change or the stillness of the input video signal, so that the switching of the resolution is hardly recognized by human eyes.

According to the eighth embodiment of the present invention,
The generated AV stream is stored in a DVD together with its management information.
-When writing to a storage medium such as a RAM, a VOB, which is a management unit of an AV stream, is switched in synchronization with a change in resolution, thereby facilitating consistency between the AV stream and the management information. It is possible to suppress the number of resolution switching to a specified number or less so as not to become too large.

Further, according to the ninth embodiment of the present invention,
By setting a limit on the resolution to be used for switching the resolution in accordance with the target average rate and the designation from the operator, there is an effect that encoding can be performed with an image quality according to the encoding conditions and the preference of the operator.

[0114]

As described above, according to the present invention, since the resolution is changed in accordance with the characteristics of the input video signal, the amount of generated codes can be greatly reduced while minimizing image quality deterioration. An encoding device can be provided.

[Brief description of the drawings]

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

FIGS. 2A to 2D are schematic diagrams illustrating examples of selectable resolutions.

FIG. 3 is a flowchart of resolution determination in the image encoding device of FIG. 1;

FIGS. 4A and 4B are graphs for explaining the definition of the highest frequency component in FIG. 3;

FIG. 5 is a block diagram showing a modified example of the image encoding device of FIG. 1;

FIG. 6 is a block diagram of an image encoding device according to a second embodiment of the present invention.

FIG. 7 is a graph showing an example of frequency characteristics of each of two filters in FIG. 6;

FIG. 8 is a flowchart of resolution determination in the image encoding device of FIG. 6;

FIG. 9 is a block diagram illustrating a modification of the image encoding device of FIG. 6;

FIG. 10 is a block diagram of an image encoding device according to a third embodiment of the present invention.

FIG. 11 is a flowchart of resolution determination in the image encoding device of FIG. 10;

FIG. 12 is a detailed flowchart of an activity determination step in FIG. 11;

FIG. 13 is a block diagram illustrating a modification of the image encoding device in FIG. 10;

FIG. 14 is a block diagram of an image encoding device according to a fourth embodiment of the present invention.

FIG. 15 is a schematic diagram of a DCT coefficient block for explaining the operation of the DCT coefficient analyzer in FIG.

FIG. 16 is a flowchart of resolution determination in the image encoding device of FIG. 14;

FIG. 17 is a block diagram illustrating a modification of the image encoding device in FIG. 14;

FIG. 18 is a block diagram of an image encoding device according to a fifth embodiment of the present invention.

19 is a flowchart of resolution determination in the image encoding device of FIG.

FIG. 20 is a block diagram illustrating a modification of the image encoding device in FIG. 18;

FIG. 21 is a block diagram of an image encoding device according to a sixth embodiment of the present invention.

22A and 22B are diagrams illustrating an operation example when the resolution is not changed in the image encoding device in FIG. 21, where FIG. 22A illustrates encoding difficulty X, FIG.
(C) shows the time transition of each of the accumulated errors D.

23A and 23B are diagrams illustrating an operation example in the case where the resolution is changed in the image encoding device in FIG. 21, where FIG. 23A illustrates an encoding difficulty X, FIG. 23B illustrates an assigned bit amount Bt, and FIG. Indicates the time transition of each of the accumulated errors D.

24A and 24B are diagrams illustrating an operation example of the image encoding device in FIG. 21 when the resolution is changed, where FIG. 24A illustrates the encoding difficulty X, FIG. 24B illustrates the assigned bit amount Bt, and FIG. Indicates the time transition of each of the cumulative difficulty levels AX.

FIG. 25 is a block diagram of an image encoding device according to a seventh embodiment of the present invention.

FIG. 26 is a block diagram of an image encoding device according to an eighth embodiment of the present invention.

27A and 27B are diagrams illustrating an example of a GOP structure determined by a GOP structure determiner in FIG. 26. FIG.
(B) shows a case where the newly started GOP is a closed GOP, and (c) shows a case where the newly started GOP is an I-frame start GOP.

FIG. 28 is a block diagram of an image encoding device according to a ninth embodiment of the present invention.

FIG. 29 is a block diagram illustrating a modification of the image encoding device in FIG. 28.

[Explanation of symbols]

 REFERENCE SIGNS LIST 100 delay unit 101 frequency band analyzer 102 resolution determiner 103 resolution converter 103 a sampler 104 blocker 105 DCT converter 106 quantizer 107 variable length encoder 110 compression encoder 201, 202 filter 204, 205 Output value comparator 207 Resolution determiner 301 Activity measurer 302 Resolution determiner 401 DCT coefficient analyzer 402 Resolution determiner 501 Average quantization step calculator 502 Generated code amount counter 503 Resolution determiner 610 Compression encoder 620 Code amount Controller 621 Generated bit amount detector 622 Cumulative error calculator 623 Bit allocator 624 Encoding difficulty detector 625 Difficulty change detector 630 Resolution change unit 631 Resolution converter 632 Cumulative difficulty calculator 633 Resolution determiner 40 Still scene detector 641 Scene change detector 650 GOP structure determiner 651 VOB switch 652 MPEG encoder 653 Parameter management unit 654 Storage media 660 Selector 661 Input device AX Cumulative difficulty Ba Average target code amount Bg Generated bit amount Bt Allocated bit amount D Cumulative error Di Initial cumulative error Dmax Maximum cumulative error X Encoding difficulty ΔX Difficulty change detection signal

 ──────────────────────────────────────────────────続 き Continuing from the front page (72) Inventor Hideki Fukuda 1006 Kazuma Kadoma, Osaka Prefecture Matsushita Electric Industrial Co., Ltd. 72) Inventor Iwasaki Shiro 1006 Kazuma Kadoma, Osaka Prefecture F-term (reference) 5C059 KK01 LB00 MA00 MA23 MC11 ME01 PP04 SS01 SS06 SS11 TA06 TB04 TC00 TC18 TC19 TD01 TD04 TD11 UA11 UA11 UA13 UA34 5J064 AA01 BA09 BA13 BA16 BC16 BC24 BD02

Claims (46)

[Claims] 1. A frequency band analyzing means for outputting a frequency component value of an input video signal, and detecting a highest frequency of the frequency component value output from the frequency band analyzing means, wherein the highest frequency is predetermined. Resolution determining means for determining a resolution according to which of the plurality of frequency bands belongs, resolution changing means for changing and outputting the resolution of the input video signal according to the determination of the resolution determining means, An image encoding apparatus, comprising: encoding means for compression-encoding the video signal output from the resolution changing means. 2. The image encoding apparatus according to claim 1, wherein the input video signal is a digital signal, and the resolution changing unit changes the number of pixels of the input video signal according to the determined resolution. An image encoding device comprising a resolution conversion unit for outputting. 3. The image encoding device according to claim 1, wherein the input video signal is an analog signal, and the resolution changing unit samples the input video signal at a sampling frequency according to the determined resolution. An image coding apparatus comprising a sampling means for converting a digital video signal into a digital video signal and outputting the digital video signal. 4. The image encoding device according to claim 1, wherein the resolution determining unit is configured to provide a higher resolution to the predetermined plurality of frequency bands as the highest frequency belongs to a higher frequency band. An image encoding apparatus for determining a resolution. 5. The image encoding apparatus according to claim 1, wherein the resolution determining means determines a threshold value of the frequency component value which is an output of the frequency band analysis means. An image coding apparatus for detecting a highest frequency which is equal to or higher than a value. 6. The image encoding apparatus according to claim 1, wherein the highest resolution determined by said resolution determining means is the same as the resolution of said input video signal. 7. A plurality of filter means for performing frequency filtering having different frequency characteristics on an input video signal, and comparing output values of the plurality of filter means with a predetermined threshold value, Comparing means for outputting a result; resolution determining means for determining a resolution from a comparison result of the comparing means; and resolution for changing and outputting the resolution of the input video signal in accordance with the determination of the resolution determining means. An image encoding apparatus, comprising: a changing unit; and an encoding unit for compressing and encoding a video signal output from the resolution changing unit. 8. The image encoding device according to claim 7, wherein the input video signal is a digital signal, and the resolution changing unit changes the number of pixels of the input video signal according to the determined resolution. An image encoding device comprising a resolution conversion unit for outputting. 9. The image encoding apparatus according to claim 7, wherein the input video signal is an analog signal, and the resolution changing unit samples the input video signal at a sampling frequency according to the determined resolution. An image coding apparatus comprising a sampling means for converting a digital video signal into a digital video signal and outputting the digital video signal. 10. The image encoding apparatus according to claim 7, wherein the plurality of filter units are low-pass filters each having a different cutoff frequency, and the resolution determining unit outputs an output of the plurality of filter units. An image encoding apparatus characterized in that, when there is a value larger than the predetermined threshold value, the resolution is determined such that the higher the cutoff frequency of the filter means, the higher the resolution. 11. The image encoding device according to claim 7, wherein the highest resolution determined by said resolution determining means is the same as the resolution of said input video signal. 12. An activity measuring means for obtaining a variance value of a pixel value with respect to an input video signal, and a resolution is determined by which of a predetermined range the variance value output from the activity measuring means belongs to. Resolution determining means for determining; resolution changing means for changing and outputting the resolution of the input video signal in accordance with the determination of the resolution determining means; and compression encoding of the video signal output from the resolution changing means. An image encoding apparatus comprising: an encoding unit for determining the resolution, wherein the resolution determining unit determines the resolution such that the higher the variance value output from the activity measuring unit, the higher the resolution. Image encoding device. 13. The image encoding device according to claim 12, wherein the input video signal is a digital signal, and wherein the resolution changing unit changes the number of pixels of the input video signal according to the determined resolution. An image encoding device comprising a resolution conversion unit for outputting. 14. The image encoding device according to claim 12, wherein the input video signal is an analog signal, and the resolution changing unit samples the input video signal at a sampling frequency according to the determined resolution. An image coding apparatus comprising a sampling means for converting a digital video signal into a digital video signal and outputting the digital video signal. 15. The image encoding apparatus according to claim 12, wherein said activity measuring means calculates and outputs a variance of pixel values for each predetermined unit, and said resolution determining means includes said activity measuring means. The average value or the maximum value of the variance value of each unit, which is the output of the unit, is determined over a predetermined video unit period, and the resolution is determined based on which of the predetermined ranges the average value or the maximum value belongs to. An image encoding device, characterized in that: 16. The image encoding apparatus according to claim 12, wherein the highest resolution determined by said resolution determining means is the same as the resolution of said input video signal. 17. A resolution changing means for changing and outputting a resolution of an input video signal, a block dividing means for dividing an output of the resolution changing means into blocks having a predetermined size, and Orthogonal transform means for orthogonally transforming the output of the means into a frequency domain; code string generating means for transforming the orthogonal transform coefficient output from the orthogonal transform means into a code string; and output of the orthogonal transform means. Orthogonal transform coefficient analyzing means for analyzing the frequency distribution of the orthogonal transform coefficient, and resolution determining means for determining the resolution used by the resolution changing means based on the analysis result output from the orthogonal transform coefficient analyzing means. An image encoding device, comprising: 18. The image encoding apparatus according to claim 17, wherein the input video signal is a digital signal, and wherein the resolution changing unit changes the number of pixels of the input video signal according to the determined resolution. An image encoding device comprising a resolution conversion unit for outputting. 19. The image encoding apparatus according to claim 17, wherein the input video signal is an analog signal, and the resolution changing unit samples the input video signal at a sampling frequency according to the determined resolution. An image coding apparatus comprising a sampling means for converting a digital video signal into a digital video signal and outputting the digital video signal. 20. The image coding apparatus according to claim 17, wherein the orthogonal transform coefficient analysis unit sets a plurality of zones obtained by dividing the orthogonal transform coefficient into frequency components, and outputs an output of the orthogonal transform unit. Counting which of the plurality of zones the highest frequency component of the orthogonal transform coefficient belongs to for each block over a predetermined video unit period, wherein the resolution determining means determines a resolution based on the counting result. An image encoding device characterized by the above-mentioned. 21. The image coding apparatus according to claim 20, wherein said resolution determining means determines the resolution such that the higher the count number of the high frequency component zone, the higher the resolution. Device. 22. The image encoding apparatus according to claim 17, wherein the highest resolution determined by said resolution determining means is the same as the resolution of said input video signal. 23. A resolution changing unit for changing the resolution of an input video signal and outputting the same, a block dividing unit for dividing an output of the resolution changing unit into blocks having a predetermined size, and the block division. Orthogonal transform means for orthogonally transforming the output of the means into a frequency domain; quantizing means for quantizing an orthogonal transform coefficient output from the orthogonal transform means using a variable quantization step; A code string generating means for converting an output of the means into a code string; a generated code amount counting means for obtaining and outputting a cumulative value of a code amount generated in the code string generating means within a predetermined period; Means for calculating and outputting an average value of the quantization step used in the quantization step within the predetermined period, and a generated code amount output from the generated code amount counting means. A complexity value, which is a product of the accumulated value and the average value of the quantization step output from the average quantization step calculation means, is determined, and the resolution used by the resolution changing means is determined based on the complexity value. And a resolution determining unit for determining the resolution, wherein the resolution determining unit is higher as the complexity value belongs to a larger value range among a plurality of predetermined ranges. An image coding apparatus characterized in that the resolution is determined so as to be the resolution. 24. The image encoding device according to claim 23, wherein the input video signal is a digital signal, and wherein the resolution changing unit changes the number of pixels of the input video signal according to the determined resolution. An image encoding device comprising a resolution conversion unit for outputting. 25. The image encoding device according to claim 23, wherein the input video signal is an analog signal, and the resolution changing unit samples the input video signal at a sampling frequency according to the determined resolution. An image coding apparatus comprising a sampling means for converting a digital video signal into a digital video signal and outputting the digital video signal. 26. The image encoding apparatus according to claim 23, wherein the highest resolution determined by said resolution determining means is the same as the resolution of said input video signal. 27. Compression encoding means for compressing and encoding an input video signal so as to output an encoded bit stream having a variable bit rate according to a variable allocated bit amount; A cumulative error calculating means for sequentially adding a difference between a generated bit amount for each predetermined period of the encoded bit stream and calculating a cumulative error; and Bit allocation means for controlling the amount of update, and resolution changing means for changing the resolution of the input video signal to low when the cumulative error exceeds a predetermined amount. Image encoding device. 28. The image coding apparatus according to claim 27, wherein said resolution changing means restores said resolution when said accumulated error falls below another specified amount smaller than said specified amount. Image encoding device. 29. The image encoding apparatus according to claim 27, wherein said resolution changing means selects one of three or more levels of resolution according to the magnitude of said accumulated error. Encoding device. 30. Compression encoding means for compressing and encoding an input video signal so as to output an encoded bit stream having a variable bit rate according to a variable allocated bit amount, and a given average target amount and A cumulative error calculating means for sequentially adding a difference between a generated bit amount for each predetermined period of the encoded bit stream and calculating a cumulative error; and Bit allocation means for controlling the amount of update; encoding difficulty detecting means for detecting the encoding difficulty indicating the complexity of the input video signal for each of the predetermined periods; and a plurality of encoding difficulty levels. Of the input video signal in accordance with at least one of the magnitude of the cumulative error and the magnitude of the cumulative difficulty. The image coding apparatus is characterized in that a resolution changing means for changing the Zodo. 31. The image encoding apparatus according to claim 30, wherein the resolution changing unit is configured to determine that the cumulative error exceeds a predetermined first specified amount or the cumulative difficulty is a predetermined amount. 2. An image coding apparatus according to claim 1, wherein the resolution of the input video signal is changed to a lower value when the specified amount exceeds 2. 32. The image encoding device according to claim 30, wherein the resolution changing unit is configured to determine that the accumulated error is smaller than a third designated amount smaller than the first designated amount, or the accumulated difficulty is smaller than the third designated amount. An image encoding apparatus, wherein the resolution is restored when a value falls below a fourth designated amount smaller than the second designated amount. 33. The image encoding device according to claim 30, wherein said resolution changing means changes the resolution of said input video signal to be low when said accumulated error exceeds a predetermined upper limit designated amount, and said accumulation difficulty is increased. An image encoding apparatus, wherein the resolution is restored when the degree falls below a predetermined lower limit specified amount. 34. A compression encoding unit for compressing and encoding an input video signal, a scene change detection unit for detecting a scene change of the input video signal, and the input video signal in synchronization with the scene change. And a resolution changing means for changing the resolution of the image. 35. A compression encoding unit for compressing and encoding an input video signal, a still scene detection unit for detecting a still scene of the input video signal, and the input unit when the still scene is detected. An image encoding apparatus comprising: a resolution changing unit configured to change a resolution of a video signal. 36. An input video signal is compression-encoded to an MPE.
An image encoding device comprising: a compression encoding unit for outputting a G stream; and a resolution changing unit for changing a resolution of the input video signal, wherein the compression encoding unit includes GO
GOP for determining P (group of pictures) structure
An image coding apparatus comprising a structure determining unit, wherein a GOP is switched at the same time as the resolution is changed. 37. The image encoding device according to claim 36, wherein the GOP structure determining unit is configured to determine the GOP immediately after the resolution is changed.
An image encoding apparatus, wherein an OP is a closed GOP that does not refer to a previous GOP. 38. The image encoding device according to claim 36, wherein the GOP structure determining unit determines a GOP immediately after the resolution is changed.
An image encoding apparatus characterized in that an OP starts from an I (intra) frame. 39. The image encoding apparatus according to claim 36, wherein the compression encoding unit is a VOB that is a set of the GOPs.
An image encoding device, comprising: a VOB switching unit for switching a video object, wherein the VOB is switched at the same time as the resolution is changed. 40. The image coding apparatus according to claim 39, wherein said VOB switching means does not switch the VOB when the number of VOBs reaches a predetermined maximum number of VOBs. apparatus. 41. The image encoding apparatus according to claim 39, wherein said VOB switching means does not switch the VOB until a specified time has elapsed after switching the VOB once. 42. The image encoding apparatus according to claim 41, wherein the specified time is a total recording time obtained from a total capacity of a recording medium on which the MPEG stream is to be recorded and a target average rate of the MPEG stream. Predetermined V
An image coding apparatus characterized by being obtained by dividing by the maximum number of OBs. 43. The image encoding apparatus according to claim 41, wherein the specified time is a remaining recording time obtained from a free space of a recording medium on which the MPEG stream is to be recorded and a target average rate of the MPEG stream. An image coding apparatus characterized by being obtained by dividing by a value obtained by subtracting the number of already recorded VOBs from a predetermined maximum number of VOBs. 44. The image encoding apparatus according to claim 39, wherein said VOB switching means inserts an end code indicating the end of the VOB into said MPEG stream when said VOB is switched. Device. 45. An MPE which compresses and encodes an input video signal.
Compression encoding means for outputting a G stream; resolution changing means for changing the resolution of the input video signal into a plurality of steps; and resolution used by the resolution changing means according to a given target average rate And a selecting means for selecting the image encoding device. 46. An MPE which compresses and encodes an input video signal.
Compression encoding means for outputting a G stream; resolution changing means for changing the resolution of the input video signal into a plurality of steps; input means for receiving input from an operator; Selecting means for selecting a resolution to be used by the resolution changing means in accordance with the input of the image encoding apparatus.