JP3286531B2 - Moving picture coding method and moving picture coding apparatus - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a moving picture coding method and a moving picture coding apparatus.

[0002]

2. Description of the Related Art MPEG (ISO11172, ISO1)
3818); In a moving image compression encoding method such as H.261, a discrete cosine transform (D / D) is performed on an input video signal (moving image signal) or difference data on which motion compensation has been performed.
Transformation to a frequency domain such as CT (Discrete Cosine Transform) is performed. Then, the frequency domain output is subjected to a quantization process using a predetermined quantization matrix. By performing variable length coding on this output, a coded output is obtained.

FIG. 13 shows a conventional MPEG encoding apparatus 10.
1 shows a block configuration. The input video signal (moving image signal) is divided into predetermined regions by the region dividing means 102. The divided area is a unit called a macro block (16 × 16 pixels in a luminance area), and is a basic unit of encoding. Motion prediction or the like is performed for each area, and the type of coding and the like are determined.

[0004] The divided area data is sent to the motion estimating means 1
03 is input. The motion prediction unit 103 performs motion prediction from another image, and when a similar area is found, encodes only the difference. And each means 1 described later
By the restoration processing of the image data in 09 to 112,
Motion prediction is performed from the image data using an image restored under the same conditions as those of the decoding device. When coding by motion prediction is performed, the output of the motion prediction unit 103 is difference data. When the encoding by the motion prediction is not performed, the pixel data is output from the motion prediction unit 103 as it is.

The output difference data and pixel data are sent to the frequency conversion means 104. Frequency conversion means 104
Performs frequency conversion by DCT. This conversion is usually performed in units of 8 × 8 pixels in a macroblock. It should be noted that in moving picture compression encoding methods other than MPEG, D
Not only the CT but also other appropriate frequency conversion methods may be used.

[0006] The calculated frequency component is supplied to a quantization means 10.
At 5, quantization is performed. The quantization means 105 uses the fact that human visual characteristics are insensitive to high frequencies,
The division of each frequency component is performed using a predetermined matrix in which the division value of the high frequency component is set large. In the quantization means 105,
By multiplying using the same matrix on the decoding device side,
High-frequency components can be restored with reduced accuracy. The matrix for performing the quantization is calculated by multiplying the quantization width obtained from the rate control unit 108 described later. If the value of the quantization width is set to a large value, it becomes possible to perform large rounding of high-frequency components, and it is possible to reduce the code amount. Therefore, the generated code amount (output code amount) is controlled by controlling the value of the quantization width.

[0007] The quantized output is supplied to a variable length coding means 1
At 06, variable length coding is performed. Variable length coding means 1
06 is output to the output buffer memory 107.
Sent to Here, the code amount of the bit stream output from the variable length coding unit 106 is not constant for each image because of the variable length coding. For this reason, in the output buffer memory 107, the difference in the code amount for each image is reduced, and the image is output as an encoded bit stream.

[0008] The rate control means 108 observes the amount of data generated at the time of variable length coding in the variable length coding means 106, and changes the quantization width in the quantization means 105 to generate the generated bit rate. Control.

Each of the means 109 to 112 internally restores data for performing motion compensation. Inverse quantization means 10
No. 9 performs multiplication under the inverse condition of quantization. Frequency inversion means 1
Numeral 10 performs inverse conversion of the frequency conversion in the frequency conversion unit 104. Since the frequency transform unit 104 performs the frequency transform by DCT, the frequency inverse transform unit 110 performs the inverse frequency transform by the inverse DCT. The motion compensator 111 performs motion compensation reverse to the encoding on the data obtained by the frequency inverse transform performed by the frequency inverse transformer 110 to restore the image data. Then, the restored data is stored in the image memory 112 and used for motion estimation of the next image.

The motion information and the information such as the macroblock encoding method are also variable-length-coded and input to the buffer memory 107, but the description is omitted here.

[0011]

SUMMARY OF THE INVENTION Variable length coding means 10
In No. 6, an encoding method adapted to data subjected to frequency conversion and quantization is adopted. In ordinary image data, the energy distribution is concentrated on the low frequency side.
In the quantized data in which the precision of the high frequency component is reduced, the energy distribution is further concentrated on the low frequency side, and the high frequency component has many “0”. Therefore, such data encoding is performed by scanning the data zigzag from the low frequency side to the high frequency side and using a correspondence table of (the number of "0") and (the data value that follows). However, the appearance frequency of these data is naturally high ("0"
), And (the subsequent data value) are small, and codes corresponding to such data are assigned codes having a short code length. Conversely, for those having an extremely large number of "0" and those having an extremely large (data value that follows), the table itself does not exist because it hardly appears in the data. Done. The exception processing is performed according to MPEG2 (ISO1381).
8) is called an escape encoding process.

FIG. 14 shows the internal configuration of the variable length coding means 106. The Run-Level processing means 301 converts the input quantized data into a combination of (the number of “0”) and (a data value that follows). Run-Level
Output data of the processing means 301 is output to the ROM 302 and each of the means 303 and 304. The ROM 302 stores a variable length encoding table. The escape code adding unit 303 adds an escape code (“000 in MPEG2”) to the output data of the run-level processing unit 301.
001 ”(6-bit fixed-length code), and converted into a code of (escape code) + (number of“ 0 ”) + (actual data). In MPEG2, the escape code is a 6-bit fixed-length code of “000001”,
(The number of "0") is a 6-bit fixed length code, and the actual data is a 12-bit fixed length code, so that a 24-bit code length is always required. The escape determination means 304 determines whether or not an escape encoding process is performed.

A multiplexer (MUX) 305 selects one of the output of the ROM 302 and the output of the escape code adding means 303 in accordance with an instruction from the escape judging means 304 to output encoded data. That is, the multiplexer 305 outputs the output of the escape code adding unit 305 as encoded data output when the escape encoding process is performed, and outputs the output of the ROM 302 as the encoded data output when the escape encoding process is not performed.

As described above, the variable length coding means 106
When the escape encoding process is performed, an encoded data output (= (escape code; which is twice as large as the input quantized data (actual data; 12-bit fixed-length code))
6 bits) + (the number of “0”; 6 bits) + (actual data) is output.

The variable length coding table stored in the ROM 302 is set assuming an ideal state. However, video signals of analog media and ordinary TV broadcasts are not ideal, and various noise components are mixed in. In many cases, the data is outside the long coding table. Escape encoding is performed on data outside the variable length encoding table.

Accordingly, the frequency of performing the escape encoding process in the variable length encoding means 106 increases,
The data amount of the encoded data output increases. As a result, the amount of data given to high-frequency components caused by noise increases, so that the amount of data that must be given to low-frequency components required for a restored image decreases, and the quality of a decoded image deteriorates. was there.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and an object of the present invention is to provide a moving picture coding method capable of effectively removing noise components and improving the quality of a decoded picture. Another object is to provide a moving picture encoding device.

[0018]

According to the first aspect of the present invention, an input moving image signal is converted into a frequency domain, and a frequency domain output is subjected to a quantization process using a predetermined quantization matrix, and a quantization process is performed. A variable-length encoding of the encoded output according to a predetermined table, and encoding data that does not match the table as exceptional processing, wherein, when performing the variable-length encoding, the frequency of occurrence of the exceptional processing The gist of the present invention is that the bandwidth of the input moving image signal is limited according to the occurrence frequency.

According to a second aspect of the present invention, an input moving image signal is converted into a frequency domain, a frequency domain output is subjected to a quantization process using a predetermined quantization matrix, and the quantized output is stored in a predetermined table. A variable-length encoding according to the following, and encodes data that does not match the table as exceptional processing, wherein when performing the variable-length encoding, the occurrence frequency of the exceptional processing is observed, The gist of the present invention is to limit the band of the frequency conversion output according to the frequency.

According to a third aspect of the present invention, an input moving image signal is converted into a frequency domain, a frequency domain output is subjected to a quantization process using a predetermined quantization matrix, and the quantized output is stored in a predetermined table. A variable-length encoding according to the following, and encodes data that does not match the table as exceptional processing, wherein when performing the variable-length encoding, the occurrence frequency of the exceptional processing is observed, The gist of the present invention is to limit the bandwidth of the output of the quantization process according to the frequency.

According to a fourth aspect of the present invention, an input moving image signal is converted into a frequency domain, a frequency domain output is subjected to a quantization process using a predetermined quantization matrix, and the quantized output is stored in a predetermined table. A variable-length encoding according to the following, and encodes data that does not match the table as exceptional processing, wherein when performing the variable-length encoding, the occurrence frequency of the exceptional processing is observed, The gist of the present invention is to change the value of a quantization matrix used in the quantization processing according to the frequency.

According to a fifth aspect of the present invention, in the moving picture coding method according to any one of the first to fourth aspects, when the exception processing is observed, only the one generated by a high-frequency component is observed. The main point is to observe.

According to a sixth aspect of the present invention, there is provided a transforming means for transforming an input video signal into a frequency domain, and a quantization for quantizing a frequency domain output of the frequency transforming means using a predetermined quantization matrix. Means, variable-length coding means for performing variable-length coding on the quantized output of the quantization means using a predetermined coding table, and coding data which does not match the table as exceptional processing; The gist of the present invention is to provide an observation unit for observing the frequency of occurrence of the exception processing, and a band limiting unit for limiting the band of the input moving image signal according to an output of the observation unit.

According to a seventh aspect of the present invention, there is provided a transforming means for transforming an input moving picture signal into a frequency domain, and a quantization for quantizing a frequency domain output of the frequency transforming means using a predetermined quantization matrix. Means, variable-length coding means for performing variable-length coding on the quantized output of the quantization means using a predetermined coding table, and coding data which does not match the table as exceptional processing; The gist of the present invention is that an observing unit that observes the frequency of occurrence of the exception processing and a band limiting unit that limits a band of an output of the frequency converting unit according to an output of the observing unit are provided.

According to an eighth aspect of the present invention, there is provided a conversion means for converting an input moving image signal into a frequency domain, and a quantization means for quantizing a frequency domain output of the frequency conversion means using a predetermined quantization matrix. Means, variable-length coding means for performing variable-length coding on the quantized output of the quantization means using a predetermined coding table, and coding data which does not match the table as exceptional processing; The gist of the present invention is that an observing unit that observes the frequency of occurrence of the exception processing and a band limiting unit that limits the band of the output of the quantizing unit according to the output of the observing unit.

According to a ninth aspect of the present invention, there is provided a transforming means for transforming an input moving image signal into a frequency domain, and a quantization means for quantizing a frequency domain output of the frequency transforming means using a predetermined quantization matrix. Means, variable-length coding means for performing variable-length coding on the quantized output of the quantization means using a predetermined coding table, and coding data which does not match the table as exceptional processing; The gist of the present invention is that observation means for observing the frequency of occurrence of the exception processing and control means for changing a value of a quantization matrix used at the time of operation by the quantization means in accordance with an output of the observation means are provided. .

According to a tenth aspect of the present invention, in the moving picture coding apparatus according to any one of the sixth to ninth aspects, the observing means observes only the exception processing caused by the high frequency component. That is the gist.

[0028]

BEST MODE FOR CARRYING OUT THE INVENTION

(First Embodiment) Hereinafter, a first embodiment of the present invention will be described with reference to the drawings. In this embodiment,
The same components as those in the conventional embodiment shown in FIGS. 13 and 14 are denoted by the same reference numerals, and detailed description thereof will be omitted.

FIG. 1 shows a block configuration of an encoding device 1 according to the present embodiment. The frequency conversion unit 104 performs frequency conversion by DCT on the input pixel data and difference data. The frequency-converted data is input to the band limiting unit 2. The band limiting unit 2 performs a band limiting process on the input data. The band limiting process includes the following.

(1) High-frequency components above a predetermined region in the input data are completely removed. (2) High-frequency components in a predetermined area or more in input data are limited. That is, the low-frequency component in the input data is passed as it is, and the high-frequency component is divided by a large value. In this way, it is possible to prevent a decoding device (not shown) from processing high-frequency components.

Here, as an index of removal or restriction of high frequency components, the number of occurrences of exception processing observed by the exception processing observation means 3 described later is used. In this case, the degree of removal and restriction of high frequency components is increased.

The output of the band limiting means 2 is supplied to the quantization means 105.
Is input to The quantization means 105 performs a quantization process.
The output of the quantization means 105 is subjected to variable length coding by the variable length coding means 106.

The encoded data output of the variable length encoding means 106 is sent to an output buffer memory 107. here,
The bit rate of the bit stream output from the variable length coding unit 106 is not constant for each image because of the variable length coding. For this reason, in the output buffer memory 107, the difference in the amount of data for each image is reduced, and the image is output as an encoded bit stream.

The rate control means 108 observes the amount of data generated at the time of variable-length coding in the variable-length coding means 106 and controls the generated bit rate by changing the quantization width in the quantization means 105. I do.

The exception processing observation means 108 observes the number of appearances of the escape coding process generated in the variable length coding means 106, and if the number of occurrences is extremely large, it is determined that many noise components are generated. Judgment is made, and high-frequency components are removed or limited by controlling the band limiting process of the band limiting means 2.

FIG. 2 shows an example of coding of ordinary quantized data. FIG. 2A shows the scan order of the quantized data. The data obtained by performing DCT on 8 × 8 pixels has a DC component at the upper left end, and the frequency increases as the number given to the pixel point increases in the direction of the arrow. In normal pixel data, the energy distribution in the frequency domain is concentrated on the low frequency side. Therefore, as shown by the arrows and the numbers given to the pixel points in FIG. 2A, the data is zigzag-scanned from the low frequency side to the high frequency side, and arranged in the data at that time (the number of “0”). ) And (subsequent data values) are encoded using a variable length encoding table.

FIG. 2B shows an example of actual quantized data. In FIG. 2C, the quantized data shown in FIG. 2B is converted into a set of (the number of “0”) and (a data value that follows), and is encoded using a variable-length encoding table. The results are shown.

In FIG. 2C, the last two data combinations A and B correspond to the escape encoding process. As described above, in the case of escape encoding processing, 24
It requires a bit code length. Therefore, both combinations A and B of each data require a code length of 24 bits,
The total code length is 48 bits. On the other hand, the sum of the code lengths of the 64 operation units (pixel points) of the quantized data is 99 bits. That is, the combination A, B of each data is 48.5% of the code length of all the quantized data.
Will be occupied. However, as described above, the data to be subjected to the escape encoding process is often a noise component. For this reason, in most cases, even if the data to be subjected to the escape encoding process is removed or restricted, the image quality of the decoded image is not affected.
Therefore, even if the lower right side, which is a high-frequency component, is removed or restricted from the quantized data shown in FIG. 2B, there is no significant effect on the image quality.
The same result can be obtained by exchanging the order of the band limitation in and the quantization processing in the quantization means 105 and performing band limitation on the quantized output. That is, in the encoding device 1, the output of the frequency conversion unit 104 is input to the quantization unit 105, the output of the quantization unit 105 is input to the band limiting unit 2, and the output of the band limiting unit 2 is The information may be input to the input unit 106.

By the way, there are the following two types of escape encoding processing in the variable length encoding means 106. As shown in combinations A and B of each data in FIG.
(The number of "0") is large. In this case, since many 0s continue until then, it can be easily determined that the noise component is caused.

(The data value that follows) is large.
In this case, the escape encoding process is performed even when (the number of “0”) is small. Therefore, since the low-frequency component is also subjected to the S-case encoding process, the image quality of the decoded image is significantly affected. That is, it is not possible to determine that the noise is caused by a noise component simply by observing only the number of appearances of the escape encoding process.

Therefore, as shown in FIG. 3, (number of "0") in the escape encoding process is observed, and the escape encoding process which occurs when (the number of "0") is larger than a predetermined value is performed. By observing the number of appearances, it is determined whether or not the escape encoding process is performed using a noise component.

Alternatively, as shown in FIG. 4, both (the number of “0”) and (the data value that follows) are observed,
(The number of “0”) is greater than a predetermined value, and
By observing the number of appearances of the escape encoding process that occurs when (the data value that follows) is smaller than a predetermined value, it is determined whether or not the escape encoding process is a noise component.

In the zigzag scan order shown in FIG. 2A, by observing the number of occurrences of the escape encoding process occurring in a predetermined order transition, it is determined whether or not the escape encoding process is a noise component. Is also good.

By the way, there are the following methods for counting (the number of "0"). [1] A method in which a means for counting (the number of “0”) is provided in the exceptional processing observation means 3.

[2] Since (the number of “0”) is a 6-bit fixed-length code, it is determined whether or not the most significant bit is “1”. It can be determined whether or not this is the case. And ("0"
Is greater than or equal to 32, it is determined that the escape encoding process is performed using a noise component.

FIG. 5 shows this state. 402 in the figure is FIG.
This is an example of the same quantized data. By multiplying this by a matrix corresponding to 502, the data of the 0 portion on this matrix disappears, and the high-frequency component is removed. In this case, an integer value of only 0 or 1 was used as a numerical value. However, this is converted into a real numerical value like a matrix shown by 601 to 603 in FIG. It is possible to do. As a result, generation of data that is considered to be caused by noise is prevented, and as a result, image quality can be improved. In this calculation, multiplication by a value of 1 or less has been described.
The same effect can be obtained even in the form of division by one or more numerical values.

Then, what kind of band limitation is to be performed is determined based on the number of occurrences of exception processing (frequency of occurrence of escape codes) observed by the exception processing observation processing 3 around a predetermined time. This is shown in FIG. Numerals 701 to 703 indicate matrices for multiplication in FIG.
To 703, the restriction of the high frequency component is strong. Using this matrix data, for example, in a matrix like 703,
High frequency components are almost completely removed.

FIG. 6 shows an example in which real values are used for the values of the band limiting matrix. 601 to 603 are respectively shown in FIG.
2 shows a matrix for multiplication in which the high frequency components 601 to 603 are more restricted. By changing in this way, the restriction on the high frequency side can be changed, and the matrix is changed in this way in accordance with the frequency of occurrence of exception processing in the variable length coding unit, thereby removing noise components. Further, in this case, an example of multiplication has been described, but of course, a configuration in which division is performed using the reciprocal of this value is also possible.

FIG. 8 shows the encoding apparatus 1 shown in FIG.
2 shows a block configuration when applied to the G encoding device 11. (Second Embodiment) Hereinafter, a second embodiment of the present invention will be described with reference to the drawings. In this embodiment,
The same components as those in the first embodiment have the same reference numerals, and a detailed description thereof will be omitted.

FIG. 9 shows a block configuration of the encoding device 21 of the present embodiment. In the first embodiment, the configuration in which the band limitation and the quantization are separated has been described. However, it is also possible to perform this in one process. As the final processing result, the same one as the encoding device described in FIG. 1 is obtained. The input pixel data is converted by the frequency conversion means 104 into data in the frequency domain. This is input to the quantization means 105. At this time, the quantization matrix used for quantization is changed depending on the frequency of exception processing that occurs during the subsequent variable-length encoding.

In the encoding apparatus shown in FIG. 1, after the pixel data is subjected to frequency transformation such as DCT by the band control means 2 as shown in FIG. . Then, the data after the processing is subjected to quantization by the quantization means in FIG. 1, and this quantization is division processing by a matrix. Therefore, the same processing can be performed by dividing the data in the frequency domain by the result of dividing the quantization matrix by the band limiting matrix. FIG.
Performs this matrix operation in the quantization matrix changing means 2
2.

When the quantization matrix is given for multiplication, that is, given as a real number of 1 or less, a processing matrix is obtained by performing multiplication with the band-limiting matrix for multiplication.
This may be multiplied by the data by the quantization means. Conversely, if the band-limiting matrix is given by a real number of 1 or less for division, by multiplying by the band-limiting matrix for division. That is, a processing matrix is obtained, and this may be divided by the quantization means with respect to the data.

In the case of encoding by MPEG, when the encoding is performed by changing the quantization matrix, each value of the matrix is given to the decoding means by inserting it into a code string. I have. However, in this case, since the purpose is to remove a high frequency band due to noise or the like by band limitation, it is necessary to insert a changed quantization matrix due to band limitation, which is an output of the quantization matrix conversion means 22, into a code string. Absent.
Therefore, in this case, the quantization matrix is encoded using the one before the band limitation.

This quantization matrix changing means has the same matrix changing rules as the band limiting means of FIG. 1 (FIGS. 6 and 7) with respect to the frequency of exception processing in the variable length coding means.
A quantization matrix that can perform this processing collectively is provided to the quantization means. Then, the quantization matrix transmitted to the decoding means uses the one which does not perform the band limiting process.

FIG. 10 shows an encoding apparatus 21 shown in FIG.
3 shows a block configuration when applied to the PEG encoding device 31. In the inverse quantization means 109 of the MPEG encoding device 31, a quantization matrix before band limitation is used.

(Third Embodiment) Hereinafter, a third embodiment of the present invention will be described with reference to the drawings. In this embodiment, the same components as those in the first embodiment have the same reference numerals, and a detailed description thereof will be omitted.

FIG. 11 shows a block configuration of the encoding device 41 of the present embodiment. In the first and second embodiments, it has been described that band limitation is performed on data after frequency conversion. However, the same effect can be expected by observing the frequency of exception processing and removing high-frequency components that become noise according to the frequency.

High frequency components of the input image data are limited by the band limiting means 2. This is performed according to the exception processing frequency of the exception processing observation unit 3. As this band limitation, an ordinary two-dimensional filter, FIR filter, etc.
Any filter can be used as long as it can perform high-frequency restriction.

After this output is input to the input processing means 43 for performing macroblock conversion and the like, this output is input to the motion prediction means 103 and subjected to processing such as motion compensation. , And are subjected to a process of quantization and variable-length encoding, and output as a code string. In this case, the exceptional processing observation performed by the exceptional processing observation means 3 is the same as that of the first embodiment. If possible, only the exceptional processing due to the high frequency component is observed, and the band limiting means 42 of the preceding stage performs the band limiting according to the frequency. Will be controlled.

In this example, it has been described that the band limitation is performed before the input processing means 43. However, it is only necessary that the band can be limited to the input moving image signal. Therefore, the band limiting unit 42
Alternatively, it may be at the subsequent stage of the motion prediction means 103.

The above embodiments may be modified as described below, and the same operation and effect can be obtained in such a case. (1) The frequency conversion means 104 uses an appropriate frequency conversion method other than DCT.

(2) The present invention is applied to a moving picture compression encoding method other than MPEG. As described above, each embodiment has been described.
The technical ideas other than the claims that can be grasped from each embodiment will be described below together with their effects.

(A) The moving picture encoding method according to any one of claims 1 to 5, wherein the frequency conversion processing is a discrete cosine transformation processing. In this way, the present invention can be applied to the MPEG encoding method.

(B) The moving picture coding apparatus according to any one of claims 6 to 10, wherein the frequency conversion processing is a discrete cosine transformation processing. In this way, the present invention can be applied to an MPEG encoding device.

(C) The input moving image signal is divided into predetermined regions, the divided output is subjected to motion prediction, and the output after the motion prediction is described in any one of claims 1 to 5. MPEG encoding method for performing signal processing according to the moving image encoding method.

In this way, the MPEG encoding method can be embodied. (D) an area dividing means (102) for dividing an input moving image signal into predetermined areas; a motion estimating means (103) for performing motion estimation on an output of the area dividing means; MP equipped with the video encoding device according to claim 1.
EG encoding device.

In this way, an MPEG encoding device can be embodied. (E) The moving picture coding method according to any one of claims 1 to 5, wherein the exception processing is an escape coding processing.

In this way, exception processing suitable for the MPEG encoding method can be executed. (F) The moving picture coding apparatus according to any one of claims 6 to 10, wherein the exception processing is an escape coding processing.

In this way, exception processing suitable for the MPEG encoding device can be executed.

[0070]

According to the present invention, there is provided a moving picture coding method capable of effectively removing noise components and improving the quality of a decoded picture. can do.

According to the present invention, there is provided a moving picture coding apparatus capable of effectively removing a noise component and improving the quality of a decoded picture. Can be.

[Brief description of the drawings]

FIG. 1 is a block diagram of a first embodiment.

FIG. 2 is a conceptual diagram for explaining the operation of the first embodiment.

FIG. 3 is a conceptual diagram for explaining the operation of the first embodiment.

FIG. 4 is a conceptual diagram for explaining the operation of the first embodiment.

FIG. 5 is a conceptual diagram for explaining the operation of the first embodiment.

FIG. 6 is a diagram showing a band limiting matrix used in the first embodiment.

FIG. 7 is a diagram showing a band limiting matrix used in the first embodiment.

FIG. 8 is a block diagram showing the first embodiment applied to an MPEG encoding apparatus.

FIG. 9 is a block diagram of a second embodiment.

FIG. 10 is a block diagram showing a configuration in which the second embodiment is applied to an MPEG encoding apparatus.

FIG. 11 is a block diagram of a third embodiment.

FIG. 12 is a block diagram showing a third embodiment applied to an MPEG encoding apparatus.

FIG. 13 is a block diagram of a conventional MPEG encoding apparatus.

FIG. 14 is a block diagram of a main part of a conventional MPEG encoding apparatus.

[Explanation of symbols]

 2 band limiting means 3 exception processing observing means as observing means 22 quantization matrix changing means 104 as control means 104 frequency transforming means 105 quantizing means 106 variable length coding means

Continuation of the front page (56) References JP-A-3-216075 (JP, A) JP-A-7-87490 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) H04N 7 / 24-7/68 H04N 1/41-1/419

Claims (10)

(57) [Claims] 1. An input video signal is converted into a frequency domain, a frequency domain output is subjected to a quantization process using a predetermined quantization matrix, and the quantized output is subjected to variable length coding according to a predetermined table.
A moving image encoding method for encoding data that does not match the table as exceptional processing, wherein when performing the variable-length encoding, the frequency of occurrence of the exceptional processing is observed, and according to the frequency of occurrence, A moving picture coding method characterized by performing band limitation on an input moving picture signal. 2. An input video signal is transformed into a frequency domain, a frequency domain output is subjected to a quantization process using a predetermined quantization matrix, and the quantized output is subjected to variable length coding according to a predetermined table.
A moving image encoding method for encoding data that does not match the table as exceptional processing, wherein when performing the variable-length encoding, the frequency of occurrence of the exceptional processing is observed, and according to the frequency of occurrence, A moving image encoding method, wherein a band is limited to an output of frequency conversion. 3. An input video signal is converted into a frequency domain, a frequency domain output is subjected to a quantization process using a predetermined quantization matrix, and the quantized output is subjected to variable-length coding according to a predetermined table.
A moving image encoding method for encoding data that does not match the table as exceptional processing, wherein when performing the variable-length encoding, the frequency of occurrence of the exceptional processing is observed, and according to the frequency of occurrence, A moving picture coding method, wherein a band is limited to an output of a quantization process. 4. An input video signal is converted into a frequency domain, a frequency domain output is subjected to a quantization process using a predetermined quantization matrix, and the quantized output is subjected to variable length coding according to a predetermined table.
A moving image encoding method for encoding data that does not match the table as exceptional processing, wherein when performing the variable-length encoding, the frequency of occurrence of the exceptional processing is observed, and according to the frequency of occurrence, A moving picture coding method characterized by changing a value of a quantization matrix used in a quantization process. 5. The moving picture coding method according to claim 1, wherein, when observing the exception processing, only a signal generated by a high-frequency component is observed. Method. 6. A transforming means for transforming an input moving image signal into a frequency domain, a quantizing means for quantizing a frequency domain output of the transforming means using a predetermined quantization matrix, and A variable-length encoding unit that encodes the quantized output in a predetermined encoding table using variable length encoding, and encodes data that does not match the table as exception processing; and observes the frequency of occurrence of the exception processing in the variable length encoding unit. A moving image encoding apparatus comprising: an observing unit that performs band-limiting on the input moving image signal according to an output of the observing unit. 7. A transforming means for transforming an input moving image signal into a frequency domain, a quantizing means for quantizing a frequency domain output of the transforming means using a predetermined quantization matrix, and A variable-length encoding unit that encodes the quantized output in a predetermined encoding table using variable length encoding, and encodes data that does not match the table as exception processing; and observes the frequency of occurrence of the exception processing in the variable length encoding unit. A moving image encoding apparatus, comprising: an observing unit that performs band limiting on an output of the conversion unit according to an output of the observing unit. 8. A transforming means for transforming an input moving image signal into a frequency domain, a quantizing means for quantizing a frequency domain output of the transforming means using a predetermined quantization matrix, and A variable-length encoding unit that encodes the quantized output in a predetermined encoding table using variable length encoding, and encodes data that does not match the table as exception processing; and observes the frequency of occurrence of the exception processing in the variable length encoding unit. A moving image encoding apparatus comprising: an observation unit that performs band limitation on an output of the quantization unit in accordance with an output of the observation unit. 9. A transforming means for transforming an input moving image signal into a frequency domain, a quantizing means for quantizing a frequency domain output of the transforming means using a predetermined quantization matrix, and A variable-length encoding unit that encodes the quantized output in a predetermined encoding table using variable length encoding, and encodes data that does not match the table as exception processing; and observes the frequency of occurrence of the exception processing in the variable length encoding unit. A moving image encoding apparatus comprising: an observing unit that performs the operation; and a control unit that changes a value of a quantization matrix used by the quantizing unit at the time of operation in accordance with an output of the observing unit. 10. The moving picture encoding apparatus according to claim 6, wherein said observation means observes only exception processing caused by a high-frequency component.