JP4539028B2 - Image processing apparatus, image processing method, recording medium, and program - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
  The present invention relates to an image processing apparatus and an image processing method.,recoding mediaAnd an image processing apparatus suitable for use in a case where it is possible to re-encode the corresponding data using information relating to encoding performed in the past on the corresponding data, and Image processing method,recoding mediaAnd the program.
[0002]
[Prior art]
For example, in a system that transmits a moving image signal to a remote place such as a video conference system and a videophone system, in order to efficiently use a transmission path, line correlation and interframe correlation of video signals are used. The image signal is compression encoded.
[0003]
When the image signal is compression-encoded, encoding is performed so that the generated bit stream has a predetermined bit rate. However, in actual operation, it may be necessary to convert the bit rate of the bit stream due to the circumstances of the transmission path.
[0004]
Further, for example, when the transmitted image signal is edited at a broadcasting station, editing is performed in units of seconds. Therefore, it is preferable that the image information of a frame is independent from the image information of other frames. Therefore, a long GOP with a large number of frames constituting a GOP (Group of Picture) that is a set of frames in which information is correlated so that image quality does not deteriorate even when transferred at a low bit rate (for example, 3 to 9 Mbps). And a Short GOP transferred at a high bit rate (18 to 50 Mbps) and having a small number of frames constituting the GOP need to be mutually converted.
[0005]
For example, a system capable of frame editing after encoding uncompressed data into MPEG Long GOP stream data will be described with reference to FIG.
[0006]
An SDI (Serial Digital Interface) -ASI (Asynchronous Serial Interface) conversion apparatus 1 encodes a supplied SDI input image so as to be an MPEG Long GOP (ASI stream), and encodes the encoded MPEG Long GOP. Output stream data. SDI is an uncompressed digital video / audio transmission system based on point-to-point transmission, and is defined by ANSI (American National Standards Institute) / SMPTE (Society of Motion Picture and Television Engineers) 259M. ing.
[0007]
The ASI-SDTI CP (Serial Data Transport Interface Contents Package) conversion device 2 once decodes the supplied MPEG Long GOP stream data by the decoding unit 21, and then all the intra frames (All Intra) at the encoding unit 22. ), And all the encoded intra-frame stream data (SDTI CP stream) is output to the frame editing apparatus 3 of the SDTI CP interface. SDTI CP is a global standard for a transmission method for transmitting MPEG data in real time (synchronized transfer), standardized as SMPTE 326M by the promotion of the Pro-MPEG Forum.
[0008]
The stream data frame-edited by the frame editing device 3 is supplied to the SDTI CP-ASI conversion device 4. The SDTI CP-ASI conversion apparatus 4 once decodes all the supplied intra-frame stream data by the decoding unit 31, and then encodes and encodes the MPEG Long GOP in the encoding unit 32. MPEG Long GOP stream data (ASI stream) is output.
[0009]
A system capable of encoding an input image into a MPEG long GOP at a high bit rate, decoding it, and re-encoding it into a low bit rate MPEG long GOP will be described with reference to FIG.
[0010]
The Long GOP encoding device 51 once decodes the supplied SDI input image by the decoding unit 61, and then encodes and encodes the high-bit-rate MPEG Long GOP in the encoding unit 62. MPEG Long GOP stream (ASI stream) data is output. The long GOP encoding device 52 once decodes the supplied high bit rate MPEG long GOP by the decoding unit 71, and then encodes the low bit rate MPEG long GOP by the encoding unit 72. The encoded low bit rate MPEG Long GOP stream (ASI stream) data is output.
[0011]
As described above, when encoding and decoding of image information are repeated, if the encoding parameter used for each encoding changes, the image information deteriorates. In order to prevent the deterioration of the image information, there is a technique that can suppress the deterioration of the image due to re-encoding by using the encoding history information inserted in the user data area of the picture layer of the bit stream ( For example, see Patent Document 1).
[0012]
[Patent Document 1]
JP 2000-059788 A
[0013]
For example, a case where encoding history information is used in a system capable of performing frame editing after encoding uncompressed data into MPEG Long GOP stream data will be described with reference to FIG. In addition, the same code | symbol is attached | subjected to the part corresponding to the case in FIG. 1, The description is abbreviate | omitted suitably.
[0014]
That is, the ASI-SDTI CP converter 101 is supplied with an MPEG Long GOP (ASI stream) generated by being encoded by the SDI-ASI converter 1 similar to FIG.
[0015]
Since the MPEG Long GOP is composed of three types of picture types (I picture, P picture, and B picture) having different encoding features, video data obtained by decoding the picture is also represented by a frame. There is a picture, P picture, and B picture. Therefore, when this video data is re-encoded with the MPEG Long GOP, if the video data having the feature of I picture, P picture or B picture is encoded with different picture types, Image degradation may occur. For example, if video data that is a B picture that tends to be more distorted than an I picture and a P picture is encoded as an I picture before decoding, the surrounding picture is predicted using the distorted I picture as a reference picture. Since it is encoded, the image quality deteriorates.
[0016]
In order not to cause such image quality degradation due to re-encoding, the ASI-SDTICP conversion apparatus 101 once decodes the supplied MPEG Long GOP stream data by the decoding unit 111, and then the encoding unit 112 When encoding to be an intra frame, all the parameters such as the picture type and the quantization value of the encoding executed in the past, that is, the encoding by the SDI-ASI conversion apparatus 1, are all SDTI-CP streams of the intra frame. It is added as history information (History data) of SMPTE 328M and supplied to the frame editing apparatus 3.
[0017]
The stream data subjected to frame editing by the frame editing device 3 is supplied to the SDTI CP-ASI conversion device 102. The SDTI CP-ASI conversion apparatus 102 decodes the supplied stream data of all intra frames with history information by the decoding unit 121. The encoding unit 122 re-encodes to a long GOP using necessary parameters such as a picture type and a quantization value included in the decoded history information, and outputs the result.
[0018]
Further, as described with reference to FIG. 2, in a system that can encode an input image into an MPEG LongGOP at a high bit rate, decode it, and re-encode it into an MPEG LongGOP at a low bit rate. A case where image degradation due to re-encoding is not caused will be described with reference to FIG. 2 corresponding to those in FIG. 2 are denoted by the same reference numerals, and description thereof will be omitted as appropriate.
[0019]
That is, the LongGOP encoding device 131 that has received the MPEG Long GOP stream (ASI stream) data encoded by the LongGOP encoding device 51 decodes the high bit rate MPEG Long GOP by the decoding unit 141. In this case, necessary encoding parameters are acquired, and the decoded video data and the acquired encoding parameters are supplied to the encoding unit 142. The encoding unit 142 encodes the video data using the supplied encoding parameters so as to be a low bit rate MPEG long GOP, and the encoded low bit rate MPEG long GOP stream ( ASI stream) data is output.
[0020]
[Problems to be solved by the invention]
As described above, using history information or encoding parameters, encoding by reusing past encoding information (picture type, motion vector, quantization value, etc. of past encoding) By doing so, it is possible to prevent image quality degradation. However, when the bit rate is lowered by re-encoding as in the system described with reference to FIG. 4, if the quantized value is used as it is, the bit rate cannot be lowered, and the quantized value is reused. Can not do it. In the system described with reference to FIG. 3, for example, depending on the VBV buffer state at the time of encoding a stream including an edit point, encoding may not be performed by reusing the quantized value.
[0021]
As described above, when the history information or the encoding parameter is used and the past encoding information is reused for encoding, the history value or the quantization value among the information included in the encoding parameter is used. When it cannot be reused, the quantized value is determined by rate control such as TM5 used in normal coding. That is, the quantized value used for encoding is different from the quantized value of past encoding.
[0022]
Furthermore, if the encoding device or transcoder that has been encoded in the past is different from the encoding device or transcoder that performs re-encoding, and the rate control method differs depending on the device, There is a high possibility that the quantized values will be different in the encoding performed.
[0023]
When the quantized value used in re-encoding is different from the quantized value in past encoding, image quality degradation is caused depending on the difference. For example, when there is a difference between the quantization values of about twice the quantization value in the past encoding, the image quality is significantly deteriorated, and 2 for the quantization value in the past encoding. If the difference between the quantization values is twice or more, the image quality is further significantly deteriorated.
[0024]
Regarding this cause, for example, the following can be considered.
[0025]
When the video data to be encoded contains a lot of high-frequency components such as noise and fine textures, DCT coefficients that do not become zero even when quantized are widely distributed to higher order. For example, consider a case where the result of quantizing a certain higher-order coefficient in an intra macroblock is 1. In re-encoding, if this part is quantized with a doubled quantization value, the result is 0.5, and becomes 1 again by rounding. Further, since the inverse quantization is performed at the time of decoding and the doubled quantization value is multiplied, the coefficient value becomes twice the coefficient value at the time of past encoding. Therefore, when inverse DCT is performed, high-frequency components such as noise and fine texture of video data are emphasized as compared with input video data at the time of past encoding.
[0026]
In addition, when the same part is quantized using a quantization value more than twice in re-encoding, the coefficient is rounded to 0, resulting in a decrease in resolution. In such a case, in an encoding apparatus that performs quantization calculation with truncation, the quantization value used in re-encoding is slightly larger than the past quantization value, and the quantization result is 0. Therefore, all the data that is 1 in the past becomes 0, the resolution drops, and the image deteriorates.
[0027]
In particular, when a portion that was an I picture in past encoding is re-encoded as an I picture again, such high-frequency component emphasis and image quality deterioration due to resolution reduction are likely to occur. Therefore, in the re-encoded video data, the high frequency component is emphasized only in the I picture portion or the resolution is lowered. Therefore, only the portion corresponding to the I picture is emphasized in the reproduced video data. As a result, flicker occurs.
[0028]
When the bit rate is reduced, the quantized value tends to increase, so that the above-described image quality degradation is likely to occur. Even if the bit rate remains the same, if the input video data at the time of re-encoding is different from the input video data at the time of past encoding, the bits of I picture, P picture, and B picture There will be a difference in the amount allocation. In such a case as well, there is a possibility that the quantized value may be different from that in the past encoding. Therefore, depending on the magnitude of the difference from the past quantized value, image quality degradation may occur. End up.
[0029]
The present invention has been made in view of such a situation, and in the case of encoding by reusing parameters in past encoding, image quality deterioration occurs when the quantized value cannot be reused. It is to be able to prevent this.
[0030]
[Means for Solving the Problems]
  The first image processing apparatus of the present invention includes a receiving unit that receives an encoding parameter used in past encoding of the image data, and a quantization value included in the encoding parameter received by the receiving unit. Used when encoding the image data without using itCompared to the quantization matrix defined in the coding standard, the value of the high frequency AC component was reduced.Storage means for storing a quantization matrix;Based on a determination result of determining whether to reuse a quantization value included in the encoding parameter received by the receiving means, and a determination result of the determination means,When the quantization value included in the encoding parameter received by the receiving unit is reused, the image data is encoded by reusing the quantization matrix included in the encoding parameter received by the receiving unit. AndBased on the determination result of the determination means,An encoding unit for encoding the image data using a quantization matrix stored in the storage unit when the quantization value included in the encoding parameter received by the receiving unit is not reused; Prepare.
[0031]
  In the case of encoding the image data without reusing the encoding parameter received by the receiving unit, the encoding unit uses the quantization matrix that is set as a default in the encoding standard. Data can be encoded.
[0032]
  The quantization matrix is an intraQuantizationIt can be a matrix or a non-intra quantization matrix.
[0035]
  The quantization matrix stored in the storage means can be configured such that the value of the high frequency AC component is limited to twice or less the DC component.
[0036]
  The quantization matrix stored in the storage means may be such that the value of the high frequency AC component and the value of the low frequency AC component are the same or substantially the same.
[0037]
  The encoding means may encode the image data in accordance with the MPEG standard.
[0039]
  Said determination meansIs included in the encoding parameter received by the receiving means when the bit rate for encoding the image data is smaller than the bit rate indicated in the encoding parameter received by the receiving means. It can be determined that the value is not reused.
[0040]
  Said determination meansDetermines that the quantization value included in the encoding parameter received by the receiving unit is not reused when image data encoded as an I picture in past encoding is encoded as an I picture. Can be.
[0041]
  In the first image processing method of the present invention, the receiving means receives the encoding parameter used in the past encoding for the image data, and the storage means is the quantized value included in the received encoding parameter. Used when encoding the image data without reusingCompared to the quantization matrix defined in the coding standard, the value of the high frequency AC component was reduced.Store the quantization matrix,A determination unit determines whether or not to reuse the quantization value included in the encoding parameter received by the reception unit;The encoding means isBased on the determination result of the determination means,When reusing the quantization value included in the received encoding parameter, the quantization data included in the received encoding parameter is reused, and the image data is encoded.Based on the determination result of the determination means,If the quantization value included in the received encoding parameter is not reused, the image data is encoded using the stored quantization matrix.
[0042]
  The first program of the present invention is received by the receiving means for receiving the encoding parameters used in the past encoding for the image data in the image processing apparatus for encoding the image data. This is used when the image data is encoded without reusing the quantization value included in the encoding parameter.Compared to the quantization matrix defined in the coding standard, the value of the high frequency AC component was reduced.Storage means for storing a quantization matrix;Based on a determination result of determining whether to reuse a quantization value included in the encoding parameter received by the receiving means, and a determination result of the determination means,When the quantization value included in the encoding parameter received by the receiving unit is reused, the image data is encoded by reusing the quantization matrix included in the encoding parameter received by the receiving unit. AndBased on the determination result of the determination means,An encoding unit for encoding the image data using a quantization matrix stored in the storage unit when the quantization value included in the encoding parameter received by the receiving unit is not reused; It functions as an image processing apparatus provided.
[0043]
  In the first image processing apparatus and method, and the program of the present invention, an encoding parameter used in past encoding for the image data is received, and a quantization value included in the received encoding parameter is calculated. Used when encoding the image data without reuseCompared to the quantization matrix defined in the coding standard, the value of the high frequency AC component was reduced.A quantization matrix is stored,It is determined whether to reuse the quantization value included in the received encoding parameter, and based on the determination result,When reusing the quantization value included in the received encoding parameter, the image data is encoded using the quantization matrix included in the received encoding parameter;Based on the judgment result,When the quantization value included in the received encoding parameter is not reused, the image data is encoded using the stored quantization matrix.
[0044]
  The second image processing apparatus of the present invention includes a receiving unit that receives an encoding parameter used in past encoding for the encoded stream, and a quantization value included in the encoding parameter received by the receiving unit. Used when re-encoding the encoded stream without re-useCompared to the quantization matrix defined in the coding standard, the value of the high frequency AC component was reduced.Storage means for storing a quantization matrix;Based on a determination result of determining whether to reuse a quantization value included in the encoding parameter received by the receiving means, and a determination result of the determination means,When the quantization value included in the encoding parameter received by the receiving unit is reused, the quantization stream included in the encoding parameter received by the receiving unit is reused to convert the encoded stream into the encoded stream. Re-encode andBased on the determination result of the determination means,Re-encoding for re-encoding the encoded stream using a quantization matrix stored in the storage means when the quantization value included in the encoding parameter received by the receiving means is not reused Means.
[0045]
  In the second image processing apparatus of the present invention, the receiving means receives the encoding parameter used in the past encoding for the encoded stream, and the storage means is the quantization included in the received encoding parameter. Used when re-encoding the encoded stream without reusing valuesCompared to the quantization matrix defined in the coding standard, the value of the high frequency AC component was reduced.Store the quantization matrix,Determining means determines whether to reuse the quantization value included in the encoding parameter received by the receiving means;The re-encoding means isBased on the determination result of the determination means,When re-using the quantization value included in the received encoding parameter, re-encoding the encoded stream using the quantization matrix included in the received encoding parameter;Based on the determination result of the determination means,If the quantization value included in the received encoding parameter is not reused, the method includes re-encoding the encoded stream using the stored quantization matrix.
[0046]
  According to a second program of the present invention, in the image processing apparatus for re-encoding the encoded stream, the computer receives reception parameters for receiving the encoding parameters used in the past encoding for the encoded stream, and the reception Used when re-encoding the encoded stream without re-using the quantized value contained in the encoding parameter received by the meansCompared to the quantization matrix defined in the coding standard, the value of the high frequency AC component was reduced.Storage means for storing a quantization matrix;Based on a determination result of determining whether to reuse a quantization value included in the encoding parameter received by the receiving means, and a determination result of the determination means,When the quantization value included in the encoding parameter received by the receiving unit is reused, the quantization stream included in the encoding parameter received by the receiving unit is reused to convert the encoded stream into the encoded stream. Re-encode andBased on the determination result of the determination means,Re-encoding for re-encoding the encoded stream using a quantization matrix stored in the storage means when the quantization value included in the encoding parameter received by the receiving means is not reused And function as an image processing apparatus.
[0047]
  In the second image processing apparatus and method, and the program of the present invention, an encoding parameter used in past encoding for the encoded stream is received, and a quantization value included in the received encoding parameter Used when re-encoding the encoded stream without reusingCompared to the quantization matrix defined in the coding standard, the value of the high frequency AC component was reduced.A quantization matrix is stored,It is determined whether to reuse the quantization value included in the received encoding parameter, and based on the determination result,When the quantization value included in the received encoding parameter is reused, the encoded stream is re-encoded using the quantization matrix included in the received encoding parameter. ,Based on the judgment result,If the quantization value included in the received encoding parameter is not reused, the encoded stream is re-encoded using the stored quantization matrix.
[0060]
Hereinafter, embodiments of the present invention will be described with reference to the drawings.
[0061]
FIG. 5 is a block diagram showing the configuration of a system capable of frame editing after encoding uncompressed data to MPEG Long GOP stream data to which the present invention is applied.
[0062]
In addition, the same code | symbol is attached | subjected to the part corresponding to the conventional case demonstrated using FIG. 3, and the description is abbreviate | omitted suitably. That is, in a system capable of frame editing after encoding uncompressed data into MPEG Long GOP stream data to which the present invention is applied, instead of the SDTI CP-ASI converter 102, the SDTI CP -It is basically configured in the same manner as the conventional case described with reference to FIG. 3 except that the ASI conversion device 151 is provided. The SDTI CP-ASI conversion device 151 is replaced with the encoding unit 122. The SDTI CP-ASI is provided with an encoding unit 161 that can select an optimal quantization matrix and execute a quantization process based on whether or not the quantization value can be reused. The configuration is basically the same as that of the conversion device 102.
[0063]
The ASI-SDTI CP converter 101 is supplied with an MPEG Long GOP (ASI stream) generated by being encoded by the SDI-ASI converter 1.
[0064]
The MPEG LongGOP is composed of three types of pictures (I picture, P picture, and B picture). When the ASI-SDTI CP conversion apparatus 101 decodes the supplied MPEG Long GOP stream data once by the decoding unit 111 and then encodes all the intra frames in the encoding unit 112, In the process, when these stream data are re-encoded with LongGOP, the video data having the feature of I picture, P picture, or B picture is encoded with another picture type. In order to prevent this, the parameters such as the picture type and the quantization value of the encoding performed in the past, that is, the encoding by the SDI-ASI converter 1, are all used as the history information (History data) of the SMPTE 328M. The data is added onto the CP stream and supplied to the frame editing device 3.
[0065]
The stream data with history information edited by the frame editing device 3 is supplied to the SDTI CP-ASI conversion device 151. The SDTI CP-ASI conversion device 151 decodes the supplied stream data of all intra frames with history information by the decoding unit 121. The encoding unit 161 re-encodes to LongGOP using parameters such as picture type and quantization value included in the decoded history information as necessary, and outputs the result.
[0066]
FIG. 6 is a block diagram illustrating a configuration of the encoding unit 161.
[0067]
The history extraction unit 171 extracts the history information from the SMPTE 328M SDIT CP stream with history information decoded by the decoding unit 121, supplies the history information to the rate setting unit 177, and supplies the video stream to the video rearrangement unit 172. Supply. The history information includes, for example, information related to encoding executed in the past, such as a picture type, a quantization value, a motion vector, or a quantization matrix.
[0068]
The video rearrangement unit 172 rearranges each frame image of the sequentially input image data as necessary, and corresponds each frame image of the image data to a luminance signal of 16 pixels × 16 lines and a luminance signal. Macro block data divided into macro blocks constituted by color difference signals to be generated is generated and supplied to the calculation unit 173 and the motion vector detection unit 174.
[0069]
The motion vector detection unit 174 receives the input of the macroblock data, calculates the motion vector of each macroblock based on the macroblock data and the reference image data stored in the frame memory 183, and serves as motion vector data. And sent to the motion compensation unit 182.
[0070]
The computing unit 173 performs motion compensation on the macroblock data supplied from the video rearranging unit 172 based on the image type of each macroblock. Specifically, the calculation unit 173 performs motion compensation in the intra mode for the I picture, performs motion compensation in the forward prediction mode for the P picture, and bidirectional for the B picture. Motion compensation is performed in the prediction mode.
[0071]
Here, the intra mode is a method in which a frame image to be encoded is used as transmission data as it is, and the forward prediction mode is a prediction residual between a frame image to be encoded and a past reference image as transmission data. The bidirectional prediction mode is a method in which transmission residuals are prediction residuals between a frame image to be encoded and past and future reference images.
[0072]
First, when the macroblock data is an I picture, the macroblock data is processed in the intra mode. That is, the calculation unit 173 sends the macroblock of the input macroblock data as it is to the DCT (Discrete Cosine Transform) unit 175 as calculation data. The DCT unit 175 converts the input operation data into a DCT coefficient by performing a DCT transform process that is orthogonal transform, and sends this to the quantization unit 176 as DCT coefficient data.
[0073]
The quantization unit 176 performs quantization processing on the input DCT coefficient data based on the quantization value Q and the quantization matrix supplied from the rate setting unit 177, and uses the VLC (Variable Length) as the quantization DCT coefficient. Code; variable length coding) unit 178 and inverse quantization unit 179. The quantization unit 176 calculates the quantized DCT coefficient Sq using the following equation (1).
[0074]
Sq = ((32 × S) / q) / (2 × Q) (1)
[0075]
However, in Equation (1), the DCT coefficient component value before quantization is S, the component value of the quantization matrix is q, and the multiplication result is rounded off.
[0076]
The quantized DCT coefficient data sent to the inverse quantization unit 179 is subjected to an inverse quantization process using the same quantization step size as that of the quantization unit 176, and is sent to the inverse DCT unit 180 as DCT coefficient data. The inverse quantization unit 179 calculates the inverse quantization DCT coefficient Siq using the following equation (2).
[0077]
Siq = (2 × Sq × q × Q) / 32 (2)
[0078]
The inverse DCT unit 180 performs inverse DCT processing on the supplied DCT coefficient data, and the generated computation data is sent to the computation unit 181 and stored in the frame memory 183 as reference image data.
[0079]
When the macroblock data is a P picture, the calculation unit 173 performs motion compensation processing using the forward prediction mode for the macroblock data. When the macroblock data is a B picture, the arithmetic unit 173 performs the bidirectional prediction mode for the macroblock data. Performs motion compensation processing by.
[0080]
The motion compensation unit 182 performs motion compensation on the reference image data stored in the frame memory 183 according to the motion vector data, and calculates forward prediction image data or bidirectional prediction image data. The calculation unit 173 performs subtraction processing on the macroblock data using the forward prediction image data or the bidirectional prediction image data supplied from the motion compensation unit 182.
[0081]
That is, in the forward prediction mode, the motion compensation unit 182 reads the reference image data by shifting the read address of the frame memory 183 according to the motion vector data, and uses the reference image data as the forward prediction image data. It supplies to the calculating part 181. The calculation unit 173 subtracts the forward prediction image data from the supplied macroblock data to obtain difference data as a prediction residual. Then, the calculation unit 173 sends the difference data to the DCT unit 175.
[0082]
The calculation unit 181 is supplied with the forward prediction image data from the motion compensation unit 182. The calculation unit 181 adds the forward prediction image data to the calculation data supplied from the inverse DCT unit, thereby making a reference. Image data is locally reproduced and output to the frame memory 183 for storage.
[0083]
In the bidirectional prediction mode, the motion compensation unit 182 reads the reference image data by shifting the read address of the frame memory 183 according to the motion vector data, and uses the reference image data as the bidirectional prediction image data. It supplies to the calculating part 181. The calculation unit 173 subtracts the bidirectional prediction image data from the supplied macroblock data to obtain difference data as a prediction residual.
Then, the calculation unit 173 sends the difference data to the DCT unit 175.
[0084]
Bidirectional prediction image data is supplied from the motion compensation unit 182 to the calculation unit 181, and the calculation unit 181 adds the bidirectional prediction image data to the calculation data supplied from the inverse DCT unit. Image data is locally reproduced and output to the frame memory 183 for storage.
[0085]
Thus, the image data input to the encoding unit 161 undergoes motion compensation prediction processing, DCT processing, and quantization processing, and is supplied to the VLC unit 178 as quantized DCT coefficient data. The VLC unit 178 performs variable length coding processing based on a predetermined conversion table for the quantized DCT coefficient data, and sends the variable length coded data obtained as a result to the buffer 184. The buffer 184 buffers the supplied variable length encoded data and then outputs it.
[0086]
The rate setting unit 177 constantly monitors the accumulation state of the variable-length encoded data stored in the buffer 184, and based on the occupation amount information indicating the accumulation state or the history information supplied from the history extraction unit 171. A quantization step size is determined.
[0087]
At this time, the rate setting unit 177 performs encoding, which will be described later, among the quantization matrices included in the history information supplied from the history extraction unit 171 or a plurality of quantization matrices stored in the memory 185. The one that matches the condition is selected.
[0088]
In the memory 185, in addition to the intra quantization matrix and the non-intra (inter) quantization matrix (FIG. 12) which are the defaults in MPEG2 shown in FIG. 7, the past encoding shown in FIG. The intra quantization matrix used when the quantized value of the information is not reusable is stored.
[0089]
As described above, when re-encoding using past encoding information, for example, the bit rate is lowered by re-encoding, or depending on the state of the VBV buffer at the time of encoding the stream including the edit point In some cases, the quantization value in the past encoding cannot be reused. In order to prevent image quality degradation caused by the fact that quantization values in past encoding cannot be reused, the intra quantization matrix that can protect the frequency component by reducing the value of the AC component is re-encoded. It may be used.
[0090]
FIG. 7 shows an intra quantization matrix that is the default in MPEG2. The intra quantization matrix in FIG. 7 is set so that the higher the AC component, the larger the value. This is because the human vision is insensitive to the degradation of the high-frequency component image, and by dividing the high-frequency component by a larger value, the data can be displayed without degrading the appearance of the image. This is because it can be greatly reduced. The DCT coefficient before quantization tends to have a large value for the DC component and a small value for the AC component. Therefore, by using a quantization matrix that takes a larger value as the frequency becomes higher, as shown in FIG. In many cases, the quantized coefficients are such that the high frequency components are almost zero. In MPEG2, the maximum value that each component of the quantization matrix can take is 255.
[0091]
However, when the video data contains a lot of high frequency components such as noise and fine texture, there are many cases where the coefficient remains in the high frequency component of the DCT coefficient before quantization and the value is 1. When such video data is decoded and re-encoded, it is possible to use the same quantization matrix as the past encoding, and if the quantization value is the same as the past encoding, the quantum The converted coefficient also becomes 1, and can be decoded again to the original value. However, as described above, if the quantized value is, for example, about twice that of past encoding, the value after decoding after re-encoding is greatly different from the value before past encoding. May occur.
[0092]
In this way, the image quality is significantly deteriorated due to the occurrence of value fluctuations in many components. Furthermore, if re-encoding is repeated for such video data, the same change in the decoded value is also repeated, so that the image quality further deteriorates every time re-encoding is performed. End up.
[0093]
Therefore, a high value is not set even in the high frequency AC component, and the high frequency component is recoded by using a quantization matrix that is configured with the same or substantially the same value as the low frequency component. Since the value immediately before dividing by the quantized value can be set to a somewhat large value at the time of conversion, it is possible to prevent the influence of the fluctuation of the decoded value due to the fluctuation of the quantized value described above. For example, as shown in FIG. 8, use a quantization matrix in which all components other than the DC component defined as “8” in the MPEG2 standard are set to 16, which is twice the value of the DC component. Thus, it is possible to avoid the influence of the fluctuation of the decoded value due to the fluctuation of the quantization value described above.
[0094]
When the default intra quantization matrix in MPEG2 shown in FIG. 7 is used, the DCT coefficient component value before quantization at the time of past encoding is 9, and the quantization matrix value of the corresponding component Is q = 34 and the quantization value is Q = 4, the quantization result is Sq = ((32 × 9) / 34) / (2 × 4) = 1 from the above-described equation (1). Then, when inverse quantization is performed, Sq = (2 × 1 × 34 × 4) / 32 = 9 is obtained from the above-described equation (2).
[0095]
At this time, when the quantization value Q becomes Q = 8 from Q = 4 at the time of re-encoding, the quantization result is similarly obtained from Equation (1) from Sq = ((32 × 9 ) / 34) / (2 × 8) = 1, and inverse quantization results in Sq = (2 × 1 × 34 × 8) / 32 = 17 from equation (2), and the DCT coefficient component value before quantization This value is significantly different from 9, which is.
[0096]
On the other hand, for example, when the quantization matrix of FIG. 8 is used for re-encoding, since the quantization matrix value of the corresponding component is q = 16, the quantization value Q is changed from Q = 4 to 2 Even if the double Q = 8, the quantization result at the time of re-encoding is Sq = ((32 × 9) / 16) / (2 × 8) = 1 from the equation (1), and the inverse quantum Then, from Equation (2), Siq = (2 × 1 × 16 × 8) / 32 = 8. That is, when the quantization matrix in FIG. 8 is used for re-encoding, an inverse quantization result that is not significantly different from the DCT coefficient component value 9 before quantization can be obtained. It can be seen that it is less susceptible to fluctuations in the decoded value due to fluctuations.
[0097]
Therefore, when the quantized value cannot be reused, a large value is not set even in the high frequency AC component as shown in FIG. 8 (the value of the high frequency AC component is twice that of the DC component). By using an intra quantization matrix that is limited to the following and is the same as or substantially equal to the value of the low-frequency AC component, image quality degradation after re-encoding can be suppressed.
[0098]
Here, the intra quantization matrix of FIG. 8 is not set to a large value even in the high frequency AC component, and the high frequency component is the same as or substantially equivalent to the low frequency component. Although described as an example of the quantization matrix, the intra quantization matrix used in the present invention is not limited to the intra quantization matrix of FIG. 8, and the AC component is not limited to the intra quantization matrix of FIG. Even if they are not all the same value, for example, a sufficiently small value may be used so that the high frequency component is not more than twice the DC component.
[0099]
The rate setting unit 177 performs encoding using history information based on a signal indicating a user's operation input supplied from an operation input unit (not shown) or whether history information is supplied from the history extraction unit 171. Determine whether it is possible to do this. As described above, the history information supplied from the history extraction unit 171 includes, for example, a picture type, a quantization value, a motion vector, or a quantization matrix.
[0100]
When the rate setting unit 177 determines that the encoding cannot be performed using the history information, the rate setting unit 177 reads the intra quantization matrix described with reference to FIG. 7 used for normal encoding from the memory 185, and This is supplied to the quantization unit 176.
A method for determining the quantized value Q when it is determined that encoding cannot be performed using history information will be described later.
[0101]
If the rate setting unit 177 determines that encoding can be performed using history information, can the quantization value be reused based on the history information supplied from the history extraction unit 171? Judge whether or not.
[0102]
Here, whether or not the quantized value can be reused can be set in advance by the user as to whether or not the quantized value can be reused. The information indicating whether or not the quantized value can be reused may be included in the image frame, the image frame in which the previous encoding indicated in the history information is performed, Whether the image frame at the time of encoding matches the position and size, whether the bit rate in the previous encoding shown in the history information is smaller than the bit rate of the current encoding, Alternatively, it may be determined based on a predetermined condition such as whether or not the chroma format in the previous encoding is larger than the current chroma format.
[0103]
If it is determined that the quantization value can be reused, the rate setting unit 177 supplies the quantization matrix included in the history information to the quantization unit 176. On the other hand, if it is determined that the quantized value cannot be reused, the rate setting unit 177 reads the quantization matrix described with reference to FIG. 8 from the memory 185 in order to prevent image degradation. , And supplied to the quantization unit 176.
[0104]
Further, when the history information is not used, or when the quantization value included in the history information cannot be reused, the rate setting unit 177 generates the generated code amount of the macroblock actually generated rather than the target generated code amount. When the number of generated codes is large, the quantization step size is increased to reduce the amount of generated codes, and when the actual generated code amount is smaller than the target generated code amount, the quantization step size is decreased to increase the generated code amount. Has been made.
[0105]
That is, the rate setting unit 177 obtains the buffer occupancy of the virtual buffer by assuming the transition of the accumulation state of the variable-length encoded data stored in the VBV buffer provided on the decoder side, and calculates the quantized value Q Is calculated and supplied to the quantization unit 176 together with the quantization matrix described with reference to FIG. 7, which is used for normal encoding.
[0106]
The buffer occupancy d (j) of the virtual buffer in the jth macroblock is expressed by the following equation (3), and the buffer occupancy d (j + 1) of the virtual buffer in the j + 1th macroblock is The buffer occupation amount d (j + 1) of the virtual buffer in the j + 1-th macroblock is expressed as the following equation (5) by subtracting the equation (4) from the equation (3). The
[0107]
d (j) = d (0) + B (j−1) − {T × (j−1) / MBcnt}
... (3)
[0108]
Here, d (0) is the initial buffer capacity, B (j) is the number of encoding generation bits in the j-th macroblock, MBcnt is the number of macroblocks in the picture, and T is the target generation in units of pictures. Code amount.
[0109]
d (j + 1) = d (0) + B (j) − (T × j) / MBcnt
... (4)
[0110]
d (j + 1) = d (j) + {B (j) -B (j-1)}-T / MBcnt
... (5)
[0111]
Note that when the macroblock in the picture is divided into an intra slice portion and an inter slice portion, for example, the rate setting unit 177 assigns the target to each macro block in the intra slice portion and the macro slice in the inter slice portion. The generated code amounts Tpi and Tpp can be set individually.
[0112]
Therefore, the rate setting unit 177 substitutes the buffer occupation amount d (j + 1) and the constant r shown in the equation (6) into the equation (7), thereby quantizing the index data Q of the macroblock (j + 1). (J + 1) is calculated and supplied to the quantization unit 176.
[0113]
r = (2 × br) / pr (6)
Q (j + 1) = d (j + 1) × (31 / r) (7)
Here, br is a bit rate and pr is a picture rate.
[0114]
The quantization unit 176 determines the quantization step size in the next macroblock based on the quantization value Q supplied from the rate setting unit 177, and quantizes the DCT coefficient data according to the quantization step size.
[0115]
As a result, the quantization unit 176 can quantize the DCT coefficient data with the optimum quantization step size for the target generated code amount of the next picture calculated based on the actual generated code amount.
[0116]
Thus, the quantization unit 176 can quantize the buffer 184 so that the buffer 184 does not overflow or underflow according to the data occupancy of the buffer 184, and quantize the decoder side VBV buffer so that it does not overflow or underflow. Quantized DCT coefficient data can be generated.
[0117]
In the above description, the case where the encoding process is performed in units of pictures has been described. However, the case where the encoding process is performed in units of slices or macroblocks instead of in units of pictures is basically. Similarly, the encoding process is executed.
[0118]
Also, when the input image described with reference to FIG. 4 is encoded into an MPEG LongGOP at a high bit rate, and is decoded and re-encoded into a low bit rate LongGOP, an image for re-encoding The present invention can be similarly applied to a system in which deterioration is not caused. FIG. 9 is a diagram illustrating an example in which a VBV buffer is prevented from failing when an input image is encoded into a MPEG long GOP at a high bit rate and is decoded and re-encoded into a low bit rate long GOP. FIG. 2 is a block diagram showing the configuration of a system configured to prevent image degradation due to re-encoding. Note that portions corresponding to those in FIG. 4 are denoted by the same reference numerals, and description thereof will be omitted as appropriate.
[0119]
That is, the system of FIG. 9 includes a LongGOP encoding device 201 instead of the LongGOP encoding device 131, and the LongGOP encoding device 201 uses the optimal quantization matrix instead of the encoding unit 142 to quantize. The configuration is basically the same as that of the LongGOP encoding device 131 except that the encoding unit 211 capable of encoding is provided.
[0120]
When the Long GOP encoding apparatus 201 receives the MPEG Long GOP stream (ASI stream) data encoded by the Long GOP encoding apparatus 51, the decoding unit 141 decodes the MPEG Long GOP having a high bit rate. In addition, necessary encoding parameters are acquired, and the decoded video data and the acquired encoding parameters are supplied to the encoding unit 211. The encoding unit 211 encodes the video data so as to become a low bit rate MPEG long GOP using the supplied encoding parameters as necessary, and encodes the encoded low bit rate MPEG. Outputs long GOP stream (ASI stream) data.
[0121]
FIG. 10 is a block diagram illustrating a configuration of the encoding unit 211. In FIG. 10, the same reference numerals are given to the portions corresponding to the encoding unit 161 in FIG. 6, and description thereof will be omitted as appropriate.
[0122]
That is, the encoding unit 211 is the same as that shown in FIG. 6 except that the history extraction unit 171 is omitted, and a parameter input unit 221 that acquires the parameters supplied from the decoding unit 141 and supplies them to the rate setting unit 177 is provided. The encoding unit 161 basically has the same configuration.
[0123]
The rate setting unit 177 constantly monitors the accumulation state of the variable-length encoded data stored in the buffer 184, and based on the occupancy information indicating the accumulation state or the parameter information supplied from the parameter input unit 221. A quantization step size is determined.
[0124]
The rate setting unit 177 performs encoding using the parameter information based on a signal indicating a user operation input supplied from an operation input unit (not shown) or whether parameter information is supplied from the parameter input unit 221. Determine whether to do it. The parameter information supplied from the parameter input unit 221 includes, for example, a picture type, a quantization value, a motion vector, a quantization matrix, and the like, similar to the history information described above.
[0125]
If the rate setting unit 177 determines that the encoding cannot be performed using the parameter information, the rate setting unit 177 reads the quantization matrix described with reference to FIG. To the conversion unit 176.
[0126]
If the rate setting unit 177 determines that encoding can be performed using the parameter information, can the quantization value be reused based on the parameter information supplied from the parameter input unit 221? Judge whether or not.
[0127]
Here, whether or not the quantized value can be reused can be set in advance by the user as to whether or not the quantized value can be reused. The information indicating whether or not the quantized value can be reused may be described in the image frame that has been subjected to the previous encoding indicated in the parameter information, Whether the image frame at the time of encoding matches the position and size, whether the bit rate in the previous encoding indicated in the parameter information is smaller than the bit rate of the current encoding, Alternatively, it is determined according to a predetermined condition such as whether or not the chroma format in the previous encoding is larger than the current chroma format.
[0128]
Then, when it is determined that the quantization value can be reused based on the parameter information, the rate setting unit 177 supplies the intra quantization matrix included in the parameter information to the quantization unit 176. On the other hand, if it is determined that the quantized value cannot be reused, the rate setting unit 177 reads the intra quantization matrix described with reference to FIG. 8 from the memory 185 in order to prevent image degradation. To the quantization unit 176.
[0129]
The quantization unit 176 quantizes the DCT coefficient supplied from the DCT unit 175 based on the quantization value and the quantization matrix supplied from the rate setting unit 177.
[0130]
Next, the quantization matrix determination process 1 executed in the encoding unit 161 in FIG. 6 or the encoding unit 211 in FIG. 10 will be described with reference to the flowchart in FIG.
[0131]
In step S1, the rate setting unit 177 is based on a signal indicating a user operation input supplied from an operation input unit (not shown), history information from the history information extraction unit 171 or parameter input from the parameter input unit 221. Whether or not the encoding is executed is determined by reusing the past encoding information described in the history information or the parameter.
[0132]
If it is determined in step S1 that the past encoding information is not reused and the encoding is not performed, the rate setting unit 177 is used in normal encoding in step S2 and will be described with reference to FIG. The intra quantization matrix that has been read is read from the memory 185 and supplied to the quantization unit 176, and the processing is terminated.
[0133]
If it is determined in step S1 that the past encoding information is to be reused, the rate setting unit 177 is set in advance in step S3 such that the quantized value cannot be reused. Whether the quantization value can be reused in the history information or parameter information, and the previous encoding indicated in the history information or parameter information is executed. Whether or not the position of the image frame and the image frame at the time of re-encoding coincide with the position and size, the bit rate in the previous encoding indicated in the history information and parameter information is the current encoding Whether the bit rate is smaller than the current bit rate, or whether the chroma format in the previous encoding is larger than the current chroma format. Based on whether or not the quantization value determines whether reusable.
[0134]
If it is determined in step S3 that the quantized value is reusable, the rate setting unit 177 is supplied from the history information supplied from the history information extraction unit 171 or the parameter input unit 221 in step S4. The intra-quantization matrix in the past encoding included in the parameters is supplied to the quantization unit 176, and the process ends.
[0135]
If it is determined in step S3 that the quantization value is not reusable, in step S5, the rate setting unit 177 stores the intra quantization matrix stored in the memory 185 and described with reference to FIG. The data is supplied to the quantization unit 176, and the process is terminated.
[0136]
By such processing, when the quantized value used for encoding is different from the quantized value of the past encoding, when the bit rate is reduced or the input video data at the time of re-encoding is encoded in the past Even when the input video data is different from the current input video data, it is possible to prevent flicker caused by high frequency components being emphasized or image resolution being lowered, particularly in an I picture.
[0137]
Further, in the processing described above, a large value is not used for the AC component (particularly, the high frequency part of the AC component) when the quantized value is not reusable among the information regarding the past encoding. The intra quantization matrix has been selected and described as being used for quantization, but not only the intra matrix but also the non-intra (inter) quantization matrix can be reused by the same processing. The optimum one may be selected based on whether or not there is.
[0138]
That is, in addition to the non-intra quantization matrix that is the default in the MPEG2 TM5 shown in FIG. 12, the coding unit 161 in FIG. 6 or the memory 185 in the coding unit 211 in FIG. In addition, a non-intra quantization matrix that does not use a large value even in a high-frequency AC component is further stored, and based on whether the quantization value can be reused by the rate control unit 177, The optimal one can be selected.
[0139]
FIG. 12 shows a non-intra quantization matrix that is the default in TM2 of MPEG2. The non-intra quantization matrix of FIG. 12 also uses a characteristic that human vision is insensitive to deterioration of the high-frequency component image, so that the high-frequency component is divided by a larger value. The AC component of the region is set to have a larger value.
[0140]
Even in the non-intra quantization matrix, a large value is not used even in the high frequency AC component, and the high frequency component is the same as the low frequency component, or by using a quantization matrix configured with substantially the same value, At the time of re-encoding, the value immediately before dividing by the quantized value becomes large, so that it is possible to avoid the influence of the fluctuation of the decoded value due to the fluctuation of the quantized value. For example, as shown in FIG. 13, by using a quantization matrix in which all the components are 16, it is possible to avoid the influence of the fluctuation of the decoded value due to the fluctuation of the quantization value described above.
[0141]
Here, FIG. 13 shows an example of a non-intra quantization matrix in which a high value is not used even in the high frequency AC component and the high frequency component is configured with the same or substantially the same value as the low frequency component. Although the description has been given using the quantization matrix in which all components are the same, the non-intra quantization matrix used in the present invention is not limited to the non-intra quantization matrix of FIG. Even if not all components or all AC components have the same value as in the intra quantization matrix, the high-frequency component has a sufficiently small value that is, for example, less than twice the DC component, It is only necessary to use a value whose high frequency component is smaller (for example, 32 or less) than the non-intra quantization matrix described with reference to 12.
[0142]
Next, referring to the flowchart in FIG. 14, based on whether or not the quantized value is reusable, which is executed in the encoding unit 161 in FIG. 6 or the encoding unit 211 in FIG. 10. A description will be given of the quantization matrix determination process 2 for selecting an intra quantization matrix and a non-intra quantization matrix.
[0143]
In step S21, the rate setting unit 177 is based on a signal indicating a user's operation input supplied from an operation input unit (not shown), history information from the history information extraction unit 171, or parameter input from the parameter input unit 221. Whether or not the encoding is executed is determined by reusing the past encoding information described in the history information or the parameter.
[0144]
If it is determined in step S21 that the past encoding information is not reused, the rate setting unit 177 is used in normal encoding in step S22 and will be described with reference to FIG. The intra quantization matrix and the non-intra quantization matrix described with reference to FIG. 12 are read from the memory 185 and supplied to the quantization unit 176, and the process is terminated.
[0145]
If it is determined in step S21 that the past encoding information is to be reused, the rate setting unit 177 is set in advance in step S23, for example, as a quantized value cannot be reused. Whether the quantization value can be reused in the history information or parameter information, and the previous encoding indicated in the history information or parameter information is executed. Whether or not the position of the image frame and the image frame at the time of re-encoding coincide with the position and size, the bit rate in the previous encoding indicated in the history information and parameter information is the current encoding A predetermined condition such as whether or not the bit rate is smaller than the current chroma format or whether or not the chroma format in the previous encoding is larger than the current chroma format Based on whether or not satisfied, the quantization value determines whether reusable.
[0146]
If it is determined in step S23 that the quantized value is reusable, the rate setting unit 177 is supplied from the history information supplied from the history information extraction unit 171 or the parameter input unit 221 in step S24. The intra quantization matrix and the non-intra quantization matrix in the past encoding included in the parameters are supplied to the quantization unit 176, and the process is terminated.
[0147]
If it is determined in step S23 that the quantization value is not reusable, in step S25, the rate setting unit 177 stores the intra quantization matrix described in FIG. The non-intra quantization matrix described with reference to FIG. 13 is supplied to the quantization unit 176, and the process ends.
[0148]
By such processing, when the quantized value used for encoding is different from the quantized value of the past encoding, when the bit rate is reduced or the input video data at the time of re-encoding is encoded in the past Even when it is different from the input video data at the time, in particular, it is possible to prevent flicker caused by a decrease in the resolution of the image of the I picture and to prevent deterioration of the image even in the non-intra part. can do.
[0149]
In the above description, it is assumed that a quantization matrix (standard default) used as a standard in the encoding device is used as a quantization matrix used when information on past encoding is not reused. Whatever value the quantization matrix (for example, quantization matrix A) used when information about past coding is not reused (for example, the highest possible in the MPEG standard for high-frequency components) In the present invention, the information relating to past coding is reused, but the quantization matrix used in the case where the quantization value of past coding is not reused (even if it has a value of 255). For example, the quantization matrix B) is a quantization matrix having a smaller high frequency component than the quantization matrix A (at least Can achieve the effect that if a quantization matrix) composed of a value less than 255, to prevent image deterioration.
[0150]
That is, in the case where the quantization value used for encoding is different from the quantization value of the previous encoding, the quantization matrix B is used for quantization, which is compared with the case where the quantization matrix A is used. Flicker that occurs especially when the resolution of the I picture image drops, even when the rate is low or when the input video data at the time of re-encoding is different from the input video data at the time of past encoding. It is possible to prevent image deterioration.
[0151]
In the present invention, for example, only a P picture is used without using a B picture that causes a reordering delay and an I picture with a large amount of generated code, and this P picture is converted into an intra-frame consisting of several slices. The present invention can also be applied to a case where low delay encoding is performed so that encoding can be performed without reordering by dividing the slice into inter slices including all remaining slices.
[0152]
In the present invention, all frame images are all P pictures as low delay coding. For example, within the frame size of 45 macroblocks in the horizontal direction and 24 macroblocks in the vertical direction, 2 macroblocks in the vertical direction and 45 macros in the horizontal direction from the top of the frame image Even when the block area is set as one intra slice part and all others are set as inter slice parts, the intra slice part is set as one vertical macro block and 45 horizontal macro block areas. The present invention can also be applied to the case of forming a region having a size of.
[0153]
Further, in the above-described embodiment, the case where the present invention is applied to the encoding unit 161 or the encoding unit 211 that performs compression encoding by the MPEG method has been described, but the present invention is not limited to this. The present invention may also be applied to encoding devices using other various image compression methods.
[0154]
In the above-described embodiment, each conversion device or LongGOP encoding device that converts stream data has been described as having a decoding unit and an encoding unit. The present invention can be applied even when the units are configured as independent devices as a decoding device and an encoding device, respectively.
[0155]
That is, in the above-described embodiment, each conversion device and LongGOP encoding device has been described as converting stream data. However, for example, as shown in FIG. The decoding device 251 for converting into a signal and the encoding device 252 for encoding a baseband signal and converting it into stream data may be configured as independent devices. Further, even when the decoding device 251 does not completely decode the supplied stream data and the corresponding encoding device 252 partially encodes the corresponding part of the incompletely decoded data, this The invention is applicable.
[0156]
For example, when the decoding device 251 performs only decoding and inverse quantization on the VLC code and has not performed inverse DCT transformation, the coding device 252 performs quantization and variable length coding processing, but does not perform DCT transformation. No processing is performed. The present invention can be applied in determining whether or not to reuse the quantization value in the quantization of the encoding device 252 that performs such partial encoding (encoding from an intermediate stage). Needless to say.
[0157]
Furthermore, when the baseband signal completely decoded by the decoding device 251 is encoded to an intermediate stage (for example, DCT transformation and quantization are performed but variable length encoding processing is not performed), Since the decoding device 251 has not completely decoded (for example, only decoding and inverse quantization are performed on the VLC code, and inverse DCT conversion is not performed), the data encoded to the middle stage is encoded. The present invention can also be applied to a case where the encoding device 252 further encodes to an intermediate stage (for example, quantization is performed but variable length encoding processing is not performed).
[0158]
Furthermore, the present invention can also be applied to a transcoder 261 configured by an encoding device 251 that performs such partial decoding and an encoding device 252 that performs partial encoding. Such a transcoder 261 is used, for example, when an editing device 262 that performs editing such as splicing is used.
[0159]
The series of processes described above can be executed by hardware, but can also be executed by software. In this case, for example, the SDTI CP-ASI converter 151 and the Long GOP encoder 201 are configured by a personal computer 301 as shown in FIG.
[0160]
In FIG. 16, a CPU (Central Processing Unit) 311 performs various processes according to a program stored in a ROM (Read Only Memory) 312 or a program loaded from a storage unit 318 to a RAM (Random Access Memory) 313. Execute. The RAM 313 also appropriately stores data necessary for the CPU 311 to execute various processes.
[0161]
The CPU 311, ROM 312, and RAM 313 are connected to each other via a bus 314. An input / output interface 315 is also connected to the bus 314.
[0162]
The input / output interface 315 includes an input unit 316 including a keyboard and a mouse, an output unit 317 including a display and a speaker, a storage unit 318 including a hard disk, and a communication unit 319 including a modem and a terminal adapter. It is connected. The communication unit 319 performs communication processing via a network including the Internet.
[0163]
A drive 320 is connected to the input / output interface 315 as necessary, and a magnetic disk 331, an optical disk 332, a magneto-optical disk 333, a semiconductor memory 334, or the like is appropriately mounted, and a computer program read from them is loaded. Installed in the storage unit 318 as necessary.
[0164]
When a series of processing is executed by software, a program constituting the software executes various functions by installing a computer incorporated in dedicated hardware or various programs. For example, a general-purpose personal computer is installed from a network or a recording medium.
[0165]
As shown in FIG. 16, the recording medium is distributed to supply a program to the user separately from the main body of the apparatus, and includes a magnetic disk 331 (including a floppy disk) on which the program is stored, an optical disk 332 ( Package media including CD-ROM (compact disk-read only memory), DVD (including digital versatile disk), magneto-optical disk 333 (including MD (mini-disk) (trademark)), or semiconductor memory 334 In addition to being configured, it is configured by a ROM 312 in which a program is stored and a hard disk included in the storage unit 318, which is supplied to the user in a state of being incorporated in the apparatus main body in advance.
[0166]
In the present specification, the step of describing the program stored in the recording medium is not limited to the processing performed in chronological order in the order in which they are included, but is not necessarily processed in chronological order, either in parallel or individually. The process to be executed is also included.
[0167]
In the present specification, the system represents the entire apparatus constituted by a plurality of apparatuses.
[0168]
【The invention's effect】
Thus, according to the present invention, image data can be encoded. In particular, when quantization values used in past encoding cannot be reused, quantization can be performed using a quantization matrix in which high frequency components are configured with values less than 255. It is possible to prevent deterioration in image quality.
[0169]
According to another aspect of the present invention, image data can be encoded, and information used in past encoding can be reused, but past quantization can be performed when a quantized value cannot be reused. Quantization can be performed using a quantization matrix that is configured with a smaller value than the quantization matrix used when the information used in can not be reused, so that degradation of image quality is prevented. Can do.
[0170]
In addition, according to another aspect of the present invention, image data can be converted, and in the encoding process in the conversion process, when the quantized value used in the past encoding cannot be reused, the high frequency component is Since quantization can be executed using a quantization matrix configured with a value of less than 255, it is possible to prevent deterioration in image quality.
[0171]
  Furthermore, according to another aspect of the present invention, image data can be converted, and information used in past encoding is reused in encoding processing in conversion processing, but quantization values cannot be reused. In this case, it is possible to perform quantization using a quantization matrix configured with a smaller value than the quantization matrix used when information used in past encoding cannot be reused. Deterioration can be prevented.
Furthermore, according to another aspect of the present invention, re-encoding can be performed. In particular, it is determined whether a quantized value in the past encoding process is reusable, and is determined not to be reusable. In such a case, since the re-encoding is performed using the intra quantization matrix having a smaller AC component value, it is possible to prevent the deterioration of the image quality.
[Brief description of the drawings]
FIG. 1 is a diagram for explaining a conventional system in which re-encoding is performed when frame editing is performed.
FIG. 2 is a diagram for explaining a conventional system capable of re-encoding by changing the bit rate of MPEG Long GOP.
FIG. 3 is a diagram for describing a case where encoding history information is used in a conventional system in which re-encoding is performed when frame editing is performed.
FIG. 4 is a diagram for explaining a case where encoding history information is used in a conventional system that can re-encode by changing the bit rate of MPEG Long GOP.
FIG. 5 is a block diagram showing a system configuration in which re-encoding is performed when frame editing is applied to which the present invention is applied.
6 is a block diagram showing a configuration of an encoding unit in FIG. 5. FIG.
FIG. 7 is a diagram for explaining an intra quantization matrix that is a default in MPEG2.
FIG. 8 is a diagram for explaining an intra quantization matrix used in the present invention.
FIG. 9 is a block diagram showing the configuration of a system that can re-encode by changing the bit rate of MPEG Long GOP to which the present invention is applied;
10 is a block diagram showing a configuration of an encoding unit in FIG. 9. FIG.
FIG. 11 is a flowchart for explaining quantization matrix determination processing 1;
FIG. 12 is a diagram for explaining a non-intra quantization matrix that is a default in MPEG2.
FIG. 13 is a diagram for explaining a non-intra quantization matrix used in the present invention.
FIG. 14 is a flowchart for explaining quantization matrix determination processing 2;
FIG. 15 is a diagram for explaining a configuration of a different apparatus to which the present invention is applicable.
FIG. 16 is a block diagram illustrating a configuration of a personal computer.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 SDI-ASI conversion apparatus, 3 frame editing apparatus, 51 Long GOP encoding apparatus, 61 decoding part, 62 encoding part, 101 ASI-SDTI conversion apparatus, 111 decoding part, 112 encoding part, 121 decoding part, 141 decoding , 151 SDTI CP-ASI converter, 161 encoding unit, 171 history extraction unit, 172 image rearrangement unit, 173 calculation unit, 174 motion vector detection unit, 175 DCT unit, 176 quantization unit, 177 rate setting unit, 178 VLC unit, 179 inverse quantization unit, 180 inverse DCT unit, 181 operation unit, 182 motion compensation unit, 183 frame memory, 184 buffer, 185 memory, 201 Long GOP encoding device, 211 encoding unit, 221 parameter input unit

Claims (15)

In an image processing apparatus that encodes image data,
Receiving means for receiving encoding parameters used in past encoding of the image data;
Compared with the quantization matrix defined in the encoding standard, which is used when the image data is encoded without reusing the quantization value included in the encoding parameter received by the receiving means. Storage means for storing a quantization matrix having a reduced AC component value;
Determining means for determining whether or not to reuse the quantization value included in the encoding parameter received by the receiving means;
In the case where the quantization value included in the encoding parameter received by the receiving unit is reused based on the determination result of the determining unit, the quantization matrix included in the encoding parameter received by the receiving unit Is used to encode the image data,
When the quantization value included in the encoding parameter received by the receiving unit is not reused based on the determination result of the determining unit, the image is obtained using the quantization matrix stored in the storage unit. An image processing apparatus comprising: an encoding unit that encodes data. In the case of encoding the image data without reusing the encoding parameter received by the receiving unit, the encoding unit uses the quantization matrix that is set as a default in the encoding standard. The image processing device according to claim 1, wherein the data is encoded. The image processing apparatus according to claim 1, wherein the quantization matrix is an intra quantization matrix or a non-intra quantization matrix. The image processing apparatus according to claim 1, wherein the quantization matrix stored in the storage unit has a high-frequency AC component value limited to twice or less of a DC component. The image processing apparatus according to claim 1, wherein the quantization matrix stored in the storage unit has a high-frequency AC component value and a low-frequency AC component value that are the same or substantially the same. The image processing apparatus according to claim 1, wherein the encoding unit encodes the image data according to an MPEG standard. The determination means is included in the encoding parameter received by the receiving means when the bit rate for encoding the image data is smaller than the bit rate indicated in the encoding parameter received by the receiving means. The image processing apparatus according to claim 1, wherein it is determined that the quantized value to be reused is not reused. The determination unit does not reuse the quantization value included in the encoding parameter received by the receiving unit when the image data encoded as the I picture in the past encoding is encoded as the I picture. The image processing apparatus according to claim 1. In an image processing method of an image processing apparatus for encoding image data,
Receiving means receives encoding parameters used in past encoding for the image data;
Compared to a quantization matrix defined in a coding standard, which is used when the storage means codes the image data without reusing the quantization value included in the received coding parameter. Stores a quantization matrix with a small AC component value,
Determining means determines whether or not to reuse the quantized value included in the encoding parameter received by the receiving means;
The encoding means is
When the quantization value included in the received encoding parameter is reused based on the determination result of the determination means, the quantization matrix included in the received encoding parameter is reused to recycle the quantization value included in the received encoding parameter. While encoding image data,
If the quantization value included in the received encoding parameter is not reused based on the determination result of the determination means, the step of encoding the image data using the stored quantization matrix An image processing method including: Computer
In an image processing apparatus that encodes image data,
Receiving means for receiving encoding parameters used in past encoding of the image data;
Compared with the quantization matrix defined in the encoding standard, which is used when the image data is encoded without reusing the quantization value included in the encoding parameter received by the receiving means. Storage means for storing a quantization matrix having a reduced AC component value;
Determining means for determining whether or not to reuse the quantization value included in the encoding parameter received by the receiving means;
In the case where the quantization value included in the encoding parameter received by the receiving unit is reused based on the determination result of the determining unit, the quantization matrix included in the encoding parameter received by the receiving unit Is used to encode the image data,
If the quantization value included in the encoding parameter received by the receiving unit is not reused based on the determination result of the determining unit, the image is obtained using the quantization matrix stored in the storage unit. A program that causes an image processing apparatus to function as an image processing apparatus. A recording medium on which the program according to claim 10 is recorded. In an image processing apparatus that re-encodes an encoded stream,
Receiving means for receiving encoding parameters utilized in past encoding for the encoded stream;
Compared to a quantization matrix defined in an encoding standard, which is used when re-encoding the encoded stream without reusing the quantization value included in the encoding parameter received by the receiving means, Storage means for storing a quantization matrix in which the value of the high frequency AC component is reduced;
Determining means for determining whether or not to reuse the quantization value included in the encoding parameter received by the receiving means;
In the case where the quantization value included in the encoding parameter received by the receiving unit is reused based on the determination result of the determining unit, the quantization matrix included in the encoding parameter received by the receiving unit To re-encode the encoded stream,
When the quantization value included in the encoding parameter received by the receiving unit is not reused based on the determination result of the determining unit, the code is obtained using the quantization matrix stored in the storage unit. An image processing apparatus comprising: a re-encoding unit that re-encodes the encoded stream. In an image processing method of an image processing apparatus for re-encoding an encoded stream,
Receiving means receives encoding parameters used in past encoding for the encoded stream;
Compared with a quantization matrix defined in an encoding standard, which is used when the storage means re-encodes the encoded stream without reusing the quantization value included in the received encoding parameter. , Memorize the quantization matrix that reduced the value of the high frequency AC component,
Determining means determines whether or not to reuse the quantized value included in the encoding parameter received by the receiving means;
The re-encoding means is
When the quantization value included in the received encoding parameter is reused based on the determination result of the determination means, the quantization matrix included in the received encoding parameter is reused to recycle the quantization value included in the received encoding parameter. Re-encode the encoded stream,
Based on the determination result of the determination unit, if the quantization value included in the received encoding parameter is not reused, the encoded stream is re-encoded using the stored quantization matrix. An image processing method including the step of: Computer
In an image processing apparatus that re-encodes an encoded stream,
Receiving means for receiving encoding parameters utilized in past encoding for the encoded stream;
Compared to a quantization matrix defined in an encoding standard, which is used when re-encoding the encoded stream without reusing the quantization value included in the encoding parameter received by the receiving means, Storage means for storing a quantization matrix in which the value of the high frequency AC component is reduced;
Determining means for determining whether or not to reuse the quantization value included in the encoding parameter received by the receiving means;
In the case where the quantization value included in the encoding parameter received by the receiving unit is reused based on the determination result of the determining unit, the quantization matrix included in the encoding parameter received by the receiving unit To re-encode the encoded stream,
When the quantization value included in the encoding parameter received by the receiving unit is not reused based on the determination result of the determining unit, the code is obtained using the quantization matrix stored in the storage unit. A program that functions as an image processing apparatus comprising re-encoding means for re-encoding an encoded stream. A recording medium on which the program according to claim 14 is recorded.

Images