JP4178521B2 - Encoded video signal recording method and video signal encoding apparatus - Google Patents

Description

The present invention relates to recording and playback of video signals compressed and encoded by moving picture experts group (MPEG) or the like, and in particular, video signals recorded in real time by a transport stream are played back at high speed and video played back at high speed. The present invention relates to an encoded video signal recording method and video signal encoding apparatus for performing re-encoding for converting a signal image size and recording at high speed.

  2. Description of the Related Art Conventionally, video signal recording apparatuses and reproducing apparatuses using an encoding method such as MPEG have been developed. MPEG video standards are ISO / IEC1172-2, ITU-T-H. 262, ISO / IEC13818-2. The MPEG system standard is ITU-TH. 222, ISO / IEC13818-1. The MPEG stream includes a programming stream used for DVD and the like, and an MPEG-2 multiplexing standard used for digital broadcasting and the like, that is, a transport stream standard. Transport streams are also used for disc media such as Blu-rays.

  Regarding video signal recording devices, hybrid recording devices equipped with both optical desk media recorders and large-capacity hard disk drive recorders have been developed. In a hybrid recording device, for example, a method of first recording on a hard disk drive at a high rate, then converting to a rate and re-recording on the disk is also used. The rate conversion is usually aimed at conversion to a low rate, and the rate is set according to the capacity of the recording medium to be re-recorded. In particular, when re-encoding at a low rate and recording, it is common to re-encode the video signal with a reduced resolution by converting the image size of the video in order to avoid a major video corruption. . In the first compression encoding and recording, a video signal input in real time is encoded and recorded in real time. However, it is desirable that the next re-recording can be performed at a high speed in accordance with the recordable speed of the disc. It is desirable that re-recording can be performed at the same speed as a recording medium capable of recording at high speed. In some cases, re-recording is further compressed by re-encoding processing, rate-converted, and recorded on the disc.

Japanese Patent Application Laid-Open No. H10-228561 discloses a method of converting and re-encoding an image size. As a method of reducing the amount of signal processing without degrading the accuracy of motion vector detection for motion prediction compensation when the image size of the encoded image data is converted and re-encoded, the encoded image data is When decoding, temporarily store a motion vector, perform a motion vector whose resolution has been converted based on the stored motion vector, select a motion vector to be used for re-encoding from the converted motion vector, Re-encoding processing is performed.
JP-A-10-336672

  However, conventionally, a re-encoding method for performing processing at a recording speed on a recording medium capable of recording at high speed has not been realized. Although re-encoding by the method disclosed in Patent Document 1 can extract a motion vector when decoding image data, the processing time is increased by performing motion compensation prediction using a plurality of candidate vectors. In addition, sufficient performance of high-speed re-encoding has not been obtained, for example, since motion vector scaling (normalization related to resolution conversion) is required during re-encoding. The processing speed required at the time of re-encoding needs, for example, 8 × speed or 16 × speed, which is the DVD recording speed. There is a further need to reduce the processing time for performing high-speed re-encoding.

Accordingly, the present invention has been made to solve the above-described problems, and compresses and encodes an input video signal in real time and records it on a built-in recording medium, and also performs the recorded encoding. An object of the present invention is to provide an encoded video signal recording method and video signal encoding apparatus capable of reproducing a signal and generating a further re-encoded signal and recording it on a portable recording medium at a high speed. To do.

In order to solve the above problems, the present invention provides the following method and apparatus.
(1) A video packet in a predetermined stream format obtained by obtaining a motion vector of an input video signal and encoding a first compressed video signal obtained by encoding at a first compression rate using the motion vector. after recording on the first recording medium is inserted into said first recording medium the first compressed video signal recorded by decoding to obtain a decoded video signal, the following prior Symbol decoding video In a method for recording an encoded video signal, wherein a second compressed video signal obtained by encoding a signal at a second compression rate higher than the first compression rate is recorded on a second recording medium.
Based on the motion vectors had use in generating the first compressed video signal a first step of generating conversion motion vectors used for coding of the second compressed video signal,
The transformed motion vector is inserted into an invalid packet positioned in front of the video packet into which the first compressed video signal related to the transformed motion vector is inserted, and the first motion video inserted into the video packet A second step of recording on the first recording medium together with a compressed video signal ;
A third step of obtaining the transformed motion vector from the first recording medium and obtaining a transformed video signal in which the size of the decoded video signal obtained by decoding the first compressed video signal is reduced ;
Said conversion using motion vectors, records how encoded video signal characterized by comprising further a fourth step of the converting video signal is encoded to obtain a second compressed video signal.
(2) In the fourth step, when the converted motion vector is not obtained in the third step, encoding is performed using a motion vector generated from the converted video signal. (1) A method for recording an encoded video signal according to (1).
(3) A video packet in a predetermined stream format obtained by obtaining a motion vector of an input video signal and encoding the first compressed video signal obtained by encoding at a first compression rate using the motion vector. Each time it is inserted and recorded on the first recording medium, the first compressed video signal recorded on the first recording medium is decoded to obtain a decoded video signal, and then the decoded video In a video signal encoding apparatus for recording a second compressed video signal obtained by encoding a signal at a second compression rate higher than the first compression rate on a second recording medium,
Means for generating a transformed motion vector used for encoding the second compressed video signal based on the motion vector used when generating the first compression rate video signal;
The transformed motion vector is inserted into an invalid packet positioned in front of the video packet into which the first compressed video signal related to the transformed motion vector is inserted, and the first motion video inserted into the video packet Means for recording on the first recording medium together with a compressed video signal;
Means for obtaining the converted motion vector from the first recording medium and obtaining a converted video signal obtained by reducing the size of the decoded video signal obtained by decoding the first compressed video signal;
Means for encoding the transformed video signal to obtain the second compressed video signal using the transformed motion vector;
A video signal encoding apparatus comprising:
(4) The means for obtaining the second compressed video signal, when the converted motion vector is not obtained by the means for obtaining the converted video signal, performs encoding using the motion vector generated from the converted video signal. The video signal encoding apparatus as set forth in (3), wherein

According to the present invention, a first step of generating a transformed motion vector used for encoding the second compressed video signal based on the motion vector used when generating the first compression rate video signal; A second step of recording the converted motion vector on the first recording medium in addition to the already recorded first compressed video signal; and reproducing the first recording medium to obtain the converted motion vector. A third step of obtaining a converted video signal by converting the size of the decoded video signal obtained by decoding the first compressed video signal, and encoding the converted video signal using the converted motion vector And the fourth step of obtaining the second compressed video signal, and the main part comprising the fourth step, the input video signal is compressed and encoded in real time and recorded on the built-in recording medium. To generate a re-coded signal further compressed by reproducing the recorded encoded signals, the recording method of encoded video signals can be recorded at high speed on a portable recording medium, and realize a video signal encoding apparatus it can.

A video signal encoding apparatus according to each embodiment of the present invention will be described below with reference to FIGS.
FIG. 1 is a block diagram showing a schematic configuration example of a video signal encoding apparatus according to the first embodiment of the present invention.
FIG. 2 is a block diagram showing a configuration example of the video encoder unit according to the first embodiment of the present invention.
FIG. 3 is a block diagram showing a configuration example of the multiplexing processing unit according to the first embodiment of the present invention.
FIG. 4 is a diagram showing an example of inserting an encoded information packet according to the first embodiment of the present invention.
FIG. 5 is a diagram showing a packet configuration example of a transport stream according to the embodiment of the present invention.
FIG. 6 is a diagram showing a description example of a motion vector information packet according to the embodiment of the present invention.
FIG. 7 is a diagram showing a description example of motion vector information data according to the embodiment of the present invention.
FIG. 8 is a block diagram illustrating a configuration example of the separation processing unit according to the first embodiment of the present invention.
FIG. 9 is a block diagram illustrating a configuration example of the video decoder unit according to the first embodiment of the present invention.
FIG. 10 is a diagram illustrating an example of rearrangement data performed by the re-encoded motion vector information rearrangement unit according to the first embodiment of the present invention.
FIG. 11 is a block diagram illustrating a configuration example of the motion compensation prediction unit according to the first embodiment of the present invention.
FIG. 12 is a block diagram illustrating a configuration example of a motion compensation prediction unit according to the second embodiment of the present invention.
FIG. 13 is a block diagram illustrating a configuration example of a motion compensation prediction unit according to the third embodiment of the present invention.
FIG. 14 is a diagram showing the concept of hierarchical motion vector detection according to the third embodiment of the present invention.
FIG. 15 is a diagram illustrating an example of detecting a hierarchical motion vector according to the third embodiment of the present invention.

  The video signal encoding device is used to further compress the recorded signal and re-record it on a portable recording medium when the input video signal is compressed and encoded in real time and recorded on a built-in recording medium. Re-encoding motion vectors are generated and recorded together on the built-in recording medium. A decoded video signal obtained by decoding a compression-coded signal recorded on a built-in recording medium is re-encoded using a re-encoding motion vector previously recorded at a higher compression rate. The re-encoding speed enables high-speed re-encoding at the speed of a portable recording medium capable of recording a high-speed signal, and a high-compression encoded signal can be recorded on the portable recording medium at a high speed.

The configuration of the video signal encoding device will be described.
The video signal encoding apparatus shown in FIG. 1 includes a video encoder unit 1, a multiplexing processing unit 2, a recording medium unit 3, a decoded image size conversion unit 4, a separation processing unit 5, and a video decoder unit 6. . The recording medium unit 3 includes a hard disk 31 and a DVD (Digital versatile Disc) 32 as recording media. A video signal is input to the video encoder unit 1 via the switch S1. The recording mode information obtained by the user's operation is input to the video encoder unit 1, the multiplexing processing unit 2, the decoded image size conversion unit 4, the separation processing unit 5, and the video decoder unit 6. The video decoder unit 6 outputs a decoded video signal. An audio ES (Elementary Stream) is output from the separation processing unit 5.
2 includes a first image memory 11, a subtractor 12a, an adder 12b, a DCT (discrete cosine transform) unit 13, a quantizer 14, a VLC (variable length coding) unit 15, a motion compensation. It comprises a predictor 16, an inverse quantizer 17, an inverse quantizer 18, and a second image memory 19. A video signal is input to the first image memory 11. Re-encoded motion vector information and recording mode information are input to the motion compensation predictor 16. A video ES (Elementary Stream) is output from the VLC unit 15. Motion vector information is output from the motion compensation predictor 16.

The multiplex processing unit 2 shown in FIG. 3 includes a TS packet generator 21, a TS (Transport Stream) -MUX (multiplex) unit 22, and a motion vector information TS packet generator 23.
8 includes a re-encoded motion vector information extractor 51, a TS-DEMUX (demultiplex) unit 52, and an ES extractor 53.
9 includes a VLD (variable length decoding) unit 61, an inverse quantizer 62, an inverse DCT unit 63, an adder 64, a motion compensation predictor 65, a buffer memory 66, and a re-encoded motion vector. An information sorter 67 is included.
The motion compensation predictor 16 shown in FIG. 11 includes an image size conversion circuit 161, a motion vector detection circuit 162, a motion compensation circuit 163, and a motion vector information positive circuit 164.

The operation of the video signal encoding device will be described.
First, as recording mode information, a video signal of a program obtained by reception is recorded in the hard disk 31 in the high image quality mode, and mode information for recording the program on the DVD 32 by compression encoding so that the program can be recorded on one side is illustrated. Not input from the operation unit. Next, the switch S1 is turned on to the a side, and the video signal of the received program is input to the video encoder unit 1. In the video encoder unit 1, a video signal input in real time is encoded in real time with high image quality. The motion vector of the input video signal is obtained, and the motion-compensated image signal is encoded by DCT (discrete cosine transform) operation and variable length encoding. The motion vector obtained at the time of encoding is used when a further compressed encoded signal is generated by re-encoding and recorded on the DVD 32 at an arbitrary time after real-time recording.

  The video encoder unit 1 encodes a video elementary stream (ES) obtained by encoding, an audio ES obtained by encoding an audio signal of a received program with an audio encoder (not shown), a motion vector used during re-encoding, and The recording mode information is supplied to the multiplexing processing unit 2, where the inputted signals are multiplexed to generate a transport stream (TS). The generated TS is input to the hard disk 31 of the recording medium unit 3 and recorded there.

  The TS signal recorded on the hard disk 31 is reproduced as necessary. The demultiplexing unit 5 demultiplexes the signals multiplexed by the multiplexing unit 2. The video ES obtained by the separation is decoded by the video decoder unit 6. The video signal obtained by decoding is displayed on the monitor TV. The audio ES obtained by the separation is decoded by an audio decoder (not shown), and the audio signal obtained by the decoding is supplied to a speaker (not shown) to be sounded. All the recorded parts recorded on the hard disk 31 or the designated part of the recorded program are recorded on the DVD 32 by the user. The DVD 32 has a predetermined recording capacity and is a portable recording medium capable of high-speed recording such as 8 × speed or 16 × speed.

Here, the video signal recorded with high quality on the hard disk 31 is converted into a video signal having a data amount that can be recorded on one side of the DVD 32, re-encoded, and recorded on the DVD 32. Since the recording is not performed while viewing the program, it is desirable that the recording be performed as fast as possible.
The TS recorded on the hard disk 31 is read at high speed. The read TS is separated by the separation processing unit 5. The video ES and the re-encoding motion vector obtained by the separation are supplied to the video decoder unit 6. In the video decoder 6, the input video ES is decoded to obtain a decoded video signal. The encoded video signal is input to the decoded image size conversion unit 4 and converted into an image having a small number (size) of pixels.

The switch S2 is set to the b side when the data amount to be recorded on the DVD 32 can be reduced, and the switch S2 is turned to the a side when the data amount needs to be further reduced. The decoded video signal or the size-converted decoded image is input to the video decoder unit 6 via the switch S1 inserted on the b side. The re-encoding motion vector is generated as re-encoded motion vector information rearranged in a predetermined order by the video decoder unit 6 and input to the video encoder unit 1.
In the video encoder unit 1, the input decoded video signal or the size-converted decoded image and the re-encoded motion vector information are input together, and the re-encoded video ES is generated at a high speed. The multiplexing processing unit 2 generates a signal in which the input video ES and audio ES are multiplexed. A multiplexed signal is recorded on the DVD 32 of the recording medium unit 3 at a high speed.

This will be described in detail.
At the time of the first real time recording in the video encoder unit 1 shown in FIG. 2, the input video signal is encoded in real time, and a signal recorded on the hard disk 31 is output. Furthermore, the generated motion vector is converted into motion vector information for re-encoding that is used at the time of the second high-speed recording in which the signal recorded on the hard disk 31 is re-encoded and recorded on the DVD 32 and output. .
At the first real-time encoding, motion vector detection performed by so-called MPEG (moving picture experts group) encoding, orthogonal transform by DCT, zigzag scanning, and variable length encoding are performed to obtain a video elementary stream (ES ) Is generated and output together with re-encoding motion vector information.

  First, image data for a plurality of frames is stored in the first image memory 11 as video signals input during real-time recording. The stored image data is data in which luminance data is 16 × 16 pixels and color data (Cb / Cr) is blocked in units of 16 × 8 pixels. The color data is sub-sampled every other pixel in the vertical direction to form a block image of 8 × 8 pixels. Blocked image data includes image data read from the first image memory 11 in the motion compensation prediction 16, and locally decoded reproduced image data of the frame (field) read from the second image memory 19. It becomes more. In the motion compensated prediction 16, the movement positions of all block images of the previous frame (field) with respect to the image data input from the first image memory 11 are calculated by searching from a range of 32 × 32 pixels, for example. An inter-field (inter-field) prediction motion vector is detected. The search for a motion vector is a process that requires the most computation time at the time of encoding. Motion compensation is performed using the motion vector and the reproduced image data read from the second image memory 19. Motion compensated image data, motion vectors, and prediction modes are output from the motion compensated predictor 16. The motion vector information generated when the recording mode is real-time recording is vector information that can be used at the next re-encoding. That is, the motion vector information when re-encoding is performed with size conversion is converted into a motion vector amount based on the number of pixels based on the image size to be converted. The converted motion vector is supplied to the VLC unit 15.

The video signal is encoded at the time of real-time encoding in the same manner as defined in MPEG-2 video. The subtracter 12 subtracts the image data read from the first image memory 11 and the motion compensated video data to obtain difference image data. The differential image data is input to the DCT unit 13, subjected to discrete cosine transform, and DCT coefficients are output. The quantizer 14 receives DCT coefficients, performs quantization, outputs quantized image data, and outputs quantization information related to the quantization table. The VLC unit 15 receives quantized data, motion vectors, and prediction modes. A video elementary stream (ES) obtained by variable-length encoding them is output.
An I (Intra-coded) picture and a P (Predictive-coded) picture are necessary because they are used as reference image data for motion compensation prediction for a B (Bidirectionally predictive-coded) picture. In the inverse quantizer 17, the quantized data is inversely quantized and a DCT coefficient is output. In the inverse DCT unit 18, DCT coefficients are input and inverse DCT is performed to obtain a differential reproduction image. The adder 12b adds the difference reproduction image and the motion compensation image data to obtain reproduction image data. The second image memory 19 receives and stores reproduced image data.

In the video encoder unit 1, re-encoding accompanied by size conversion of an input video signal will be described.
First, the size of the image data read from the first image memory 11 is converted. Similarly, the size of the locally decoded reproduced image data of the frame (field) read from the second image memory 19 is converted. Re-encoded motion vector information generated at the time of real-time encoding is used as a motion vector related to the inter-frame (inter-field) prediction of the size-converted image data. The re-encoded motion vector for each predetermined block is generated as motion vector information together with the encoded picture type. The input re-encoded motion vector is used, and motion compensation is performed based on the picture type of the video signal to be re-encoded. As a picture type, a P picture is predicted from the previous I picture or the previous P picture. A B picture is predicted from an I picture or a P picture on both sides of the past and the future in terms of time.
The re-encoding performed by the video encoder unit 1 is an operation for encoding an input size-converted video signal using a re-encoded motion vector generated at the time of real-time encoding and subjected to size conversion processing. . The signal itself obtained by encoding is the same as a signal generated by normal MPEG-2 or the like. However, since a search operation obtained by calculating a motion vector is unnecessary, it is different in that high-speed encoding is possible.

  In the TS packet generator 21 shown in FIG. 3, the input video ES and audio ES are multiplexed to generate an AV transport stream (TS). In the TS packet generator 23, the input re-encoding motion vector information is generated as a motion vector information TS packet. The TS-MUX unit 22 generates and outputs a transport stream in which TS packets describing motion vector information are inserted during the invalid packet period of the AV transport stream.

  FIG. 4 shows a transport stream output from the multiplexing processing unit 2 with motion vector information TS packets inserted. Motion vector information TS packets are inserted in the period of invalid packets indicated as N0, N1,..., Nn after the TS packets, followed by video and audio packets. By placing the motion vector information before the video packet related to the motion vector information, management of the video stream and the motion vector information is facilitated. When the recording mode is high-speed re-encoded recording, the motion vector information packet is not generated and is not multiplexed with the transport stream.

  FIG. 5 shows a packet structure of a transport stream output from the multiplexing processing unit 2. The structure is the same as that of a TS packet used in MPEG2. The motion vector information for re-encoding is 184 bytes during the invalid packet period in which NullpacketID (0x1fff) is set in PID (Packet Identification), unit start display is “1”, and adaptation field control is set to “01” (payload only). It is inserted into the payload and transmitted.

  FIG. 6 shows a format example of the payload signal. Following the 1-byte data identification, re-encoding motion vector information is transmitted. The data identification describes whether the payload of the invalid packet is motion vector information or the payload of a normal invalid packet.

FIG. 7 is a configuration example of motion vector information data. Motion vector information includes picture types related to pictures I, P, and B, image size identification at the time of re-encoding, prediction modes following each macro block address, and re-encoding motion vectors such as motion vectors for each macro block. Configured for each.
It is also possible to create and transmit re-encoding motion vectors for a plurality of image sizes. In this case, re-encoding motion vectors are described for each different re-encoded image size. When the motion vector information exceeds 184 bytes of the payload, it is described over a plurality of null packets.

  In the separation processing unit 5 in FIG. 8, each of the video packet, the audio packet, and the invalid packet is separated from the transport stream read from the recording medium by the TS-DEMUX unit 52 and output. The re-encoded motion vector information extractor 51 re-encoded image based on the re-encoded image size identification of the invalid packet separated by the TS-DEMUX unit 52 based on the recording mode information set for re-encoding. A size identification and re-encoding motion vector is extracted and output as re-encoded motion vector information. The ES extractor 53 extracts each of the video ES and audio ES from the input video and audio packets and outputs them.

In the video decoder unit 6 shown in FIG. 9, the variable length code signal of the input video ES is decoded by the VLD unit 61 for each block, and motion vector information and quantized image data are output. In the inverse quantizer 62, the quantized image data is inversely quantized to obtain DCT image data (DCT coefficient). In the inverse DCT unit 63, the DCT coefficient is subjected to inverse DCT processing to obtain difference image data. The difference image data is input to the adder 64 and is added to the image data from the motion compensation predictor 65 to obtain a decoded video signal. It is stored in the buffer memory 66. The motion compensation predictor 65 receives the image data stored in the buffer memory 66 and the motion vector information from the VLD unit 61, performs motion compensation based on these input signals, and adds the decoded video signal to the adder. 64.
The re-encoded motion vector rearranger 67 generates and outputs re-encoded motion vector information used at the time of re-encoding and recording based on the input re-encoded motion vector information and recording mode information. The buffer memory 66 stores re-encoded motion vector information.

  FIG. 10 shows a temporal relationship between the re-encoded motion vector information input to the video decoder unit 6 and the re-encoded motion vector information output from the video decoder unit 6. The decoded video signal stored in the buffer memory 66 is read out and output in the order of the normal video signal based on the picture replacement order shown in FIG. The picture input to the video decoder unit 6 (decoder) and the picture output from the video decoder unit 6 are interchanged as shown in FIG. The time relationship between the picture and the re-encoding motion vector in the video encoder unit 1 is matched, and the re-encoding operation is performed.

  In the motion compensated predictor 16 of the first embodiment shown in FIG. 11, when the recording mode is real-time recording, b is selected for each of the switches S3 and S4 since it is the first encoding. Image data that has not undergone image size conversion is input to the motion vector detection circuit 162. The motion vector detection circuit 162 detects a motion vector in block units for a predetermined search range with respect to the input image data, using the input reproduced image as reference data. Motion vector candidates obtained by detection are output. In S5, a is selected, and the motion compensation circuit 163 inputs the motion vector and the reproduced image data, performs motion compensation, and determines a final motion vector. The motion compensated image data generated thereby and the motion vector / prediction mode are output.

  The image size conversion circuit 161 receives image data, which is converted into re-encoded image size data described in the recording mode information and output. A is selected for each of the switches S3 and S4, and the reproduced image data after the image size conversion is input to the motion vector detection circuit 162. The motion vector detection circuit 162 performs motion vector detection on a block basis for a predetermined search range with respect to the input image data, using the input reproduced image as reference data, and outputs a motion vector candidate obtained by the detection. Is done. In S5, a is selected, and the motion compensation circuit 163 receives the motion vector and the reproduced image data subjected to image size conversion, and performs motion compensation. The final motion vector is determined and the motion vector / prediction mode is output in the format shown in FIG.

  In the case where the recording mode is re-encoding, the switch S1 is connected to b and the switch S2 is connected to a in the image size conversion circuit 161, and the size is converted by the decoded image size conversion unit 4 (S2 is b When connected to, reproduced image data is input. Since the motion vector detection circuit 162 does not detect the motion vector of the reproduced image data, the image size conversion circuit 161 does not perform the image size conversion of the reproduced image data. For each of S3 and S4, a is selected. However, when the image size at the time of re-encoding is the same as that at the time of the first encoding, b is selected for each of S3 and S4. In S5, b is selected. Re-encoded motion vector information and reproduced image data are input to the motion compensation circuit 163. Motion compensation is performed, and motion compensated image data and motion vector / prediction mode are output. The motion vector information generation circuit 164 does not generate motion vector information. As described above, since motion vector detection is not performed during re-encoding, the processing time for motion compensation prediction can be significantly reduced. Re-encoding of reproduced image data can be executed at high speed.

A second embodiment of the present invention will be described with reference to FIG. Parts having the same functions as those of the first embodiment are denoted by the same reference numerals and description thereof is omitted.
The configuration will be described.
The motion compensation predictor 16a shown in the figure is different from the motion compensation predictor 16 shown in FIG. 11 only in that the motion compensation circuit 163 includes a motion compensation circuit 163a. The motion compensated predictor 16a can record the re-encoded video signal on the DVD 32 even when a part or all of the re-encoding motion vector cannot be generated due to insufficient calculation time during real-time encoding. It is what. That is, at the time of re-encoding, the motion vector detection circuit 162 has a function of detecting motion vectors in units of blocks in a predetermined search range with respect to re-encoded image data using the input reproduction image as reference data. Therefore, it has a function of obtaining and outputting motion vector candidates independently.

The operation will be described.
First, the motion compensation circuit 163a receives the motion vector output from the motion vector detection circuit 162 and re-encoded motion vector information obtained at the time of real-time encoding. Motion compensation is performed based on the two motion vectors, and a motion vector used at the time of re-encoding is determined. Motion compensated image data and motion vector / prediction mode are output. When the image size at the time of re-encoding is the same as that at the time of the first encoding, b is selected in each of the switches S3 and S4, and the image size conversion of the image data and reproduction image data in the image size conversion circuit 161 is performed. I will not.

As a result, although the processing time of motion compensation prediction at the time of re-encoding cannot be reduced for a reproduced video portion for which re-encoded motion vector information is not obtained, motion vector detection is performed on the decoded video signal. Even when partially re-encoded motion vector information is not obtained, it is possible to prevent an accident in which recording on the DVD 32 fails due to the information.
In addition, since the motion vector detection operation for the decoded video signal is performed in a time of about real time, the high speed is reduced according to the ratio of the real time processing, but in that case, motion compensation with increased candidates is performed. Therefore, the accuracy of motion compensation prediction can be improved.

As a motion vector detection method suitable for real-time processing, there is a hierarchical motion vector detection method.
A third embodiment of the present invention will be described with reference to FIGS. Parts having the same functions as those of the first embodiment are denoted by the same reference numerals and description thereof is omitted.
The configuration will be described.
The motion compensated predictor 16b shown in FIG. 13 is different from the motion compensated predictor 16 shown in FIG. 11 in that the motion vector detection circuit 162 includes a hierarchical motion vector detection circuit 162a and a motion vector memory 162b. . That is, the motion compensated predictor 16b uses the hierarchical motion vector detection circuit 162a for motion vector detection, and sets the size of the motion vector detection image data of each layer and the size of the reproduced image data as a reference to the image data size at the time of re-encoding. The difference is that the converted motion vector is detected.

The operation will be described.
In the hierarchical motion vector detection method, the input image to be subjected to motion compensation and the comparison image are subsampled in the horizontal direction and the vertical direction for each layer, and a small reduced image is created in the upper layer, and the reduced image A wide range of motion vectors is required. In the lower layer, the motion vector range obtained in the upper layer is converted into a vector in the lower layer, and a further search is performed with the position of the obtained motion vector as the center to obtain a highly accurate motion vector. The motion vector is efficiently obtained in a wide search range.

  The hierarchical motion vector detection concept shown in FIG. 14 shows an example of hierarchical motion vector detection using three layers. Motion vectors are detected for the three layers MVD1, MVD2, and MVD3 from the top. The input image and the reference image used for each layer are images of MVD1 and MVD3, which are reduced images subsampled from the upper layer, and the same magnification as MVD2. First, a motion vector is detected in block units for a predetermined search range centered on (0, 0) in MVD1. In MVD2, the reduction ratio and position of each motion vector obtained in the upper layer are corrected, and a motion vector in a block unit is detected for a predetermined search range around the position indicated by the motion vector obtained in the upper layer. . In MVD3, the reduction ratio and position of the motion vector obtained in MDV2 are corrected, and the motion vector is detected in block units for a predetermined search range with the position indicated by the motion vector obtained in MDV2 as the center.

  Here, in the real-time recording mode, for example, if the image size of the reduced image MVD1 or MVD2 is set to the image size after size conversion used at the time of re-encoding, the reduced image converted by the image size conversion circuit 161 is reproduced. An encoding motion vector is obtained. A is selected for both the switches S3 and S4, and the hierarchical motion vector detection circuit 162a detects a motion vector for the input image data. The detected motion vector is temporarily stored in the motion vector memory 162b. In real time recording, a is selected in the switch S5, and the motion vector is also output to the motion compensation circuit 163. The motion compensation circuit 163 performs motion compensation based on the input motion vector and the reproduced image data that has undergone image size conversion. A final motion vector at the set image size is determined, and a motion vector / prediction mode is output.

  Next, in the hierarchical motion vector detection circuit 162a, the motion vector candidates stored in the motion vector memory and obtained previously are read out and normalized to the current image size. With the position indicated by the vector candidate as the center (0,0), a motion vector is detected in block units for a predetermined search range for the input image data. The obtained motion vector candidates are temporarily stored in the motion vector memory 162b. This motion vector candidate is also output to the motion compensation circuit 163 via the switch S5. The motion compensation circuit 163 receives the motion vector candidate and the reproduced image data whose image size has been converted. Motion compensation is performed, a final motion vector at the image size is determined, and a motion vector / prediction mode is output. Next, the switches S3 and S4 are connected to the terminal b, and the same-size image data that is not subjected to image size conversion is input to the hierarchical motion vector detection circuit 162a. The hierarchical motion vector detection circuit 162a reads motion vector candidates stored in the motion vector memory 162b. The motion vector candidates are normalized to the current image size (same size), and motion vector detection within a predetermined search range is required.

The motion vector is input to the motion compensation circuit 163. The motion compensation circuit 163 receives the motion vector and the size-converted image data, and performs motion compensation. The final motion vector is determined and the motion vector / prediction mode is output. The motion vector information generation circuit 164 generates and outputs the motion vector information shown in FIG. The motion vector obtained in the hierarchy that handles the reduced image can be used as a motion vector at the time of re-encoding performed by reducing the image size.
At the time of re-encoding, a video ES whose image size has been converted may be generated using the image size conversion circuit 161 instead of the decoded image size conversion unit 4.

  The hierarchical motion vector can be detected by the method shown in FIG. In other words, the method shown in FIG. 14 generates motion vectors using images of three layers, whereas the method shown in FIG. 15 uses a reduced image having a different size for each image area. Generate. This is a method in which a motion vector candidate is determined by MVD1 for a region containing a lot of motion components, and a motion vector candidate is determined by MVD2 for a region having a small motion component. In the method shown in FIG. 15, in addition to the operation in the first embodiment, it is possible to reduce the motion vector detection processing step for each image size in motion compensation prediction at the time of the first encoding.

According to the video signal encoding apparatus described in the first to third embodiments, the second video signal is generated based on the motion vector obtained when the input video signal is encoded into the first compressed video signal. A re-encoding motion vector used when re-encoding and generating the compressed video signal is generated, the first compressed video signal and the re-encoding motion vector are recorded on the built-in recording medium, and the built-in recording The medium is reproduced to obtain a re-encoding motion vector, and the decoded video signal obtained by decoding the first compressed video signal is resized to obtain a converted video signal, and the re-encoding motion vector is used. Since the converted video signal is encoded to obtain a second compressed video signal, the input video signal is compressed and encoded in real time and recorded on a built-in recording medium. Chemical The generate a re-coded signal further compressed by reproducing can be realized a video signal coding apparatus capable of high-speed recording on a portable recording medium.
In the case where the generation time of the re-encoding motion vector is insufficient at the time of real-time encoding, the first recorded in the hard disk 31 can be provided by providing a function for generating the re-encoding motion vector that is insufficient at the time of re-encoding. It is possible to prevent a generation error of the second compressed video encoded signal generated based on the compressed video signal.
Further, if the real-time video encoder has a function of detecting a hierarchical motion vector and a layer that handles the same number of pixels as the size-converted image is provided in the hierarchy, the first compressed video signal and the second compressed video are provided. Motion vectors for both signals can be created easily and with high accuracy.

Although MPEG-2 is mainly described as an encoding method, the above-described method is also applied to other compression encoding that requires time to calculate a vector for the encoding method performed by reducing temporal redundancy. Can be applied.
Also, the case where a DVD 32 recorder is built in has been described. However, the DVD recording apparatus transmits the TS signal in the format shown in FIGS. 4 to 7 to a DVD recording apparatus provided outside the video signal encoding apparatus. The second compressed video signal can be recorded while obtaining the re-encoding motion vector information from the signal transmitted as the stuffing information and performing the re-encoding at high speed. In this case, an external device that does not have the function using the re-encoding motion vector information ignores the stuffing bits, so that the second video signal can be normally generated and recorded by the conventional method.
Furthermore, since the re-encoding motion vector information is recorded as stuffing information on the hard disk 31 together with the first video signal, the hard disk 31 has a separate area for storing the re-encoding motion vector information. Area management such as managing
In the above example, the TS is recorded on the DVD. A PS (Program Stream) is recorded on a normal DVD. The DVD 32 is described as an example of a portable recording medium. The recording medium is not limited to a DVD, and any other recording medium such as a semiconductor memory may be used as long as it can record a compression-coded signal.
Furthermore, the second compressed video signal may be recorded in another area of the hard disk 31 instead of being recorded on the DVD 32. In that case, the encoded signal compressed with the second code amount can be stored as a library. Any desired second video signal recorded on the hard disk 31 can be copied to a portable recording medium and reproduced at the destination.

  Video signal for high-speed playback of compressed and encoded video signals recorded in real time, converting the image size of the video signals that were played back at high speed, and re-recording the compressed encoded signals on a portable recording medium at high speed It can be applied to an encoding device.

It is a block diagram which shows the schematic structural example of the video signal encoding apparatus which concerns on the 1st implementation of this invention. It is a block diagram which shows the structural example of the video encoder part which concerns on the 1st implementation of this invention. It is a block diagram which shows the structural example of the multiplexing process part which concerns on the 1st implementation of this invention. It is a figure which shows the example of insertion of the encoding information packet which concerns on 1st implementation of this invention It is a figure which shows the packet structural example of the transport stream which concerns on implementation of this invention. It is a figure which shows the example of a description of the motion vector information packet which concerns on implementation of this invention. It is a figure which shows the example of description of the motion vector information data based on implementation of this invention. It is a block diagram which shows the structural example of the separation process part which concerns on the 1st implementation of this invention. It is a block diagram which shows the structural example of the video decoder part which concerns on the 1st implementation of this invention. It is a figure which shows the example of rearrangement data made by the re-encoding motion vector information rearrangement part which concerns on 1st implementation of this invention. It is a block diagram which shows the structural example of the motion compensation estimation part which concerns on the 1st implementation of this invention. It is a block diagram which shows the structural example of the motion compensation estimation part which concerns on the 2nd implementation of this invention. It is a block diagram which shows the structural example of the motion compensation estimation part which concerns on the 3rd implementation of this invention. It is a figure which shows the concept of the hierarchical motion vector detection which concerns on the 3rd implementation of this invention. It is a figure which shows the example of a detection of the hierarchical motion vector which concerns on the 3rd implementation of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Video encoder part 2 Multiplexing process part 3 Recording medium part 4 Decoded image size conversion part 5 Separation processing part 6 Video decoder part 11 1st image memory 12a Subtractor 12b Adder 13 DCT unit 14 Quantizer 15 VLC 16, 16 a, 16 b Motion compensation predictor 17 Inverse quantizer 18 Inverse quantizer 19 Second image memory 21 TS packet generator 22 TS-MUX unit 23 Motion vector information TS packet generator 31 Hard disk 32 DVD
51 Re-encoded motion vector information extractor 52 TS-DEMUX unit 53 ES extractor 61 VLD unit 62 Inverse quantizer 63 Inverse DCT unit 64 Adder 65 Motion compensation predictor 66 Buffer memory 67 Re-encoded motion vector information rearrangement 161 Image size conversion circuit 162 Motion vector detection circuit 162a Motion vector hierarchical motion vector detection circuit 162b Motion vector memory 163, 163a Motion compensation circuit 164 Motion vector information positive circuit

Claims (4)

A motion vector of an input video signal is obtained , and a first compressed video signal obtained by encoding at a first compression rate using this motion vector is inserted for each video packet in a predetermined stream format. after recording on the first recording medium and to obtain a first decoded video signal by decoding said first compressed video signal recorded on the recording medium, the pre-Symbol decoded video signal to the next In a method of recording an encoded video signal, the second compressed video signal obtained by encoding at a second compression rate higher than the first compression rate is recorded on a second recording medium.
Based on the motion vectors had use in generating the first compressed video signal a first step of generating conversion motion vectors used for coding of the second compressed video signal,
The transformed motion vector is inserted into an invalid packet positioned in front of the video packet into which the first compressed video signal related to the transformed motion vector is inserted, and the first motion video inserted into the video packet A second step of recording on the first recording medium together with a compressed video signal ;
A third step of obtaining the transformed motion vector from the first recording medium and obtaining a transformed video signal in which the size of the decoded video signal obtained by decoding the first compressed video signal is reduced ;
A method of recording an encoded video signal, comprising: a fourth step of encoding the converted video signal using the converted motion vector to obtain the second compressed video signal. 2. The fourth step is characterized in that if the converted motion vector is not obtained in the third step, encoding is performed using a motion vector generated from the converted video signal. Recording method of encoded video signal. A motion vector of an input video signal is obtained, and a first compressed video signal obtained by encoding at a first compression rate using this motion vector is inserted for each video packet in a predetermined stream format. Then, after recording on the first recording medium, the first compressed video signal recorded on the first recording medium is decoded to obtain a decoded video signal, and then the decoded video signal is In a video signal encoding apparatus for recording a second compressed video signal obtained by encoding at a second compression rate higher than the first compression rate on a second recording medium,
Means for generating a transformed motion vector used for encoding the second compressed video signal based on the motion vector used when generating the first compression rate video signal;
The transformed motion vector is inserted into an invalid packet positioned in front of the video packet into which the first compressed video signal related to the transformed motion vector is inserted, and the first motion video inserted into the video packet Means for recording on the first recording medium together with a compressed video signal;
Means for obtaining the converted motion vector from the first recording medium and obtaining a converted video signal obtained by reducing the size of the decoded video signal obtained by decoding the first compressed video signal;
Means for encoding the transformed video signal to obtain the second compressed video signal using the transformed motion vector;
A video signal encoding apparatus comprising: The means for obtaining the second compressed video signal performs encoding using the motion vector generated from the converted video signal when the converted motion vector is not obtained in the means for obtaining the converted video signal. 4. A video signal encoding apparatus according to claim 3, wherein

Images