JP2007228096A - Digital moving picture recording apparatus and digital moving picture recording method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To furthermore reduce a compression coding processing amount before start of recording to a recording medium. <P>SOLUTION: Before starting recording to the recording medium 106, in the digital moving picture recording method, a data format section 203 sets a coding mode to a monitor mode wherein coding is not controlled by a coding efficiency (S40), the data format section 203 set to the monitor mode transmits a control signal to a prediction coding section 202 to be fixed to an in-frame prediction coding mode (S50), and an in-frame prediction direction is fixed to each coding block (S60). <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a digital moving image recording apparatus and a digital moving image recording method for selectively encoding a moving image signal on a frame-by-frame basis using an intra-frame predictive encoding method or an inter-frame predictive encoding method.

  Advances in digital signal processing technology and LSI technology make it possible to compress and encode large amounts of information such as moving images, still images, and audio, and to record and reproduce them on magnetic recording media, optical recording media, and semiconductor recording media. Yes. Development of video recording apparatuses such as an imaging apparatus that easily captures and records video using such a technique and a digital disk recorder that records analog broadcasting has been performed.

  The MPEG system is widely used as a representative example of the moving picture encoding system. FIG. 5 shows an example of a conventional digital moving picture recording apparatus using MPEG technology, and FIG. 6 shows the encoded data. Moving image data is supplied to the MPEG encoding unit 301 from a video input unit 101 such as a camera. In the MPEG encoding unit 301, as shown in FIG. 6, the correlation between the I picture 311 of the intra-frame encoding method that is normally encoded using the spatial correlation in one frame and the temporally previous frame A P picture 312 that is encoded by using B and a B picture 312 of an inter-frame encoding scheme that is encoded by using the correlation between preceding and subsequent frames are used. The video encoded by the MPEG encoding unit 301 is supplied to the recording unit 103 and stored in the recording medium 106. Also, in this video recording apparatus, the encoded stream output unit is adapted so that the encoded stream encoded by the MPEG encoding unit 301 conforms to the interface format such as Ethernet (registered trademark), USB (Universal Serial Bus), IEEE 1394, etc. The data is processed at 107 and output from the encoded stream output terminal 108 to the outside (television in FIG. 5).

  Next, H.264, which is a moving picture coding system that has been studied jointly by ITU-T and ISO MPEG. A frame prediction mode selection method used in a H.264 / MPEG-4 AVC (hereinafter referred to as JVT) test model will be described. In JVT, intra-frame predictive coding is performed in order to improve the compression rate by using pixel correlation between blocks. The block unit to be predicted is 4 × 4 pixels or 16 × 16 pixels. Complex images can be predicted in small block units of 4 × 4 pixels, and flat images can be predicted efficiently by predicting in large macroblock (MB) units of 16 × 16 pixels.

  There are many modes for intra-frame prediction, and a mode that does not generate a code amount is usually selected using a predetermined method (described later). This mode selection for intra-frame prediction greatly affects the coding processing amount. Next, the encoding processing amount of I, P, and B pictures will be described. The main processing of I pictures is intra-frame predictive coding, and the main processing of P pictures is to estimate prediction images from one reference image in addition to I picture processing and to determine prediction modes for each image block. In addition to the processing of the P picture, the main processing of the B picture is to perform an operation for estimating a prediction image from one reference image and determining a prediction mode for each image block.

  In motion picture intraframe prediction (hereinafter referred to as intra prediction), it is important how to obtain this mode at high speed without reducing the coding efficiency. Intra prediction mode selection methods include the method described in Non-Patent Document 1 (hereinafter referred to as “Rate-Distortion Optimum Method”) and the method described in Non-Patent Document 2 (hereinafter referred to as Edge). ) Called a component-based method). The Rate-Djstortion optimization method described in Non-Patent Document 1 is a method for selecting an encoding mode, a motion vector, and the like in each macroblock in an optimal form from the viewpoint of encoding efficiency.

  Here, the intra 4 × 4 mode will be briefly described. In the conventional video coding standard, the intra prediction is performed in the transform coefficient region, whereas in JVT, it is performed in the spatial region from the pixel values in the vicinity of the already adjacent left or upper adjacent block. That is, if the 4 × 4 pixel block to be intra-predicted is the pixels a to p in FIG. 7, the 4 × 4 pixel block has a pixel value in the vicinity of the already adjacent left or upper adjacent block. Predicted from A to L. In this prediction, there are 9 (= 1 + 8) prediction modes in total, that is, prediction using the average value of the pixel values A to L (DC prediction) and prediction from the eight directions shown in FIG. In FIG. 9, 0 (vertical) to 2 (DC prediction) to 8 (horizontal-up) indicate prediction modes.

  Next, for each of the nine encoding modes and motion vectors, a Lagrangian multiplier λ is used for the bit R required for encoding and the distortion D (or a simulation thereof) at that time. To select a case where these are minimized. That is, the mode that minimizes the evaluation function J = D + λR is selected. According to this Rate-Distortion optimal method, it is known that the best coding efficiency can be obtained. In order to obtain the value of the evaluation function J, it is necessary to obtain the bit R necessary for the encoding and the distortion D at that time, so that prediction error calculation, conversion, Quantization, bit number calculation, inverse quantization, inverse transformation, and quantization error calculation are required.

  Next, the method based on the edge component described in Non-Patent Document 2 is a method based on the idea that the prediction direction of the optimal intra prediction mode substantially matches the edge direction of the block. First, create what is called an edge map by applying the Sobel operator to each pixel of the entire screen. This means that A = | Dx | + | Dy | and θ = arctan (Dx / Dy) with respect to the Sobel operator values (Dx, Dy) obtained by applying each pixel in the horizontal and vertical directions. Is obtained. Thereafter, based on the number of intra prediction direction candidates X (excluding DC prediction), all angles are divided into X, and A is added as a histogram of the class to which θ belongs. This process is performed for each pixel in the block of interest, and the class that finally has the largest histogram is handled as the prediction direction of the block. In the case of the intra 4 × 4 mode, directions in the vicinity of both are also treated as candidates. Finally, the Rate-Distortion optimal method is applied to all of the candidate modes (three for intra 4 × 4, one for the others) and the DC prediction mode. That is, the evaluation function J = D + λR is calculated, and the prediction mode that minimizes it is selected. Non-Patent Document 2 below describes that the encoding efficiency is reduced when the directions of both neighbors are not treated as candidates.

  In such a video recording apparatus, when an instruction to start recording on the recording medium 106 is given while an encoded stream including an encoded video is transmitted from the encoded stream output terminal 108, for example, FIG. Let us consider a case where the recording start is instructed at the recording start time T in FIG. 1, that is, at the timing corresponding to the picture 312 in the encoded picture group G04. At this time, the encoding processing amount of the MPEG encoding unit 301 in FIG. 5 does not change at all before and after the recording start time T.

By the way, in Patent Document 1 below, in order to prevent the loss of a reproduced image when recording on a recording medium is started during transmission of an encoded image signal, the ratio of I pictures before the start of recording is set higher than that after the start of recording. It has been disclosed to reduce P pictures or B pictures that are missing during reproduction by increasing the number. In this case, since the number of I pictures increases before the start of recording, the amount of encoding processing is smaller than that after the start of recording because I <P <B as shown in FIG.
JP 2004-194173 A (Abstract) Gary S., Thomas W, and Keng-Pang L. “Joint Model Reference Encoding Methods and Decoding Concealment Methods”; JVT-lOO49dO, Sept, 2003. Feng P, et al, “Fast Mode Decision for Intra Prediction,” JVT-GO13, Mar. 2003.

  However, in the above-described Rate-Distortion optimal method, in order to select a prediction mode that minimizes the evaluation function J = D + λR, prediction error calculation, conversion, quantization, bit number calculation, Inverse quantization, inverse transformation, and quantization error calculation must be performed to obtain the bit R and the distortion D required for the encoding. Therefore, there is a problem that the calculation amount is large and much time is required. is there. Another problem is that unnecessary processing is performed by a CPU (Central Processing Unit) before starting recording on the recording medium, which wastes power. Further, in Patent Document 1, H.264, which encodes a moving image signal by selecting an optimal prediction direction from a plurality of prediction directions within a frame. Since the intra-frame predictive encoding of the H.264 / MPEG-4 AVC method is not considered, there is a problem that the processing amount cannot be reduced so much.

  The present invention has been made in view of the above problems of the prior art, and an object of the present invention is to provide a digital moving image recording apparatus and a digital moving image recording method capable of further reducing the amount of compression encoding processing before starting recording on a recording medium. And

In order to achieve the above object, the present invention provides an intra-frame predictive encoding unit that encodes a video signal by selecting any one of a plurality of prediction directions within a frame, and
Inter-frame prediction encoding means for inter-frame prediction encoding the moving image signal;
Selecting means for selecting data encoded by the intra-frame predictive encoding means or the inter-frame predictive encoding means for each frame;
In a digital moving image recording apparatus comprising recording means for recording data selected by the selection means on a recording medium,
Until the recording unit starts recording on the recording medium, the selection unit selects the data encoded by the intra-frame prediction encoding unit and fixes the prediction direction;
It is characterized by having.

In order to achieve the above object, the present invention provides an intra-frame predictive encoding step of encoding a video signal by selecting any one of a plurality of predictive directions within a frame, and
An inter-frame predictive encoding step for performing inter-frame predictive encoding of the moving image signal;
A selection step of selecting, for each frame, data encoded by the intra-frame prediction encoding step or the inter-frame prediction encoding step;
In a digital moving image recording method comprising a recording step of recording data selected in the selection step on a recording medium,
Until the recording step starts recording on the recording medium, the selecting step selects the data encoded by the intra-frame predictive encoding step and fixes the prediction direction;
It is characterized by having.

  According to the present invention, the encoding method is fixed to the intra-frame prediction encoding method and the prediction direction within the frame is fixed until recording on the recording medium is started, so that encoding before starting recording on the recording medium is performed. The amount of processing can be further reduced.

  Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing an embodiment of a digital moving picture recording apparatus according to the present invention, FIG. 2 is a block diagram showing in detail the prediction encoding unit of FIG. 1, and FIG. 3 is an intra-frame prediction encoding unit of FIG. FIG. 4 is a flowchart for explaining processing of the digital moving image recording apparatus of FIG.

  The video input unit 101 images a subject and supplies a video signal to the MPEG encoding unit 102. The MPEG encoding unit 102 converts the supplied video signal to H.264. It is encoded by the H.264 / MPEG-4 AVC system and supplied to the encoded stream output unit 107 and the recording unit 103. The inside of the MPEG encoding unit 102 includes a pixel block dividing unit 201, a predictive encoding unit 202, and a data formatting unit 203. The video signal supplied from the video input unit 101 is supplied to the predictive coding unit 202 via the pixel block dividing unit 201. As shown in detail in FIG. 2, the predictive coding unit 202 includes an intraframe predictive coding unit 205 and an interframe predictive coding unit 204, and the interframe predictive coding unit 204 includes a bidirectional predictive coding unit 204-1. The forward prediction encoding unit 204-2 and the backward prediction encoding unit 204-3 are configured. The output of the pixel block dividing unit 201 is output from the bidirectional predictive encoding unit 204-1, the forward predictive encoding unit 204-2, the backward predictive encoding unit 204-3, and the intra-frame predictive code by the switching unit (switch) S. Is supplied to the conversion unit 205 by a switching control signal from the data format unit 203.

  Details of the intra-frame prediction encoding unit 205 are shown in FIG. In the intra-frame prediction encoding unit 205, as shown in FIG. 9, in addition to nine prediction modes (Intra 4 × 4) in 4 × 4 pixels, vertical prediction mode, horizontal prediction mode, and average value in 16 × 16 pixels There are four prediction modes and a planar prediction mode (Intra 16 × 16), and there are a total of 13 prediction modes. The intra-frame prediction encoding unit 205 performs encoding using the correlation within the same frame of the video signal (I picture), and the inter-frame prediction encoding unit 204 performs encoding using the inter-frame correlation before and after the video signal ( P, B picture).

  Before the start of recording, the encoding method is not selected in the predictive encoding unit 202, and only intra-frame predictive encoding is performed. At this time, the prediction mode shown in FIG. 3 (and FIG. 9) is also fixed. The video signal encoded by the intraframe prediction encoding unit 205 is supplied to the data formatting unit 203. In addition, after the recording is started by the operation unit 105, any of the encoded video signals output from the intra-frame prediction encoding unit 205 and the inter-frame prediction encoding unit 204 from the control unit 104 via the data formatting unit 203. In accordance with the control signal for selecting one of them, a switching signal is supplied to the predictive coding unit 202, and either one of the encoded video signals output from the intraframe predictive coding unit 205 and the interframe predictive coding unit 204 is selected. . The selected encoded video signal is supplied to the data format unit 203. The data format unit 203 combines the supplied encoded video signals of a predetermined number of blocks to form one picture or a picture group, and supplies it to the encoded stream output unit 107 and the recording unit 103.

  The encoded stream output unit 107 converts the encoded video signal supplied from the data format unit 203 into a form suitable for the transmission path, and an external device such as a television monitor or a computer via the encoded stream output terminal 108. Output to. The encoded stream output unit 107 has an interface circuit corresponding to the encoded stream output terminal 108, and outputs an encoded video signal according to a predetermined protocol. As the encoded stream output terminal 108, any terminal capable of outputting an encoded stream signal, such as an IEEE 1394 terminal or a USB terminal, can be used.

  The control unit 104 includes, for example, a CPU, a ROM (Read Only Memory), and a RAM (Random Access Memory), and by executing a program stored in a recording medium such as a ROM, an optical disk, or a semiconductor memory, the entire digital moving image recording apparatus Control. Predictive mode selection processing such as intra-frame coding and inter-frame coding at the start of recording, which is characteristic of the present invention described below, is also realized by the control unit 104 executing a program and controlling each necessary unit. The Of course, part or all of the processing can be processed by hardware.

  The recording unit 103 modulates the encoded stream according to a predetermined recording method, and records it on the recording medium 106. As the recording medium 106, any recording medium such as a magnetic recording medium or an optical recording medium can be used. For example, in the present embodiment, a recording medium such as a magnetic tape, an optical disk, a hard disk drive, or a semiconductor memory is used.

  In the present invention, the coding mode for each block is controlled without changing the picture type (I, P, B picture) before and after the start of recording on the recording medium 106. This procedure will be described with reference to FIG. First, the digital moving image recording apparatus is turned on (step 10), and various settings are initialized (step 20). Next, if recording is not started in step 25, the process proceeds to step 40 and thereafter. If recording is started and recording on the recording medium 106 is being performed, the process proceeds to step 70. If not, the process proceeds to step 40 (step 30). In step 40, the data format unit 203 sets the encoding mode to a monitor mode that is not controlled by the encoding efficiency. The data format unit 203 set to the monitor mode sends a control signal to the prediction encoding unit 202 so as to “fix to the intraframe prediction encoding mode” (step 50). Further, when a control signal is sent to the predictive coding unit 202, the intra-frame prediction direction is fixed for each coding block (step 60). Step 40, step 50, step 60, and step 100 are repeated until the processing of the last encoded block in the frame is completed.

  On the other hand, in step 70, the data formatting unit 203 is set to the recording mode, and is set so that the encoding efficiency is maximized. In step 80, a coding mode with good coding efficiency is appropriately selected according to the coding block by performing a large amount of calculations. Further, in step 90, an appropriate prediction direction is selected and set for each coding block. Step 70, step 80, step 90, and step 100 are repeated until this process finishes the last process in a coding block. When all the processing of the encoded blocks in the frame is completed (step 100), it is confirmed whether recording is stopped (step 110). If the recording is stopped, the process proceeds to step 25. If the recording is continued, the process proceeds to step 30.

  As described above, the encoding mode and the prediction direction are fixed before the encoded stream starts to be recorded on the recording medium 106 in a predetermined format, and the encoding mode and the prediction direction are changed when recording on the recording medium 106 is started. Make an appropriate selection.

1 is a block diagram showing an embodiment of a digital moving image recording apparatus according to the present invention. It is a block diagram which shows the prediction encoding part of FIG. 1 in detail. FIG. 3 is a block diagram illustrating in detail an intra-frame prediction encoding unit in FIG. 2. 3 is a flowchart for explaining processing of the digital moving image recording apparatus of FIG. 1. It is a block diagram which shows the conventional digital moving image recording device. It is explanatory drawing which shows the coding data of the digital moving image recording device shown in FIG. It is explanatory drawing which shows the prediction encoding in a flame | frame. It is explanatory drawing which shows the prediction direction in a flame | frame. It is explanatory drawing which shows the prediction direction in a flame | frame. It is explanatory drawing which shows the ratio of the encoding processing amount of an I, P, B picture.

Explanation of symbols

DESCRIPTION OF SYMBOLS 101 Video input part 102 MPEG encoding part 103 Recording part 104 Control part 105 Operation part 106 Recording medium 107 Encoded stream output part 108 Encoded stream output terminal 201 Pixel block division part 202 Predictive encoding part 203 Data format part 204 Between frames Prediction coding unit 205 Intraframe prediction coding unit

Claims (2)

Intra-frame predictive encoding means for encoding a video signal by selecting any one of a plurality of prediction directions within a frame;
Inter-frame prediction encoding means for inter-frame prediction encoding the moving image signal;
Selecting means for selecting data encoded by the intra-frame predictive encoding means or the inter-frame predictive encoding means for each frame;
In a digital moving image recording apparatus comprising recording means for recording data selected by the selection means on a recording medium,
Until the recording unit starts recording on the recording medium, the selection unit selects the data encoded by the intra-frame prediction encoding unit and fixes the prediction direction;
A digital moving image recording apparatus comprising: An intra-frame prediction encoding step of encoding a video signal by selecting any one of a plurality of prediction directions within a frame;
An inter-frame predictive encoding step for performing inter-frame predictive encoding of the moving image signal;
A selection step of selecting, for each frame, data encoded by the intra-frame prediction encoding step or the inter-frame prediction encoding step;
In a digital moving image recording method comprising a recording step of recording data selected in the selection step on a recording medium,
Until the recording step starts recording on the recording medium, the selecting step selects the data encoded by the intra-frame predictive encoding step and fixes the prediction direction;
A digital moving image recording method comprising:

Images