JP2009111797A - Encoder, and recording device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To solve the problem that although, for example, a SDRAM is usable as a reference memory for an encoder, it is considered that the memory for the encoder, a memory for drawing processing, or a memory used for processes with an external device is used in common for a recording medium to reduce the cost and size, but when the SDRAM is used in common, a device needs to perform high-speed processing to secure a bandwidth for each process and functions need to be limited for a limited bandwidth. <P>SOLUTION: Disclosed is the encoder (encoding device) 100 configured to store an image read out of the reference memory in an intermediate memory and then supply it from the intermediate memory 009 to a motion compensation unit 003, and halving access to the reference memory when the image is encoded using reference data common to different frames to improve usability; and the recording device. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

The present invention relates to an encoder that compresses a moving image or a still image and converts it into a bit stream.

Image data encoding technology uses redundancy inherent in image information,
MPEG2 (ISO / IEC 13813-2, International Standard), which is compressed by orthogonal transform such as discrete cosine transform and motion compensation, is used for encoding for broadcasting and recording on recording media.
In addition, an encoding method called H.264 (ITE / ISO 14496-10 / H.264AVC) / AVC (Advanced Video Coding) that supports higher compression has also started to be adopted as an encoding method for recording on recording media. Being started. For both MPEG2 and H.264 encoding, a motion vector as a motion search result is used as encoding information.
The generation of the motion search information is performed when the encoding is performed such as intra-frame coding (intra-coded), inter-frame prediction coding (predictive-codec), bi-directional prediction coding (bidirectionally-predictive-codec), and the like. The difference from the frame is processed. The configuration of the reference frame at the time of encoding is disclosed in Patent Document 1.

JP 2007-104231 A

  However, since the common reference data is used in different frames at the time of bi-directional predictive encoding, an image stored on the same surface of the reference memory is read each time.

  The reference memory can be realized by, for example, SDRAM. However, when an encoder is incorporated in the recording apparatus, the memory used for the encoder, the memory for drawing processing, and the memory used for processing with an external device to realize cost and miniaturization Can be used in common. When SDRAM is shared, it is necessary to make the device a high-speed processing version in order to secure a bandwidth for each processing, and it is necessary to limit functions so that the bandwidth can be accommodated.

  In the present invention, in order to solve the above-described problem, an image read from the reference memory is stored in the intermediate memory and is supplied from the intermediate memory to the motion compensation unit. When encoding using common reference data in different frames, An object of the present invention is to provide an encoder and a recording apparatus that perform encoding with improved usability because the access to the reference memory can be halved.

  The above object is solved by the invention described in the claims.

  According to the present invention, an encoder and a recording apparatus with improved usability can be provided.

The following embodiment will be described as an example of encoding using the H.264 method.
However, the present invention is not limited to this, and there are MPEG4, H.261, H.263, SMPTE VC1 and the like in addition to the MPEG2 system and the H.264 system.

  Note that, in the H.264 method or the like, multiframe motion compensation is possible in which a reference frame for motion compensation can be arbitrarily selected from frames that have already been decoded.

  As a preferred embodiment of the present invention, the configuration and operation of an encoder will be described.

  The configuration of the embodiment will be described with reference to FIG. An input unit 001 that inputs image data to be encoded, a buffer unit 002 that stores the image data supplied from the input unit 001 in the buffer as an input image, and moves between the input image and a plurality of already encoded reference images Generated from a motion compensation unit 003 capable of performing compensation, a frequency conversion unit 004 that converts the frequency of an error image after motion compensation, and VLC units 005 and 006 that perform encoding with a syntax compliant with the standard, and two VLC units. A switching unit 007 that performs control to switch streams in time series, an output unit 008 that outputs an image stream supplied from the switching unit, and a part of a search range of an encoded image is read from a later-described reference memory unit 010 and stored. An encoder (encoding device) 100 including an intermediate memory 009 and a reference memory unit 010 that is a reference image storage unit that uses an encoded image as a reference image for later motion compensation is illustrated.

  For example, the buffer unit 002 and the reference memory unit 010 may be assigned to different addresses on a common SDRAM.

FIG. 2 shows an outline of the present embodiment.
In the figure, I, B, and P indicate the picture coding type (picture_coding_type), I is encoding using only information in the screen (intra-coded), and P is encoding that can also use past information. (predictive-codec), B represents a type for performing encoding (bidirectionally-predictive-codec) using past and future information.

  The upper part shows bi-directional predictive encoding. When B1 is encoded, the difference between the hatched corresponding macroblock and the search range near the corresponding macroblock of I frame: I0 and P frame: P3 is calculated. Similarly, B0 refers to I0 and P3, and B1 and B2 refer to the same I and P image data, respectively.

  In encoding, since an image encoded as shown in FIG. 3 is used as a later reference image, it must be stored in a reference memory. An example will be described in which encoding is performed using images stored in the reference memory using three reference memories.

  When bi-predictive encoding B3 in FIG. 3, for the past frame I2 and the future frame P5 from the reference memory unit 010, the image data around the corresponding macroblock is read, and the motion compensation unit 003 performs the calculation. Similarly, even when B4 is bidirectionally predicted, the same data in the reference memory unit 010 is repeatedly read in order to refer to the image data of I2 and P5 as in B3.

  FIG. 4 is a conceptual diagram in which image data read from the reference memory unit 010 is temporarily stored in the intermediate memory 009 to reduce access to the reference memory.

Unlike FIG. 3, when encoding B3, the image data read from the reference memory unit 010 is stored in the intermediate memory 009, and B2 and B3 are processed simultaneously.
At this time, the motion compensation unit 003 and the frequency conversion unit 004 may encode the two frames by multiplying the processing speed of the encoder by half the normal time and multiplying the time.
As shown in FIG. 5, it is good also as a structure of two each.

  The generated stream is output from the first VLC unit 005 and the second VLC unit 006 in time series. With this configuration, the reference image data read from the reference memory unit 010 can be halved.

  According to the present embodiment, the number of accesses to the reference memory during encoding can be reduced, and the reference memory and other buffers can be configured with a common memory, thereby making it possible to configure a user-friendly encoder.

  Next, as another preferred embodiment of the present invention, an application example to another product will be described. In the present embodiment, an application example to a recording apparatus (for example, a camcorder) will be described as a first practical product example.

  In this embodiment, for example, image data from an image sensor such as a camera or digital image content is input from the outside, and an encoding format and an encoding rate are set on a recording medium such as a hard disk (hereinafter referred to as HDD) or a magneto-optical disk. This is effective when changing and saving.

  FIG. 6 shows a block diagram when the captured image data from the camera is stored in a recording medium. The image data input from the image input interface unit 711 is encoded by the encoding device 707 to generate a stream. The generated stream is recorded by a recording medium interface unit 708 on a recording medium unit 709 such as an HDD.

  In order to play back the recorded stream, the stream is read from the HDD, decoded by the decoder 705, and if the image is output to a display, the decoded image is sent to the display unit 710.

  The reproduced image decoded by the decoder may be encoded again in order to perform bit rate conversion and format conversion of the moving image.

  This configuration makes it possible to reduce the frequency of access to the encoder memory, so that the memory used outside the encoding device 707 and the encoder memory including the reference memory can be placed on a common memory, making it easy to use. Can be provided.

  Next, a second product practical example (for example, a recorder) using this embodiment will be described.

  This embodiment is effective when analog and digital television broadcasts and programs that have already been recorded are stored in recording media such as HDDs and DVDs with different encoding formats and encoding rates.

FIG. 7 shows a block diagram when digital broadcasting is stored in a recording medium.
The digital broadcast received by the tuner unit 701 is decoded by the demodulation unit 702 and decomposed into moving image information and audio information by the demultiplexer unit 703. The moving image information is decoded by the decoder 705, and if the image is output to a display, the decoded image is sent to the display unit 710. Also, the reproduced image decoded by the decoder is encoded by an encoder to generate a stream.

  According to the above embodiment, it is possible to achieve high image quality of bit rate conversion and format conversion of moving images, and usability is improved.

Block diagram of the first embodiment Conceptual diagram of Example 1 Conventional reference memory configuration example Reference memory configuration example Reference memory configuration example Product example to which Example 1 is applied Product example to which Example 1 is applied

Explanation of symbols

100 Encoding device

Claims (8)

For encoding processing that refers to image data for a common search range in different first and second input frames,
The image data for the search range read from the reference memory is stored in the intermediate memory,
The image data read from the intermediate memory is configured to be used for encoding each of the first input frame and the second input frame,
An encoder characterized in that the number of times image data is read from a reference memory is reduced. The encoder according to claim 1, wherein
For the third input frame F2 and the fourth input frame F3 that perform bidirectional predictive encoding with reference to the first input frame F0 and the second input frame F1,
Save image data for a search range from the reference memory to the intermediate memory,
Based on the image data stored in the intermediate memory, the input frame F2 and the input frame F3 are encoded simultaneously,
An encoder characterized in that the processing time required for encoding is completed within a processing time of 2 frames. The encoder according to claim 1, wherein
For the second input frame F2 and the third input frame F3 that perform inter-frame predictive coding with the first input frame F0 as a reference target,
Save image data for a search range from the reference memory to the intermediate memory,
Based on the image data stored in the intermediate memory, the input frame F2 and the input frame F3 are encoded simultaneously,
An encoder characterized in that the processing time required for encoding is completed within a processing time of 2 frames. The encoder according to claim 1, wherein
The search range at the time of detecting vector information when performing inter-frame predictive encoding or bi-directional predictive encoding stored in the intermediate memory,
An encoder characterized by several pixels around the search target position. The encoder according to claim 1, wherein
The intermediate memory is arranged independently of the reference memory;
An encoder that shares a memory other than the reference memory and the intermediate memory. The encoder according to claim 1, wherein
An encoder characterized in that an image stream obtained by encoding and generating two different frames is output in time series within a period for encoding two frames. An image data output from an image input interface having a recording medium interface and capturing a moving image or a still image is encoded by the encoder according to claim 1, and the encoded image stream output from the encoder is recorded. A recording apparatus for recording on a medium. A recording medium interface for reading an image stream recorded on the recording medium and encoded by the first encoding method, a decoder to which the image stream read by the recording medium interface is input, and decoding by the decoder The encoded image data is input, and the encoder according to claim 1,
The recording apparatus, wherein the recording medium interface records an image stream output from the decoder and encoded by the second encoding method on a recording medium.

Images