Knowledge of video compression algorithm
The basic principle of video compression, the concept of common compression algorithm
Video compression algorithms knowledge of
the MPEG-. 1
the MPEG compressed video coding comprises three elements: I frames (I-frames), P frames (P-frames) and B frames (B-frames). In the MPEG encoding process, the compression portion of the video frame as an I frame sequence; P portion compressed into frames; there are some compression as a B frame. Method I is an intra-frame compression method, also known as "key frame" compression method. I frame compression method is based on the discrete cosine transform DCT (Discrete Cosine Transform), this algorithm is similar to the JPEG compression algorithm. I frame compression can be achieved using a compression ratio of 1/6 compression without significant signs.
Achieved with high image quality at a high compression of the compression algorithm, relying on the intra-frame compression is Bunengshixian, MPEG compression algorithm using a combination of inter and intra. The former method is a P-frame prediction algorithm, which takes into account the same information or data between adjacent frames, i.e. motion characteristics will be considered inter-frame compression. Method is based on the P-frames before the present frame and an adjacent one (I or P frame) differs from the present frame to compress data. Take the I frame and P frame compression methods may be combined to achieve higher compression and no obvious signs of compression.
However, there is only the B frame using compressed ability to achieve 200: 1 compression. Method B frames are bi-predictive inter-frame compression algorithms. When the B frames into a compressed, it is based on the previous frame adjacent to the present frame and the subsequent frame is different from the compressed data of the current frame, i.e. just before and after recording the difference between the present frame and the frame. B-frame data is only fifteen percent of the I-frame data, fifty percent below P-frame data.
A similar standard MPEG 4: 2: 2 format is adopted, the resolution of the luminance signal after the compression is 352 × 240, are the two chrominance signal resolution 176 × 120, the frame rate two different resolution information are 30 frames per second. The basic method of coding in a unit of time, and the acquired first image of the first frame is compressed as an I frame. Then for each frame thereafter, on the basis of a single frame image of the effective compression only store part of the phase changes for the subsequent frame. Inter-frame compression process often spaced intra-frame compression method, as intra (key frames) of a previous frame based compression is not, the key 15 is provided a generally every frame, so that the former can reduce a compression error associated accumulation. MPEG compression encoder first determines the current frame is an I frame or a P or B frame, and then using the corresponding algorithm to compress. After a full video sequence encoded by MPEG compression format is possible: IBBPBBPBBPBBPBBIBBPBBPBBPBBPBBI ......
compressed into a B or P frame to I frame compressed than the calculation processing needs much more time. Some encoders do not even have the B-frame compression P frames, it is clear that the compression effect is not very good.
MPEG-2
MPEG tissue compression standard MPEG-2 launched in 1994, in order to realize the possibility of video / audio services and applications interoperable. MPEG-2 standard is specified for a standard TV and HDTV digital compression scheme and system applications in a variety of layers, the coding rate from 3 to 100 megabits Mbits per second, the official standards in ISO / IEC13818 in. MPEG-2 is not a simple upgrade MPEG-1 is, MPEG-2 and made more specific predetermined further in the system and send intact. MPEG-2 is particularly suitable for encoding and transmitting a digital television broadcast, has been identified as SDTV and HDTV encoding standard. MPEG-2 is also devoted to the multiple programs of rehabilitation tap mode. MPEG-2 standard now divided into nine portions, referred to as ISO / IEC13818 international standard.
MPEG -2 image compression is the use of the image of two properties: the spatial correlation and a temporal correlation. No matter what a scene is composed of a plurality of pixels, and therefore usually present a pixel in an image with some pixels around it in the luminance and chrominance a certain relationship, such a relationship is called spatial correlation; a often show a plot of a sequence of images consisting of a number of frames constituting the continuous image, there is a certain relationship between the front and rear frame image in the image sequence, such a relationship is called the correlation time. These two correlation exists such that a large amount of redundant information in the image. Suppose we can remove the redundant information, retain only a small amount of non-relevant information is transmitted, it is possible to greatly reduce the transmission band. And the receiver using these non-related information, in accordance with certain decoding algorithm, it is possible to ensure a certain premise restore the original image in image quality. A good compression encoding scheme is to maximize the removal of redundant information in the image.
MPEG-2 encoded image is divided into three categories, called I-frames, P-frames and B-frames.
I picture using intra-frame coding mode, i.e., using only spatial correlation within a frame image without using time correlation. I using intra-frame compression, motion compensation is not used, because the I-frame does not depend on other frames, it is a random access point, when the same reference frame is decoded. I-frame is used to acquire the initialization and channel receiver, and the insertion and program switching, a relatively low compression ratio I frame image. I frame image is now a periodic sequence of images, the occurrence frequency selected by the encoder.
P and B frames using inter picture coding mode, i.e. when using the same correlation in space and time. Before the predicted P frame image using only the time, it is possible to improve the compression efficiency and image quality. P-frame picture can be included intra-coded portions, i.e., each can be forward predicted P frame macroblock can also be intra-coded. B bidirectional temporal prediction frame image, can greatly improve the compression ratio. Zhidezhuyi is employed because the next B-frame picture frame as a reference, the transmission order and display order of picture frames in MPEG-2 encoded stream codes are different.
P and B frames using inter picture coding mode, i.e. when using the same correlation in space and time. Before the predicted P frame image using only the time, it is possible to improve the compression efficiency and image quality. P-frame picture can be included intra-coded portions, i.e., each can be forward predicted P frame macroblock can also be intra-coded. B bidirectional temporal prediction frame image, can greatly improve the compression ratio. Zhidezhuyi is employed because the next B-frame picture frame as a reference, the transmission order and display order of picture frames in MPEG-2 encoded stream codes are different.
MPEG-2 encoded bit stream is divided into six levels. In order to better represent the encoded data, MPEG-2 syntax with a predetermined hierarchical structure. It is divided into six layers, from top to bottom each are: picture sequence layer, group of pictures (the GOP), image, bar macroblock, macroblock, block.
4-MPEG
MPEG -4 released in November 1998, MPEG-4 is for video, audio bit rate encoding in certain, more emphasis on interactivity and flexibility of multimedia systems. MPEG-4 standard strive to achieve two objectives: multimedia communication at low bit rates; a plurality of integrated multimedia communications industry. For this reason, MPEG-4 introduces the AV objects (Audio / Visual Objects), so that many other interaction possible:
"AV object" can be an isolated person, can also be a person's voice or some background music. It has a characteristic efficient coding and efficient storage and dissemination interoperable.
Operation of the AV MPEG -4 objects are: AV objects employed to represent the audible, visual, or a combination of audiovisual content; combinations conventional AV AV objects to generate a composite object and thereby generates an AV scene; AV data objects flexibly synthesis and synchronization multiplexes to select the appropriate network to transmit the AV data objects; agreed to receive an end user to interact AV AV object scene and the like.
MPEG-4 standard by 6 major portion configured:
① DMIF (of The Dellivery Multimedia Integration Framework)
DMIF multimedia delivery that is the general framework, it is mainly to solve the interactive network, broadcasting environment and operational issues disk applications multimedia applications. Establishing client and server interaction, and transmitted through a transmission end for multiplex bit information. By DMIF, MPEG4 can be established for each elementary stream having a bandwidth of special quality service (QoS) of the channel and oriented.
② data plane
data plane in MPEG4 can be divided into two parts: the relationship between the transmission portion and the portion of media relations.
In order to make the basic stream and AV objects appear in the same scene, MPEG4 refers to an object describing the concept of narrative (OD) and a flow chart Desktop (SMT) is. Information flow diagram of the basic OD transmission stream associated with a particular AV object. The table for each stream and a CAT (Channel Assosiation Tag) is connected, CAT can achieve a smooth transmission of the stream.
③ buffer management and real-time identification
MPEG4 system defines a decoding mode (the SDM), the decoding mode description describes an ideal processing of syntax and semantics of the bit stream decoding apparatus, it requires special buffer and real-time mode. Through effective management, better use of limited buffer space.
④ audio encoding
MPEG4 is advantageous in that - it not only supports natural sounds, and supports voice synthesis. MPEG4 audio portion of the audio coding and synthesis coding of natural sounds combined, and supports audio object features.
⑤ video coding
similar to the audio coding, MPEG4 coding is also supported for natural and synthetic visual objects. Synthetic visual objects include 2D, 3D animation and facial animation and other people.
⑥ scene at description
MPEG4 provides a set of tools, a set of objects in the scene. Synthesis of information necessary to describe the composition of the scene description, the scene description in a binary format described in BIFS (Binary Format for Scene description), said BIFS transmitted together with the AV object encoder. The main narrative describing the scene at description for each AV objects at a specific coordinate AV scene, how to organize and synchronization issues. The same time there AV AV scene objects and intellectual property protection and other issues. MPEG4 provides a wealth of AV scene for us.
Compared with the MPEG-1 and MPEG-2, MPEG-4 is more suitable for remote monitoring and interactive AV service, It is designed to have a wider adaptability and scalability: MPEG-4 transmission rate 4800-64000bps between a resolution of 176 × 144, a very narrow bandwidth can be utilized by the frame reconstruction data compression and transmission, so that it can get the best image quality with minimal data. Therefore, it will be streaming video on digital television, motion picture, Internet, multimedia real-time monitoring, mobile multimedia communications, Internet / Intranet with the visual aspects of games, interactive multimedia applications on a DVD to show their talents.
The H.264
the H.264 is the ITU-T VCEG (Video Coding Experts Group) and ISO / IEC MPEG (Moving Picture Coding Experts Group) Joint Video Team: a new digital video coding (JVT joint video team) developed standard, it is both the ITU-T H.264, but also a part 10 ISO / IEC MPEG-4 of the. January 1998 began the draft solicitation, in September 1999, completed the first draft, in May 2001 to develop its test mode TML-8, in June 2002, the JVT 5th meeting of the FCD H.264 board. Right now the standards are still under development, the first half of next year is expected to be formally adopted.
H.264 standard and have just as DPCM hybrid coding plus model transform coding. But it is a "back to basics" simplicity of design, not many options, get much better compression performance than H.263 ++; strengthening the ability to adapt to a variety of channels, a "network-friendly" structure and syntax, facilitate processing of the error and packet loss; target range wider application, to meet the different rates, different resolutions and different transmission requirements (storage) of the case; the basic system is open, without the use of the copyright.
H.264 algorithm can be conceptually divided into two layers: a video coding layer (VCL: Video Coding Layer) is responsible for efficient video content represents, Network Abstraction Layer (NAL: Network Abstraction Layer) is responsible for the appropriate manner required by the network and transmitting the data package. H.264 supports 1/4 or 1/8 pixel accuracy motion vectors. 6-tap filter may be used at 1/4 pixel accuracy to reduce high frequency noise, for the motion vector of 1/8 pixel accuracy, using more sophisticated 8-tap filter. Expected during movement, the encoder can also select "enhancing" the effect of the interpolation filter to improve prediction. H.264 entropy coding, there are two ways, one is unified for all symbols to be encoded VLC (UVLC: Universal VLC), there is a use of content adaptive binary arithmetic coding. The draft H.264 includes tools for eliminating the error, to facilitate transmission of compressed video bit error, packet loss multiple environments, such as IP or mobile channel robustness channel transmission.
Technically, H.264 standard, a plurality of the flash, as a unified symbol VLC coding, high-precision, multi-mode displacement is expected, based on an integer of 4 × 4 block transform, layered coding syntax like. These measures make H.264 algorithm has a very high coding efficiency, at the same image quality reconstruction, about 50% of the bit rate can be saved than H.263. H.264 bit stream structure of the network adaptability, add the application error recovery capabilities, can be very well adapted to IP and wireless networks.
H.264 has broad application prospects, such as multi-point communication real-time video communications, Internet video transmission, video streaming services, heterogeneous online, compressed video storage, video databases. H.264 superior performance obtained is not without cost, at the expense of greatly increased computational complexity, it is expected that the computational complexity of encoding corresponds to about 3 times the H.263, H.263 decoding complexity is roughly equivalent to the 2 times.
H.264 recommendations of the technical characteristics can be grouped into three areas, one focusing on useful and mature technology, the pursuit of higher coding efficiency, simple forms; second is to focus on adaptation to mobile and IP networks, stratified technology, the coding form and isolate the channel, substantially in the source encoder algorithms take into account many other channel characteristics; Third, in the basic framework of the hybrid encoder, its key components have done significant improvements, such as multi-mode movement is expected, intra prediction, multi-frame prediction, unified VLC, 4 × 4 two-dimensional integer transform like.
So far, H.264 has not been finalized, but because of a higher compression ratio, better channel adaptation, will be more widely used in the field of communication or storage of digital video, its development potential is limitless.