"1" development and characteristics of MPEG
1.1 MPEG-1
Before the emergence of MPEG, there were two standards for image compression, namely JPEG for still image data compression and H.261 for video telephone and conference video image compression, but they are independent of computer data standards. This requires the formulation of a unified standard for computer system, radio and television in the four aspects of image, sound, storage and transmission, which is conducive to the wide communication of various media. Therefore, MPEG came into being.
The basic task of MPEG-1 standard is that the image (including audio) data with appropriate quality must become a kind of computer data, and be compatible with the existing data (such as text, drawing and other data) in the computer, and these data must be transmitted compatible in the existing computer network, broadcast television and other communication networks. MPEG-1 standard has three components: MPEG video, MPEG audio and MPEG system. Therefore, MPEG involves the problems of video compression, audio compression and the composition and synchronization of a variety of compressed data streams. MPEG-1 standard is a coding standard for digital storage media dynamic images and their accompanying sound suitable for transmitting 1.5mbps data transmission rate. It can process various types of moving images. Its basic algorithm has a good effect on compressing the spatial resolution of 360 pixels in the horizontal direction and 288 pixels in the vertical direction, and the moving images of 24 to 30 pictures per second. Unlike JPEG, it does not define the detailed algorithm required to generate legal data stream, but provides a lot of flexibility in encoder design. In addition, a series of parameters defining the encoded bit stream and decoder are included in the bit stream itself. These characteristics allow the algorithm to be used for images with different sizes and width ratios, It can also be used in channels and devices with a wide range of operating rates.
MPEG-1 standard compression first subsamples the color difference signal to reduce the amount of data, adopts motion compensation technology to reduce inter frame redundancy, uses two-dimensional DCT transform cloud to remove spatial correlation, quantizes the DCT components, discards unimportant information, sorts the quantized DCT components according to the frequency, and encodes the DCT components with variable word length, Finally, the DC component (DC) of each data block is predicted by differential coding. The coding and decoding block diagram of MPEG video is shown in Figure 1.
1.2 MPEG-2
The full name of MPEG-2 standard is "coding of moving image and its accompanying sound", which is mainly aimed at the video and accompanying sound signals required by high definition television (HDTV), and the transmission rate is 10Mbps.
MPEG-2 standard is divided into eight parts, collectively referred to as ISO / iec1318 international standard. Part I: system, which describes the method of synthesizing transmission code stream and program code stream from multiple video, audio and data basic code streams; The second part: video, describes the video coding method; The third part: audio, which describes the audio coding method reverse compatible with MPEG-1 audio standard; The fourth part: Conformance test, which describes the software implementation method of the first, second and third parts of testing whether a coded code stream conforms to MPEG-2 standard; Part 5: digital memory - command and control, which describes the session signaling set between server and user in interactive multimedia network; Part VI; Non backward compatible audio, specifying multi-channel audio coding that is not backward compatible with MPEG-1 audio; Part VII: 10 bit video, now stopped; Part VIII: real time interface, which specifies the real-time interface for transmitting code stream.
MPEG-2 video coding standard is a series divided by level, which is divided into four "levels" according to the resolution of the encoded image: low level (LL: low level), and the pixel of the input signal is one quarter of itu-r601 format; Main level (ml: main level), the pixel of the input signal is itu-r601; Advanced-1440 (h14l: high-1440 level) is a 4:3 mode TV high-definition format; HL: high level is the high definition format of 16:9 mode TV. According to the set of coding tools used, it is divided into five "classes": simple profile (SP): only reference frame I and prediction frame p; The main class (MP: main profile) adds two-way speculation frame B than sp; Signal to clutter ratio layered class (SNR: SNR scalable profile); Spatial scalable profile (SSP); High profile (HP: high profile)“ Several combinations of "class" and "class" constitute a subset of MPEG-2 video coding standard in a specific application. For an image of an input format, a specific set of compression coding tools are used to generate a coded code stream within a specified rate range. MPEG-2 code stream is divided into six levels. From top to bottom: video sequence layer; Image group layer (GOP: group of picture); Image layer (picture); Slice; Macroblock layer and image block layer.
MPEG-2 coding process: in the case of intra coding, the encoded image only passes through DCT, and the quantizer and bit stream encoder generate the encoded bit stream without prediction loop processing. DCT is directly applied to the original image data. In the case of inter coding, the original image is first compared with the predicted image in the frame memory to calculate the motion vector, and the predicted image of the original image is generated from the motion vector and the reference frame. Then, the difference image data generated by the difference between the original image and the predicted pixel is DCT transformed, and then the output encoded bit stream is generated by the quantizer and the bit stream encoder.
1.3 MPEG-4
The goal of MPEG-4 standard is to support a variety of multimedia applications (mainly focusing on the access to multimedia information content), and the decoder can be configured on site according to different application requirements. MPEG-4 aims to provide a flexible framework and a set of open coding tools for the communication, access and management of video (audio) data.
In MPEG-4 image and video standard, the goal of video representation tool is to provide standardized core technology for the effective storage, transmission and management of texture, image and video data in multimedia environment. The ability of these tools to encode and decode atomic units of image and video content (called video object VO) is particularly emphasized. The video object with arbitrary shape is effectively represented to support the so-called content-based feature set. This feature set supports the separate encoding and decoding of content (i.e. the physical object VO in the scene). This feature not only provides strong underlying mechanism support for interactivity, but also provides favorable conditions for flexible representation and management of image or video VO content in the compression domain. MPEG-4 image and video standard uniformly supports the encoding and decoding of traditional rectangular and arbitrary shape images and videos. For content-based applications, the input image sequence may have any shape and position. The shape can be represented by an 8-bit transparent component (when a VO is composed of multiple other objects) or described by a binary mask. In addition, by using appropriate and fine object-based motion prediction tools for each physical object in the scene, the compression ratio of some video sequences can be greatly improved. The content-based coding for MPEG-4 extension can be regarded as the logical extension of the traditional vlbv kernel or HBV tool from rectangular input to arbitrary shape input. In this sense, content-based coding is a superset of vlbv and HBV kernels.
MPEG-4 standard adds seven new functions on the original basis. Features of added functions:
(1) Content based operation and bitstream editing support content-based operation and bitstream editing without coding（ 2) Mixed coding of natural and synthetic data. It provides a way to effectively combine natural video images with synthetic data (text and graphics), and supports interactive operation（ 3) Enhanced time domain random access. MPEG-4 will provide an effective random access method: in a limited time interval, an audio and video sequence can be accessed randomly according to frames or objects of any shape（ 4) Improve coding efficiency. At a comparable rate with the existing emerging standards, MPEG-4 standard will provide better subjective visual quality images（ 5) Coding of multiple concurrent data streams. MPEG-4 will provide effective multi view coding for a scene, plus multi sound channel coding and effective audio-visual synchronization. In the aspect of stereo video application, MPEG-4 will effectively describe three-dimensional natural scenery under sufficient observation viewpoint conditions by using the information redundancy caused by multi viewpoint observation of the same scene（ 6) The "flexibility and diversity" of error resilience in error prone environment means that various wired networks and various storage media are allowed. MPEG-4 will improve the error resilience, especially in low bit applications (mobile communication links) in the environment prone to serious errors. MPEG-4 is the first standard to consider channel characteristics in its audio and video representation specifications. Its purpose is not to replace the error control technology provided by the communication network, but to provide a resilience against residual errors（ 7) Content based scale variability. Content scale variability means that priority is assigned to each object in the image. Content-based scale variability is the core of MPEG-4, because once the directory and corresponding priority of the objects contained in the image are determined, the functions of other content sets are easier to implement. For very low bit rate applications, variable scale is a key factor because it provides the ability to adapt to available resources.
The above seven new functions can be divided into three categories: content-based interaction, high compression rate and flexible access modes. The first three functions are content-based interaction, the fourth and fifth are high compression mode, and the last two are flexible access modes.
1.4 MPEG-7
MPEG-7 standard is called "multimedia content description interface", which will expand the limited capabilities of existing content recognition dedicated solutions, especially including more data types. In other words, MPEG-7 will specify a standard set of descriptors for describing various types of multimedia information.
MPEG-7 standardizes the method of defining other descriptors and their structures (description schemes) and the relationship between them. This description (that is, the combination of descriptor and description scheme) will be associated with the content itself, so as to search the materials of interest to users quickly and efficiently. MPEG-7 standardizes a language used to define description schemes, namely description definition language (DDL). In addition, the AV material of relevant MPEG-7 data can be indexed and retrieved.
Like other members of the MPEG family, MPEG-7 is a standardized representation of audio and video information that meets specific needs. The MPEG-7 descriptor does not depend on the way the described content is encoded or stored. The MPEG-7 description can be attached to analog movies or pictures printed on paper. However, although the MPEG-7 description does not depend on the (coding) representation of the processed material, to a certain extent, it is developed on the basis of MPEG-4, which adopts the processing method of audio and video coding according to objects with certain temporal and spatial relationships, Therefore, it is possible to attach the description to the members (objects) in the scene by MPEG-4 coding. Therefore, MPEG-7 needs to provide different degrees in the description to realize different levels of recognition.
Because the description feature must be meaningful in the application environment, it will be different due to different user ranges and application fields. This means that the same material may be described with different types of features to match the scope of application. Of course, all these descriptions will be encoded in an efficient way, which can improve the efficiency of search. At the same time, there may be a transitional level of abstraction in the middle. Abstract level is related to the way of feature extraction. Many low-level features can be extracted automatically, while high-level features need more manual interaction.
2 future prospects of MPEG
MPEG video compression system is a complex integrated system with high technical content. Only a few powerful companies can launch commercial products in the world. Due to the complex technology and expensive equipment of MPEG video compression system, its popularity is very low so far. However, with the progress of technology, the maturity of process and the decline of price, its application is expanding. What used to cost millions of yuan to accomplish, now tens of thousands of yuan can be achieved. If you have such an MPEG video compression system, you can easily compress and store videos, photos, pictures, movies and other programs in the computer, life videos, file management and other video production.
MPEG formulates a series of standards. In fact, in many cases, no specific implementation is given. The final implementation should be realized by various manufacturers and R & D personnel. The research on MPEG mainly focuses on two aspects: (1) the research on the implementation of MPEG（ 2) Further research on image compression methods has achieved greater data compression ratio and realized the function of man-machine dialogue.
From the current MPGE standard, the author believes that it will mainly focus on the object-based processing method, that is, for different data, content and requirements, different methods will be selected according to the situation. First of all, this is to meet the most basic requirements of man-machine dialogue and the requirements of people-oriented purpose. Everyone can adopt different processing methods according to their own needs; Secondly, this is the requirement to further obtain a larger image data compression ratio. The previous compression methods based on the data itself and its exchange and statistical personality are difficult to meet the data flow speed on the highway. However, if the object-based processing method is a real model-based compression method, it can adopt different compression methods for different objects (contents), so as to obtain a huge compression ratio and meet people's visual requirements. This problem has been noticed in MPEG-4 and MPEG-7 standards, and even the research of object or content has been introduced. Therefore, the author believes that object-based image processing method will be the development direction of MPEG in the future.
MPGE video compression technology and VCD production have opened up a new development path for us. The popularization and application of MPEG video compression technology may produce a new industry, namely multimedia production. The market in this area has just started, and there is almost no space in education and training. It is an industry with great development potential, which needs to be explored. The future is an information society, and the transmission and storage of various multimedia data is the basic problem of information processing. This paper only expounds in detail from the aspect of MPEG standard, and there are still many technologies in this field to be studied and developed. It is hoped that people interested in this research will discuss it together., Read the full text, technical section
Tech supports Amazon (AWS) media services to provide quality assurance for end-to-end video
Im
Our other product:
