Napísal Ing. Ján Ivanka - Department of Electrotechnics and Measurements, Faculty of Applied Informatics, Tomas Bata University in Zlín

The article shows some basic properties of digital television’s standards that are in the present time successfully used in various technical applications. The contribution describes standards MPEG-1, MPEG-2 and MPEG-4 in more details. Technical describing of proposed standards is focused to digital image data.

INTRODUCTION
MPEG is abbreviation for Moving Picture Expert Group, the commission, which is dealing with the progression of the standards for the video compression and connected audio signal. The group is working under ISO organization – International Standards Organization.

There are few standards for the MPEG video compression, which satisfy to different areas of use. MPEG video compression is applicable in many multimedia applications, for example in video communication for low transmission in phone networks. It reaches the duality of VHS video systems in speed from 1 to 15 Mb/s (4). The standards define only syntax of code (out) signal, with the process by decoding, but they do not provide connection of the coder or decoder. The aim of the MPEG standard is definition of the source algorithm of coding with high level of the flexibility, which is useful in many different applications.

MPEG -1
This most used standard is designed with the regard for CD technology with the parameters of the maximum data flow up 1,5 Mb/s. For the common users is the most accessible compression and transference standard pro the transfer of the video and related audio signal. In principle the standard MPEG-1 is defined up to the picture size of 4095×4095x60 (60 shots per second). The MPEG-1 standard is consists of four parts:

• IS 11172-1 describes synchronization and multiplexation of video and audio signal,
• IS 11172-2 describes compression of the non-interlaced video signal,
• IS 11172-3 describes compression of the connected audio signal,
• IS 11172-4 describes testing of the unity of transferred data with the original data.

MPEG standard is exploiting mainly by digital video technology. There is usage of the line interleaving and picture format CIF. CIF has following basic attributes:

• Space distinctive aptitude 352 x 288 pictures points,
• Picture frequency 30 Hz,
• Bit rate without the compression 36 Mb/s,
• Bit rate with the compression 1,2 Mb/s.

The compression of picture information is very similar to the compression diagram of JPEG. The wireless compression part represented by Huffman coding is used for the quantized coefficients derivable from discrete cosine transformation, and also for so called kinetic vectors, which are in the diagram of MPEG (as compared with the coding of the static pictures) totally new.

In comparison with the coding of the similar static picture, there must be for the video sequence incorporated just consequential sequence of separate pictures. Basic knowledge is established on the prediction of the “movement” of separate parts of the picture from frame to frame. Moreover, with regard to area of usage of MPEG-1 this standard has to allowed accidental access to video sequence, ensure stationary picture, slow and quick review onward and back.

For a balancing the quality of the video replay together with the best possible compression ratio MPEG-1 standard works in a little bit complicated schema with three types of the pictures.

• Picture (I) with the coding inside of the picture,
• Picture (P) in the midst of the picture with the predictive coding,
• Picture (B) in the midst of the picture with interpolation coding.

In the video sequence the pictures are located according Picture 1.

Picture 1: Pictures location I, B, P in the video sequence

The first type is static picture type „I“(intraframe). These pictures present static pictures in the way, where can be seen as the unrolled classic film made by chemical process. These pictures do not contain any information about previous pictures. The existence of such pictures is obvious – for example, the first picture of the whole video must be the type „I“. The pictures I create holds and they allow incident entry to the video sequence. They have middle compression of the data.

The second type is type „P“ (predicted). For the interpretation of this picture – from view of the decompression – there is necessary to have the knowledge of the previous decoding of the picture type „I“ or type „P“. In the picture type „P“ is therefore included one of the previous picture and the picture himself also contains the moving vectors. These moving vectors determine how separate details changed compare to the previous picture, which is fully reconstructed. The pictures P have bigger compression than the pictures I.

The last type of the picture, which can be used in the compression and transmission MPEG-1 standard is type „B“ (bidirectional). The name of this picture shows, that for the decoding of this picture is necessary to have the knowledge of the most nearest types of the pictures of type „I“ or type „P“ – one of these, which in the picture sequence is located before and the other after the relevant picture of „B“. The pictures B have the largest compression of the data, but their interpolation is done by means of the past and future picture. These pictures are not used as the reference pictures.

Picture 2: Principe of the midst of the picture interpolation

MPEG -2
This standard is designed with the regard to use it in the distance and satellite transmissions with reserve of television quality. MPEG-2 standard allows the distinction up to 16383×16383 points, only limitation is, that the high and the latitude of the picture has to be divisible 16 for the better dividing to the areas for the compression.

The standard is using in the area with the highest demands for the quality of the visual signal – the digital TV. The system works under non-interlaced line spacing and by the picture form HDTV:

• Space distinctive aptitude 1920×1250 pictures points,
• Picture frequency 50 Hz,
• Bit rate without the compression 1,9 Gb/s,
• Bit rate with the compression 19 Mb/s.

Because in the television signal there is very high possibility of the cutting, there is used only the pictures of types (I) and (P). The prediction, unlike from MPEG-1, is not executed after half-pictures, but after whole pictures. There is again usage of the estimation of move.

MPEG-2 system allows the change of the picture quality, which is done by change of:

• Time distinctive ability – this is change of the picture frequency, which has the effect of change the quality of the picture and thereby the change of the transmission speed,
• Frequency distinctive ability – by the change of the latitude zone, there is possibility to regulate the details in the picture,
• Space distinctive ability – there is change of the screen of the picture.

Such interference to the quality of the picture is allowed by two-ply structure of the system according :

• Basis ply – this allows compress all tree distinction abilities.
• The ply with higher distinction

From above mentioned there is the result, that MPEG-2 has two bit flows. This solution allows work in the qualitatively different modes.

MPEG-3
This standard was originally aimed as the support for HDTV, so the television with the high distinction, indeed this area was after the corrections capable cover by the standard MPEG-2. The MPEG-3 is not used anymore.

MPEG-4
Unlike from the growing demands for data flow in the previous standards, the standard MPEG-4 is defined for the transmission of the video and the connected audio-signal in slow lines with the speed from 4800 to 64000 bits per second, therefore mainly past modems. This speed is very slow and MPEG-4 is therefore based on the distinction 176×144 points for 10 shots per second.

Advantages:
- The possibility of the deprivation compression represents the big advantage of the digitalized video and audio against the common analog technologies. By the compression sequence of the pictures is possible use so called the moving vectors, which determine how the picture changed against the other picture of the sequence. There is no need to code each picture as the static – this is advantage, which allows reach better compression proportions. In CD audio is possible to reach storage of the 12 to 13 audio CD on the one dada CD with the practically no change of the quality.

Disadvantages:
- The mere video compression is very complicated process for the memory and also for the arithmetic performance of the computer. As the decoding (replaying) is many times similar, also here the arithmetic performance of the computer is limited just clearly for the software processing of data in the form of MPEG. To reach professional outcome in this area, there is necessary to use additional hardware equipment (the special cards for playing the video).

RESULT
The dynamic picture is forming by the ground of the pictures, which vary in the space and time. The view is specific form of the signal, which is very correlative and therefore is characterized by substantial redundancy. The redundancy is part of the signal, which is redundant. In the process of the coding this redundancy is cut off and in the reverse process – the decoding, the redundancy is then restored. By the cutting off the view redundancy there is getting the compression of the view.

In the case of the dynamic views, there are two types of the redundancy – the space and the time redundancy. The space redundancy is the result of the high correlation of the picture points inside of the view. The time redundancy is based on the correlation between two pictures of the dynamic view, on the assumption that, these two pictures are from each other very similar.

The aim of this contribution was to present the basic attributes of (compression) standards of the digital video - MPEG-1, MPEG-2 a MPEG-4, which are in the present time used in the different technical applications. The technical description of mentioned standards concentrate on the digital view data, even described standards commonly specify also the standards of the voice signals (commonly multimedia).

ACKNOWLEDGMENTS
This work was supported by the Ministry of Education of the Czech Republic under grant MSM 7088352102.

LITERATURE

GAMCOVÁ, M. - MARCHEVSKÝ, S. - GAMEC, J.: Higher efficienty of motion estimation methods. In: Radioengineering. vol. 13, no. 4 (2004), p. 38-43.
JAIN, A.K.: Fundamentals of Digital Image Processing. PRENTICE HALL OF INDIA, NEW DELPHI, 1995.
KLÍMA, M., BERNAS, M., HOZMAN, J., DVOŘÁK, P.: Zpracování obrazové informace. Praha, vydavatelství ČVUT, 1996.