MPEG-2 and measurements

By Konstantin Glasman

Review contents:
MPEG-2 and measurements TV Network Monitoring Equipment of the the IMOS Company Going digital with program insertion — challenges and solutions Innovative T&M system for MPEG-2 monitoring and analysis Systems from Snell & Wilcox for testing, estimation and checking of MPEG streams Video test and measurement equipment

Space of the TV measurements

Amongst the whole MPEG family the system of compression MPEG-2 has the greatest area of application. In accordance with algorithms and procedures, specified by MPEG-2 standard, in all modern system of digital TV broadcasting the compression, or compression of data stream, is an obligatory technological component. A MPEG-2 decoder shall be in each digital TV receiver. The development of the TV is indissolubly connected with improvement of the technology and methods of TV measurements; though an introduction of data stream compression have required the development of the new methods of the data stream evaluation. However a coordinate system and landmarks in the space of TV measurements remained unchangeable and are the same, which they were designed by domestic school of TV Measurements.

The theory and practice of the TV measurements are aimed to achievement of three main goals:

• Metrological ensuring of scientific investigations, developments and system designing;
• Metrological ensuring the process of TV equipment production;
• Metrological ensuring of the production process and transmission of the TV program.

In process of the scientific investigations are required to carry out the different measurements. Moreover, the main requirement to measuring device is versatility and high accuracy. In production of the TV equipment the measurements are basically used for evaluation of the equipment parameter correspondence to the approved norms. At production and transmission of TV program the measurements are carried out for checking and adjustment of the equipment parameters, for diagnostics and prediction of failures.

There are the following completed areas of the TV measurements:

• Evaluation and measurement of the qualitative features of the TV pictures;
• Measurement of the parameters of the video signal;
• Measurement of the features of TV equipment and devices;
• Checking and measurements in the control system of TV broadcasting.

The requirements to measurements and measuring instruments are diverse and sometimes contradictory. So in each area of the TV measurements are developed instruments, in base of the functioning of which put the different methods:

• Direct measurement;
• Indirect measurement;
• Measurement in process of the TV program transmission;
• Measurement outside of TV programs;
• Measurement in real time;
• Measurement in not real time;
• Continuing measurement and evaluation;
• Discrete (selective) measurement and evaluation.

In spite of enormous achievements in sphere of the instrument development for TV measurements, in the field of evaluation of video quality the most absolute measuring instrument is a human visual system. Besides, though broadcasting systems creates the pictures for spectators, so the quality evaluation by human is decisive. Accordingly the measurements in the evaluation area of the qualitative factors of the TV pictures can be divided into two groups: objective and subjective.

The objective measurements are executed by means of special instruments. The purpose of the direct measurements is a direct evaluation of picture quality. The indirect measurements are executed with supply of special test signals.

The subjective measurements expect the evaluation of the picture quality by the observers which are spectators. These measurements of the quality are always direct since opinion of the spectators about quality of the test signal or test table with use of some scales of the subjective values should not have any meaning.

With introduction of MPEG-2 compression in two first areas of the measurements (the evaluation of picture quality and the measuring of the video signal parameters) the essential transformations have been required. Whereas the compression has brought to the appearance of the additional distortion and disturbances, the necessity of development of the new quality evaluation methods have been required. In reality the streams of compressed data are the multilayered hierarchical structures, which parameters must correspond to the approved standards. So it was appeared the necessity to develop a line of analyzers of MPEG-2 data stream.

Non-linearity of TV systems with compression

In analog broadcasting the widest expansion has got the objective indirect methods of the picture quality measurement. On base of the long-term investigations the families of the test signals were created, which have provided the high correlation degree of the results of the objective indirect measurements and direct subjective evaluation. The measurement of the test signal distortion in TV path enables to foresee the picture distortion in process of the picture transmission of the TV program in the same path, not resorting to the labour-consuming subjective test. The objective indirect methods allow to get the reliable results in digital TV system without compression also.

That fact that test signals, included in the video signal composition in the interval of the reverse motion and TV pictures are the same distortions in TV path, allows to consider the systems of the analog TV and systems of the digital TV without compression as linear systems. In this context a linearity is considered in the sense that the distortions appearing in TV path do not essentially depend from picture content that is to say from its space-time structure. The digital TV systems, in which compression is used, in this sense are nonlinear systems. The distortions, appearing in cycle of MPEG-2 compression-decompression, greatly depend from structured features of the TV picture. The usual static test signals are easy yielded to the MPEG-2 compression. Even under higher degree of compression the distortion of such test signal is low and do not give any presentation about distortion of the real TV program pictures.

Distortions in systems with compression are varied much more than in systems without compression. It is necessary to add that many disturbances, appearing on picture as a result of compression, seem as alien and artificial due to their often name is an artifact. For instance, one of the typical demonstrations of the MPEG-2 compression distortion is a block structure, at picture appearing as absolutely unnatural. Compression of the static pictures is accompanied by smaller distortion, than compression of the dynamic TV sequences. The transformations of the input picture of coder compression, which lead to reduction of correlation relationship of nearby pixels of TV frame or of the same name pixels of the nearby frames (for instance, sharpening of the picture by means of aperture correction or the increasing of the fluctuation noise level), are lead to distortion level growing and varied artifacts. The reason is in MPEG-2 system, used in broadcasting, which are functioned in mode of the constant data rate in the coder compression output.

As is well known, the compression principle is based on consequent execution of discrete, on quantization of cosine transformation factors and entropic coding of quantized element sequences. Moreover, the discrete cosine transformation is computed within the framework of picture element block with 8 × 8 pixel size. For static pictures, in which brightness and chrominance are changed smoothly that is for pictures with high degree of inter-pixel correlation, the number of nonzero factors of cosine transformation, which the only thing subjecting to transmission is small. If picture appeared as fine-structured and dynamic, when correlative relationship between element are decreased, the number of nonzero factors in block is increased. The way, by which the equalization of the data rates in these two events is provided, is the use of rougher quantization of discrete cosine transformation factors (the detachment of smaller bit amount on one factor) for fine-structured and dynamic pictures. The rougher quantization leads to growing of quantization noise and, accordingly, to high distortion and artifact level.

Thereby, the picture quality in decoder output of compression system is became variable, its value is a function of the picture content (more exactly — is a function of spatial and temporary features of the TV picture). This circumstance has forced to abandon from using of simple test signals in system with compression and go to the wide use of the subjective expertise for the reason of quality evaluation. The purpose of the multiple deep investigations is the development of the new direct objective methods of the picture quality measurement, under which the picture distortions in real medias (or distortions of the test pictures peculiar to media), are valued.

Subjective measurements

During many years of analog TV the strict formal subjective expert operations for the reason of the TV picture quality evaluation were conducted by relatively stable line of the methods, specified by Recommendation ITU-R BT.500. During some time to the group of specially selected watchers with normal vision, the TV pictures with different values of the objective factors have been demonstrated. The pictures were watched under established screen brightness and background; the spectators were placed on determined distance from TV monitor. The task of the watchers was to evaluate the picture quality with use of a certain scale and to an evaluation the detailed instruction was preceded.

The subjective measurements by the broadest image were used and they are used during the study and design of the new systems, but they are of great importance and in practical TV broadcasting. This circumstance is made conditional upon that subjective evaluation gives the direct integral evaluation of the qualitative features of the picture (strict formal when use the standard methods and quick visual evaluation under usual observation of the TV picture). The visual evaluation of picture quality was broadly used in analog TV and, certainly, will not less be broadly used in digital broadcasting.

Introduction of compression in TV broadcasting has forced to turn the greater attention on subjective measurements else since the objective indirect measurements turned out to be suitable only to evaluation of analog elements and linear digital links of the digital TV systems. As it was noted above, the quality does not remain constant in broadcasting system with compression. This quality depends on space-time structure of TV picture that is to say from its saturation by small details and dynamics. The quality inconstancy at time is done more observable in system with statistical multiplexing. The dependence of quality from the picture features and its time histories, the appearance of many new type of distortions and disturbances have caused the need of the contributing of some changes to methods of the subjective measurements, specified by Recommendation ITU-R BT.500. For instance, some revisions and changes were required to the rules of the choice of the test sequences and their length, to the condition of the comparison of the testable picture with reference picture, to the methods of the processing of experience results. The methods of the continuous visual evaluation of video sequences have been intensively developed.

The methods of the subjective measurements were broadly used at study and development of MPEG-2 compression system. Carefully examined experiments on subjective evaluation have allowed proving that at data rates, equal to 6 Mb/sec, quality of the decoded picture is at a level of studio quality in composite (NTSC/PAL/SECAM) systems of standard resolution. On its qualitative features the decoded picture in MPEG-2 system at 9 Mb/sec is equivalent to component (Y, R-Y, B-Y) picture, observed on a monitor in studio. The results of these experiments became one of the factors, which are predetermined the acceptance of MPEG-2 compression for digital TV broadcasting.

The modern methods of the subjective measurements are provided by some undoubted advantages. They are available for receiving of quality evaluation and static and dynamic pictures. They can be used for comparative evaluation of qualitative features of the picture, created by different digital and analog systems. Result of the evaluation is a scalar value in the form of averaged viewer’s evaluation, giving the integral evaluation of system efficiency in respect of ability to create the pictures with high technical quality. The additional factor, intensifying a role of the subjective evaluation of quality, is that situation that by subjective measurements only we can get the original data for the vision model and development of the methods of the objective quality evaluation, which results have a very good correspondence to the visual evaluation. This factor promotes to the study activity in the area of subjective measurement of TV picture quality.

However, the experiments on subjective quality evaluation are labor and time consuming. The experience results are the high value of scattering and they are subject to influence of the large number of difficult controlled factors. The subjective measurements can be not used for the monitoring. This brings to the need of the development of the objective measurement methods, suitable for evaluation of picture quality, created by digital TV system with compression. This is a great necessity in objective measurements for measurements under TV program transmission.

Objective measurements

The traditional test signals and static test tables are easy subjected to MPEG-2 compression without observable distortion and do not allow to get the quality evaluation of the compression coder functioning in real condition. As test pictures for objective measurements in system with compression the using of natural scenes and media have begun. The comparison of original material with the sequences, passed a codec of compression, allows giving the evaluation of a level of compression distortion. The methods, in which the comparison of the pictures at the input and output of codec of compression is used, the name “two-sided” are accepted. Other possible approach to quality evaluation does not require any presence of the original and is connected with finding of compression artifacts on decoded pictures and evaluation of their value. The methods developed within the framework of this approach, are named “one-sided”.

Keeping in mind that objective methods do not allow measuring the absolute value of the picture quality directly means that they only allow valuing the differences in level of picture quality on output of decoder of compression and a certain reference picture or video sequence. In two-way method the reference picture is present in evidence, but it is meant in one-sided methods (as reference picture can be a picture without artifact).

The results of the objective measurements must be well agreed with results of the subjective measurements for the same system and for the same test sequences or media. This requirement provides the main complexity of the development of the objective measurement methods.

Two-sided methods of the objective measurements

The comparison of the testable decoded picture with reference picture, which is often used an original that is to say the input coder picture, is mostly executed by one of two methods: comparison of the pictures and comparison of picture features.

At picture comparison the difference of reference and testing pictures is usually calculated. Than a difference is less, the compression distortion is less, than a difference is higher the distortion is more. Processing of received different arrays is directed on reception of a certain difference, which must be enough well correlated with difference of compared pictures quality. As measures of the distortion it is possible to use, for example, PSNR, which is relationship of the peak-to-peak amplitude to mean-square value of differences between original and decompressed picture. For calculation of mean-square value of differences the averaging is executed within the framework of a certain temporary interval, for instance, within the framework of frame or groups of the pictures.

From analog TV is well known that mean-square value of the deflection is not always corresponds to the subjective perception of the difference. In analog TV for reception of the best correspondence to visual evaluation the weighted functions are entered, taking into account the particularities of the visual perception of the disturbances and noises and realized by means of weighting filters and circuits. By such way this problem is resolved in system with compression. For achievement of high accuracy of the correspondence to subjective perception the signal of difference between original and decoded picture can be subjected to processing by the vision models. These models take into account the differ visibility of compression distortions and artifacts, which have the different frequency spectrum, the miscellaneous brightness and chrominance. In models it can be taken into account the dependency of artifacts from background, on which they are located since it is known that distortion, disturbances and artifacts of compression (for example, the block structure) are better noticed on the picture areas with smoothly changing brightness and on clear long contours and are less noticed on picture areas, saturated by fine-structured pieces. In this model can be used also the processing for the reason of transition to scale of the subjective values counting, when instead of PSNR the number of threshold, or hardly observable increment of the disturbances is calculated under turning from original to decoded picture (this means the transition to scale of the count of the units). The models can provide the calibration of the calculated measure of difference (for instance, PSNR), for the reason of transition to the scale of quality category (fine, well and etc.).

The investigations have shown that PSNR factor is well worked in system with narrow choice of coding parameters, for instance, when the compressed data rate is only changed because of the measurement of the quantization factor. The use of the vision models at calculation of the measure of difference of decoded picture and original provide the improvement of coincidence with results of the subjective measurements; however, all methods are provided a lower accuracy, when the question is in comparison of the systems with miscellaneous compression algorithms, as well as in the case of concatenation of codecs that is to say at quality evaluation of the picture, passed through several compression codecs.

In method of the comparison of features the spatial and temporary features of the original and decoded picture are calculated. Then the calculations will be compared; the result defines the measure of the worsening quality of decoded picture in comparison with original. This method requires the transmission of relatively small data volume in point of the measurement that increases the supply sphere of the method. As calculating feature of the picture so named “criticalness” can be used, characterizing the picture “compressibility” in coder of compression. For instance, for achievement of the constant rate of compressed data of the picture, saturated by small components, it is required the rougher quantization, than of pictures without small components. This means that fine-structured pictures are more difficult compressed and more critical to compression distortion. Than the more the criticalness, the more the compression distortions under the same data rate in compressed stream.

The two-sided methods of the evaluation of picture quality are not at whole available for monitoring, when an original is inaccessible. However these methods are successfully supplied in study, in laboratory, within the framework of closed TV system. It is possible to dispose of necessity to transfer the original in point of the measurement if compare the decoded picture with true copy of coded video sequence instead of original (this method can be used only at measurements outside of TV programs, when the known test sequence is coded).

One-sided methods of the objective measurements

The one-sided methods practically were always used in analog TV for measurement of signal/noise ratio, i. e. the ratio of peak-to-peak amplitude to the mean-square value of fluctuating noise. In digital system with compression many specific distortions, disturbances and artifacts are became apparent instead of fluctuating noise. These artifacts can be discovered in picture by means of digital processing and pattern recognition. The simpler algorithm of the processing is provided by the method, in which use the circumstance that general reason of any distortion is a quantization factor of discrete cosine transformation within the framework of small block of the picture with 8 × 8 pixel size. The direct effect of this quantization is a forming of the block structure on a picture. The block structure can be easy discovered by means of objective measurements of the decoded picture, for instance, by calculation of inter-pixel difference on the block edges. Sensitivity of the modern methods of the finding of the block structure is high; its intensity can be measured even in the case of almost whole obscurity of the block structure to watcher; thus the level of the block structure is a not bad integral feature, characterizing visibility of varied distortions and artifacts of compression.

One more known method of the one-sided objective measurements is based on processing of compressed data stream and does not require any decoding. From the pulsed-code modulation theory is known that level of quantization noise is defined by value of the quantization threshold. The noises, appearing at quantization of discrete cosine transformation factors, can be only determined in the case if the matrix of the quantization and factor of the quantization are known. Since the data are required for ensuring of the decoder functioning these data are transferred to the decoder by compressed data stream. They can be extracted from data stream without decoding so the noise quantization in system with compression can be evaluated on base of the analysis of the coded data stream; moreover for calculation neither original, nor decoded picture is not required.

The algorithms of the processing which are applied in described one-sided method, are based on the knowledge about compression principles and about distortions and artifacts appearing during compression, that does these one-sided methods simpler, however they can be used for measurements only within the framework of concrete system of compression, i. e. MPEG-2.

Test tables

In analog TV the measurement methods based on use of the test tables are broadly used. They can be used as for determination of separate features of pictures and TV path, so and for integral evaluation of reproducible TV picture quality. The measurements by means of tables are based on visual evaluation of the table element distortion and do not require the use of any special complex equipment. These measurements provide quick and it is enough proper evaluation of TV equipment features. Moreover this evaluation is well corresponds to the results of the subjective measurement when use the real pictures.

The traditional test tables can be effectively used and in digital TV without video compression. With introduction of video compression the efficiency of the use the traditional test tables is lowered, so for objective measurements as test pictures the natural scenes and media are used.

The stroke structure over the real video

However the possibility of the test tables is not yet exhausted. Recently, the elements of the test tables were offered, which can be used for evaluation of the distortions in system with compression. As the analogue of fine-grain textures, which can be present in TV picture, is a test structure in the manner of vertical black and white strips, or streaks, located on gray background and having a constant width. These structures can be superimposed on the real materials. The step of the streak structure is enough low for high criticalness to distortion of compression (two pixels). The variable feature of this structure is the modulation depth, defining the contrast of streaks. The modulation depth grows in horizontal direction by linear law beginning from zero value. The quantization of DCT factor brings to distortion and destruction of the test structure. When compression value is increased and, accordingly, the increase of the quantization step in the first place brings to destruction of streaks of low contrast. The contrast of the structure is continuously changed so a length of the destructed part of test structure, i. e. of the part, where streaking structure from black and white lines is not reproduced, is an indicator of compression distortions and artifacts. The length of untouched part of the structure is an indicator of the picture quality. The task of the watcher is in evaluation of the length of the destructed part of test structure (or untouched part) with use of a certain quality scale, located along test structure. Such streaking test structure has a high minuteness. This structure is more difficult object for compression, than a majority of the typical pictures. So the observable distortions of this structure will appeared under lower compression, than for real pictures. The signal of the generator of test table, including test structures for evaluation of the compression distortions and artifacts, is transferred on input of MPEG-2 coder. The observation of the table on the decoder output allows evaluating the picture quality in respect of compression distortion and artifacts. Such measurements can be executed outside of time of the TV program transmission only.

Pseudo random test structure distortions

The use of a test element for TV program continuous monitoring

Other possible decision is an use of the pseudorandom test structures. In process of video compression the pseudorandom structure is distorted. The collation of the features of the pseudorandom structure on the coder input and on decoder output of video compression can be used for prediction of the subjective picture evaluation. Such method is a variety of approaches to measurement by comparison of features of an original and coded picture. However since features of test structure, placed in picture on the coder input, are known, that there is no need to transmit any information about them in measurement point. The advantage of such approach is as follows: the subjective evaluation predicted with use of the small pseudorandom structure, pertains to the current fragment of the TV program directly, rather then to some typical or average statistical scene. Readings of the noise field are uncorrelated in time and space that gives the high load on coder of video compression since the possibility of the prediction is disabled. This means that distortion of the noise structure appear even in the case, when on usual TV broadcasting pictures the compression distortion and artifacts are imperceptible due to low degree of compression. The minimum side size of pseudorandom test structure of the square form is 48 pixels that allows to include this test element in picture of the media, for instance, as a fragment of the logotype and use this method for continuing monitoring during the TV program transmission.

There are test tables in the form of compressed sequences, which are included instead of compression coder. They can be used as for broadcasting, so and for development of the new equipment for systems with MPEG-2 compression. The compressed data stream, corresponding to sequence of compressed frames of the test table and is formed by generator of the transport stream, is absolutely deterministic signal, which can be used for evaluation of functioning of the decoding process and discovery of its weak points. The test streams in the form of compressed data has a very big importance at period of the digital TV implementation since such data streams, formed in full accordance with video compression standard, will play the role of the stable master signal for decoder and digital TV receiver developers. Reproduction of this master data stream by designing decoder, TV receiver or a attachment will mean that his functioning is submitted to the MPEG-2 video compression standard.

Test pattern

The technical problems, appearing in digital systems, can be imperceptible for watcher, but small change of the coding features, characteristics of the transmission channel of picture properties can bring to a failure and full abnormality of the picture playback. To value the particularities of decoder, which can bring to the potential problems under changing, for instance, of the transmission channel features, the test sequences of the special test tables can create greater difficulties for decoder, than typical TV programs (such way of the check is named as system stress). For example, the single-program transport stream in combination with test of synchronizing of video and audio signal under significant available volume of the buffer can be considered as low-level stress. The decoder will reproduce the picture and its sound accompaniment without malfunctions. However the transmission of the same test sequence in multiprogram transport stream with sound accompaniment and with use of statistical multiplexing in condition, when all programs of the stream fill or clear the buffer memory simultaneously, can be considered as high level stress. For all-round checking of compression decoders it is necessary to provide the forming of the streams with low, and with high level of the stress.

Analysis of compressed data streams

The MPEG-2 compression as algorithm of the content processing is the consecutively executed operations of discrete cosine transformation, quantization of the cosine transformation factors and the entropic coding. To this procedure in the MPEG-2 system are subjected the small blocks of the picture with 8 × 8 pixel size. The code words got as a result of processing of the 64 pixels array by three operation process, are grouped in package of block data. The data stream of video sequence is packed in the form of the complex multilayered hierarchical structure, in which the block packages are found on lower level. On following level they are combined in macro blocks. The macro block group is transferred in composition of the trailing edge. Several trailing edges form the picture of one frame and several frames form the picture groups. The top level of the compressed video data stream is the video sequence from the following successively picture groups. This sequence is elementary stream of the compressed video data. The transferring video data on each level are complemented by coding feature information and by other information and are combined in packages with uniquely specified structure, described in “Video” part of MPEG-2 specification. Also, the MPEG-2 standard defines the forming of elementary stream of audio data.

From elementary packages of video, audio and additional data the packages are cut, to them headlines are added with information, required for ensuring of the correct functioning of MPEG-2 decoder that transforms the elementary streams in packet elementary streams. By means of multiplexer the packet elementary streams of video, audio and data are combined in the united stream of compressed data. This united stream can be transferred in the form of program stream of one TV program or transport stream of one or several TV programs. The forming of program and transport streams are specified by “System” part of MPEG-2 standard.

The simple enumeration of levels of program and transport MPEG-2 stream structure has occupied a lot of place. Each level should be provided by instruments and methods for analysis of video, audio and additional data, for installing of coding parameters, as well as that exceedingly it is important, for checking of the correspondence of the data stream structure to requirements of MPEG-2 standard family. In the same way, either as under quality evaluation and measurement, the analysis of MPEG-2 data stream can be executed in real and not real-time mode; in continuous and in discrete mode; during transmission of the TV program and outside of TV program. Real time testing is usually executed in process of the monitoring for the reason of continuous analysis of the data stream key features. Temporary signal features, carrying data stream (jitter), are valued in real time. After some time the data stream can be written and subjected to more careful and more detailed analysis. Continuous measurements and analysis in real time are preferable for monitoring. If computing power of instruments does not allow analyzing a data stream at whole in real time with established level of minuteness, so selective measurements can be used, under which the analysis of some data subsets with determined periodicity is executed.

The analysis of the correspondence of transferring data streams to standard requirements can be executed during TV program on the base of any media (because there are no the restrictions, which are during quality measurement). Here the restriction is only an available computing power of the analyzers. However the analysis of the stream structure outside of the TV program transmission is also of great value. Some special test materials, strictly corresponding to all standards and protocols, can be used in this case. As was it already noted, a test material, used outside of TV program, can also contain the determined deflections from standard requirements, that is important for testing of the digital system with compression for disturbance and error resistance.

The transport MPEG-2 stream is used for delivery of the TV programs in every system of digital TV broadcasting. Within the framework of the DVB standard family the leading document TR 101 290 is released, containing the general principles of the measurements in a system of digital broadcasting. The considerable part of this document is the principles of the analysis of the MPEG-2 stream transport. The most important features of a stream are divided into three groups in accordance with their importance for the monitoring task decision. In the first group the parameters are included, which analysis allows to make sure that transport stream can be decoded by means of standard decoder. The indicators of these features indicate to synchronization loss, to errors in synch byte of the transport stream package, to errors in table of the program combining and in tables of the program schedule. This group of features has the first, i. e. the top priority. If indicators of the first group are indicated to errors, the monitoring of the rest features have not any sense.

In group of the second priority level the features are included, recommended for continuous or periodic monitoring of the transport MPEG-2 stream. The list of these features is as follows: errors in headlines of transport stream packages, errors in control sums of the tables with program information, violations of the rules of the transmission of the marks of reference time and other like violations. The third group includes the features, which are pertained to a broadcasting system (DVB in this case).

Set of features is recommended in the document TR 101 290 for monitoring of TV program in realtime. At outside of real time measurements the analysis is provided to great body of information, transferring in transport MPEG-2 stream. The minute analysis of recorded fragments of transferring data stream can be executed under installation and exploitation of the systems with video compression. For this aim the syntax of the stream is analyzed, allowing make sure that headlines and all package and tables data are formatted in accordance with requirements of the standard. The semantic analysis allows to check, are package data found in allowed limits. The check of data coordination allows to reveal the events, when data from different fields are contradicted each other. The analysis can be realized on levels of transport, program and elementary stream, on levels of the picture groups, frames, cuts, macro blocks and blocks.

About approach to measurements in broadcasting control system:
— How to select the sweetest oranges?
— If you want to buy the sweet oranges, you should select such of them, which weight is more, than their size.
(Adopt from TV program — a guide in world of goods and services)

The MPEG-2 standard family includes some thousands of features, some hundreds of the tables. Compositing and delivery of TV program in modern digital world should be accompanied by fulfillment of requirements of dozens other standards as well. And how to avoid the dead-end position of the buyer from TV program (see above), choosing the measuring instruments, capable to provide the reliability and intolerance of the digital TV broadcasting system?

The key to the development of the necessary measuring features and measuring devices for ensuring of digital TV broadcasting gives the object-oriented model of the digital TV system, designed by working group “EBU/SMPTE Task Force for Harmonized Standards for the Exchange of Media as Bitstreams”. The most important difference of SMPTE/EBU model from previous approach is that the data exchange between system objects is realized by four-level model using, which is built on the base of the model of the ISO/OSI open system mutual joining. These levels are the following: level of application, network level, canal level, physical level. The evaluation of the data stream correspondence in the digital broadcasting system to requirements of the different specifications and standards is not exhausted by analysis of the MPEG-2 data structure. MPEG-2 compression is only a level of the system application.

The multiform essence of TV broadcasting brings to the need to supply the different mechanisms, or protocols, of data exchange on each level for different application. On network level the protocols IP, ATM, Fibre Channel, and AES-3 can be used, on canal level — the protocols SDI, Ethernet, RS-422. The physical level, which defines electric and mechanical communication features, allows, for instance, the use of coaxial cable with BNC connector, twisted pair with RTJ-45 or DB-9 interface. The most important circumstance is in that all planes, layers and levels of model are built on base plane of control and monitoring, providing the control, the reliability and intolerance of the digital TV system. The realization of function of base plane of control and monitoring of the digital TV system requires some measuring devices for each level of each system element. Such approach allows to define the minimum set of the measuring devices required for forming of te metrological base for digital TV broadcasting. The temporary norms for transmission path of digital broadcasting signals, for sections of the path and measuring signals, approved at Russian Federation, are included in managing document “Temporary norms for transmission path of digital broadcasting signals, for sections of the path and measuring signals”, confirmed by order of RF Ministry of communications and informatization from September 25, 2002.

The network organization of TV broadcasting, which is typical for modern stage of the TV progress, requires the remote control of features of system of the distribution and transmission of the TV programs. The provision of measuring instrument by modems and corresponding software allows getting the complex decision, suitable both for small TV companies, and for TV networks of the national scale. The operator of the central control room for the TV program control can to call in Internet the measurement information from the register of any local TV network or peripheral TV company. In central control room the generalized measurement information on technical condition of the whole network and transmission lines is formed, that allows to take the operative decisions for realization of high-end TV broadcasting.