MPEG-2 coders and decoders

By Alexander Anufriev

Review contents:

MPEG-2 coders and decoders MPEG-1/-2 coders from Canopus Orca coder from Doremi Labs Company Optibase card for MPEG-coding/decoding Professional MPEG-2 coder with IP MGW 2000 output. Main possibilities and applications MPEG Pinnacle Studio MovieBox USB coder Coding devices of the Streampress Encode Series Equipment for MPEG coding and decoding from Scopus Network Technologies company MPEG coders from Sigma Designs Company Professional decoders MPEG-1/MPEG-2 Stradis Coders and decoders from Tandberg Television Company MPEG-2 coders from lower price ranges

The advantages of the digital television in contrast with analog TV are essential and obvious:

• There is a greater number of the programs in one and the same occupied frequency band of HF spectrum (four and more as usual);
• This is a smaller radiated power, which is necessary for provision of the same covering zone (or big to noise resistance and interferences under the same powers);
• This is the best quality of the transmission;
• This is a possibility of the creation of the terrestrial broadcasting network on one frequency (the transmission of COFDM signal under SFN (Single Frequency Network) network);
• This is a possibility of the mobile reception (with use of COFDM signal);
• This is a possibility of the simultaneous transmission of any auxiliary information.

The digital television requires the use of the digital video signals. These signals can be received at device, giving some digital signals in output (the digital television cameras, digital studio equipment), or, as this is done in most cases, by transformations of available analogue signals into digital form.

Non-compressed digital video- and audio signals are forming any big stream. On average, for transmission of one program the data transfer rate of 270 Mbit/s is required. If non-compressed data are to be sent in on-line forms, so they will occupy much greater band of the frequencies, than analogue signal. Thus it is necessary to compress these data for stream shaping of the acceptable size, for stream transmission under communication link and delivery to users. For coding and compression of video & audio signals the International standard MPEG-2 (Motion Picture Expert Group version 2) was chosen to compress a TV-program with stream from 270 Mbit/s to 5…6 Mbit/s under preservation of the same quality. The compression ratios to 4 Mbit/s and less are too possible, but here will already exist the trade-off between quality and the stream rate.

The encoding equipment aim is to convert the different formats of input video signals in united form (a transport stream — TS) — in accordance with that transport environmental, under which content will be sent. In digital TV the digital interfaces ASI DVB and SPI DVB are the most common. To date, however, the MPEG-2 coders with Ethernet output are become more popular.

Coder for compression is considered as the most responsible element in the digital signal processing chain, and to a considerable extent this coder is defined a picture stability and quality under given digital stream rate. The MPEG defines the structure of the stream and master model of decoder, but does not superimpose any restrictions on coder structure or algorithm of its functioning so there is broad field for activity of production companies to achieve the best features under minimum bit “consumption”.

Modern coders have a module design and allow using the removable blocks of input interfaces, allowing the transferring as the analogue (composite and component) so the digital video signals. If required for the following entering in multiplexer from analogue signal the teletext signals are choosing by means of additional decoder. Composite decoding brings to considerable deterioration of picture quality and therefore for use is not recommended.

In coder the important functions are carried out by a preprocessor. It realizes a digital filtering and frame synchronization and an additional temporary processing and noise suppression is carried out as well.

The standard decision for coder is considered the presence of two stereo audio channels. The coders can supply as digital AES/EBU signals, so and analogue signals. In the last case the analogue to digital conversion is realized with sampling of 18 bits/sample and sampling rate 32, 44.1 or 48 kHz. Depending on chosen mode, rate of the output stream in channel can change in the range from 32 to 384 kbit/s, providing MPEG-1 level 1 or 2. Installation of two additional audio coders that enables to organize 4 stereo or 8 mono channels in general is possible.

It is common for compression coders the presence of user data transmission channel: a low-rate asynchronous channel with a streaming up to 115.2 kbit/s and synchronous channel with a streaming up to 20 Mbit/s.

For software renovation a separate and no power dependable memory is existed, which allow keeping the previous loaded software versions. Such system allows configuring the coder with flexibility, to update software and if it is needed, to install the different options, such as statistical multiplexing, encryption, cascading, and the support of the profile 4:2:2 etc.

The most of produced coders MPEG-2 DVB can form the minimum necessary set of PSI-tables and therefore can work in single-channel mode without additional multiplexer, creating a transport stream at the input modulator directly. The coder output signal can be formed in one or several common standards. The DVB-ASI interface is used mostly; DVB-SPI and RS-422 are used more rarely. The maximum stream rate on output coder is defined by chosen profile and compression level. For instance, the stream rate for more often applicable combination MP@ML is 15 Mbit/s. For applications 4:2:2 MP@ML the maximum stream rate is 50 Mbit/s.

According to ITU (International Telecommunication Union) recommendation) the comparative tests were carried out, which were based on subjective estimation by the group of observers, which must to look through digital TV-pictures after coding in 2, 3, 4 and 5 Mbit/s streaming. These tests have shown that the best coding profile is 4:2:0 for all streams. Moreover, some experts confirm that video sequence, which is coded under profile 4:2:0 by 10 Mbit/s, has the same quality, either as when use the profile 4:2:2, but at the stream of 13 Mbit/s! So it is preferable to use the profile 4:2:0 at streaming less than 10 Mbit/s, particularly for video sequences, containing much motion. It is necessary to note also that it is difficult to catch the quality difference for video sequence, coded at these profiles with streaming above 10 Mbit/s, since quality already so high that the perceiving of the difference is very complicated. The small advantage of the presentation 4:2:2 up to 4:2:0 at streaming above 10 Mbit/s is reduced to the minimum in the case of transmission of analogue signals, converted in digital form. Considering that presently the stream with 15-20 Mbit/s for one program is not accepted (and it is expensive), so the choice of the coding profile is always falls on 4:2:0 (MP@ML).

Thereby, it is possible to select the most general settings in type MPEG-2 coder (4:2:0):

• Video resolution: Full D1; 3/4 D1; 2/3 D1; 1/2 D1; SIF; QSIF (it is necessary to choose the most suitable resolution, taken into account the stream rate (bit rate) and required coding quality in accordance with content);
• Displayed picture resolution: 720×576 pixels (the maximum for PAL), and 720×480 pixels (the maximum for NTSC). The more high resolution can provide more high accuracy, but only with account of the stream rate;
• Structure of the group of the pictures (GOP): number and the sequence of the coded I, P, B frames;
• Coding rate up to 15 Mbit/s;
• Rate of the output transport stream must be equal or above, than video- and audio streaming rates, taken together, plus the data tables. The difference between real coding rate and the output transport streaming rate withdraws by adding of zero packages (bit-stuffing);
• Sampling rate of the audio (32 or 44.1 or 48 kHz): the more the sampling rate is, the better quality of the transmission is presented, but in this case it is necessary to increase the streaming rate;
• Filter use: if the coder provided by composite video input it is possibility to choose a comb or rejection filter for dividing of chrominance and brightness signals. The other filter types can be used for noise suppression (for instance, in the case of low streaming rate to avoid the noises during transmission).

To date the overland telecommunication networks to synchronous digital hierarchy SDH have all wider spreading, as certain alternative for satellite distribution channels. It is well known, that the main interfaces for SDH-multiplexers are E1 and E3 streams. Optionally, for many modern coders the G.703 (E1, E2, E3) and DS-3 interfaces can be ordered. The transmission network of the video signals in MPEG-2 format already is in use all over the world successfully.

For streaming broadcasting of content under IP-networks the MPEG-coders with IP-output are used. These coders take video from content provider and code the signals in MPEG packages. Now the coders basically use the MPEG-2 algorithm of the coding that allows broadcasting with rate approximately from 3 to 4 Mbit/s on one stream. This rate provides the digital picture quality which is comparable with traditional analogue system.

The main features of MPEG-2 IP coders are following:

• Transmission of several channels “live” TV on IP-networks. Some models can take up to six analogue signals, then to code them in realtime mode in MPEG-2 format;
• Broadcasting on IP-network in multicast and unicast modes;
• Presence of input video interfaces: analogue composite or SDI that allows working with any signal sources;
• Coding in MPEG-2 formats with maximum resolution of 720×576;
• Presence of output interfaces 10/100BaseT Ethernet Full Duplex or Half Duplex (RJ-45); herewith the network protocols UDP Multicast/Unicast and RTP Multicast/Unicast are used.

The realization of digital compression advantages in considerable degree is depended from device type, which is restoring of original picture and audio: MPEG-2 decoders and receiver-decoders (RD). On functional features and cost this is the possibility to divide these devices into two groups: the professional devices and subscription receiving devices.

Professional decoders and RD are used for decoding of compressed signals in studio complexes preparing the programs (for monitoring, editing, standard converting), for receiving in satellite distributing system with the aim to prepare the programs or for entering in cable TV networks on main stations, for receiving of compressed signals, which are sent under synchronous and asynchronous networks, in corporative TV-system for simultaneous acceptance of content and data, as well as in digital satellite news gathering (DSNG), etc.

The main features, on which it is possible to distinguish the professional receiving device, are the following:

• Possibility of the use the large number of different input and output interfaces, including digital outputs of videodata and transport stream;
• Modularity of design allowing to change the interfaces with function change;
• Developed software with extensive menu;
• Support of the different profiles and MPEG-2 standard levels and video and audio coding formats (4:2:0/4:2:2) (AC-3, AAC);
• Support of the high input information stream rates — up to 80 Mbit/s;
• Work with more information saturated types of modulation — 8-FM and 16-QAM;
• Presence of the built-in meters of acceptance (BER) and Eb/No ratios validity;
• Possibility of remote (through Internet) control or change of software through satellite link;
• Presence of screen for displaying of parameters of the adjustment and operation, as well of receiver elements condition;
• High reliability and mean-time-to-first-failure, a possibility of the whole-day functioning and other functions.

Conditionally in RD-circuit it is possible to select three main elements: input device, decoder, output element. The input element supports the different types an interface for receiving of digital signal in desired transport environmental. The transport stream can enter on decoder either from input element, or directly from transport network through DVB-ASI or RS-422 interfaces. Decoder executes de-multiplexing and decoding of elementary video, audio, data, service data streams. The output element forms the video and audio signals and can have analogue or digital interfaces.

Many models of professional RD allow removing accepted TS through DVB-ASI interface for the following presenting on re-multiplexers for multiplexing or transmissions by FOCs. This interface can be also used for cascading of decoders, thereby allowing create the receiving-decoding systems. Presence of the DVB-ASI interface on output receiver allows to consider their not only as final equipment required for imaging of the received content on TV-screen, but use as signal sources for making software packages on main stations. The receiver-decoders well suit for this aim particularly such as, for instance, CODICO CID-3100 decoder of Scopus Company. The particularity of this receivers class is that they are do not serve for information imaging on a screen, but are intended for “opening” (de-encryption) of the programs, which are closed by different conditional access system (DCAS). For instance, CID-3100 can simultaneously decode up to 4 different conditional access systems. Herewith if with usual professional receiver it is possible to get the DVB-ASI stream only with one chosen program, that CID-3100 can give completely “open” stream. This is particularly important also for use on main stations during new software packages forming, which will then spread in open mode or be closed by other DCAS.

For receiving TV-programs, transferring through IP-network, IP-attachments (set-top box — STB) or so-called Etherbox are used. Its purpose is understandable: to take IP-traffic on Ethernet, to select MPEG-2 stream, to decode and to give the video & audio signal on TV set.

During switch on the IP-attachment executes an initialization along with “skimming” of configuration information from control system. This information includes the electronic conductor and IGMP information, which is showed to what broadcast group the selected channel belongs. When subscriber have changed the channel, STB notifies the network that subscriber anymore no needs for old broadcast stream, but must be joined to new broadcast group. Then STB takes the new MPEG-2 stream, decode it and sends analogue video to TV set. As a rule, modern IP-receivers support the MPEG-1 and MPEG-2 streams. If additional storage devices or flash-memories are required the STB features can be extended, for instance, with Java interface, allowing the installation of the interactive platforms. STB can be supplied by a conditional access system.