cable tech facts issue 108

In This Issue

• Different Types Of Digital Video Signal Formats
• MPEG-2 Profiles
• Industry News

Back To Cable Tech Facts main index.

In earlier issues of Cable Tech Facts, we discussed how the analog video signal is converted to a digital video signal. In this issue we will discuss the next steps in the digital video process - compressing and transporting.

Different Types Of Digital Video Signal Formats

The cable operator may be dealing with several different types of digital video signal formats. There could be as many as four different formats of digital video being used in a system. Each format has its advantages and disadvantages. Here is a quick review of each of the digital formats:

The cable operator will need to be able to accommodate each program source in the headend. They will also need to be able to test the signals and ensure that a quality signal is received in the headend and then re-transmitted or transported to the customer.

The problem that digital video creates is that in its uncompressed nature, it has a very wide bandwidth, almost a full 6 MHz. This would actually reduce the number of channels or frequencies that could be used in a digital system. Here is an example of what we mean.

In using an 8 or 10 bit coding system to define the signal values would leave you with a huge bitstream to try and transport to its end destination. For example, a 525-line, 8 bit, 4:2:2 component digital signal contains 720 samples per line, times some 486 active lines per frame, times 16 bits per sample (8 luminance and 8 chrominance), times approximately 30 frames per second. This would add up to 168 megabits per second. To transmit or transport this information would require a huge amount of bandwidth, extra bandwidth we don't have. We need a method of reducing this information in order for it to be useable. This is where MPEG and other video compression schemes come into place.

There are several differences between MPEG and other compression techniques. MPEG was designed and tuned for video. It uses a single color space (Y, Cr, Cb) and has built-in accommodations to handle audio signals. Secondly, MPEG almost guarantees that the bit rate you select is the bit rate your will get back out of the system. This makes MPEG very predictable. Third, and most importantly, MPEG takes advantage of two characteristics of video: the high degree of commonality between pictures and the predictability of the picture's movement.

It is important to note the differences between MPEG-1, MPEG-2, and other compression formats. Each method varies simply by the amount of detail they sample and store. MPEG-2, which is the method that seems to be coming to the top of the heap, basically delivers the same video quality as a Super VHS video recorder, which is better than the present analog quality, but isn't the top of the line. It is a compromise between high end video and the ability to provide more channels in the available bandwidth.

There are no U.S. standards established as of yet. This can cause problems when you try and hook up equipment from different manufacturers, they may not be compatible with each other. Just because one manufacturer implies they have MPEG-2 compatible equipment does not mean it will work with another manufacturer's actual MPEG-2 equipment. You must take great care in selecting your equipment or you could be chasing problems that seem to have no answer.

The Europeans have set up standards around DVB (Digital Video Broadcast), and many of the U.S. manufacturers are adopting to follow those guidelines.

Pre-Processing

The first step of compressing a digital video signal or algorithm is the pre-processing function. During this time, the intent is to remove the information that is the most difficult to code and is least important to the overall image quality. This is the key to the compression concept. The amount of information contained by a given element or event is inversely proportional to the likelihood of its occurrence: the less probable an event, the more information value it contains. An example would be a crowd surrounding a racetrack. When the camera pans the cars passing by, the details of the crowd, such as facial expressions, are removed because excessive motion is difficult to compress and is not seen by the human eye. The digital encoder begins by dividing the digital signal into 16x16 pixel macro-blocks. These macro-blocks are then divided down even further into 8x8 blocks. These blocks represent very small samples of the overall picture, but retain their orientation.

Motion Compensation

Because there is often some similarity between each frame of video, it is more efficient to actually code only the changes that occur from one frame to another. This is known as "interframe" coding. For example, if we select a black pixel in the middle of a picture of a black object, it is likely that the pixels surrounding it are black, having the similar luminance and chrominance values. Any such pixel has a low information content. If one of these pixels is white, the probability of its occurrence is low, and its informational value is high. Sending information only about the changes reduces the amount of data that has to be transmitted. In the MPEG process, bi-directional prediction is used to help forecast motion.

This step buffers several frames so that the preceding and succeeding images can be examined and compared to determine how much motion takes place and where the moving objects should be placed.

[Top]

Frequency Domain Decomposition

Depending on which MPEG method you are using, the format calls for redundancy between adjacent pixels to be eliminated and to further reduce the amount of information that has to be coded. (An MPEG-2 format eliminates more redundancy than an MPEG-4 format, hence lesser detail in the picture and less bandwidth required to transport the information). By using a process known as discrete cosine transform (DCT), pixels are decorrelated from neighboring pixels so that only the minimum amount of information is transmitted. Other "intercoding" methods include predictive coding and vector quantization. There are five profiles and four levels that make up the MPEG-2 organization. They are:

Quantizing And Coding

After quantization, which assigned discrete values to the output of the frequency domain decomposition process, the quantized information is further coded, typically using a method called entropy encoding. The most common method of entropy coding is Huffman Coding. In this method, code words are assigned to shorter codes for frequently occurring symbols, and longer codes for less frequently occurring symbols. The result is a compact digital representation of the entire picture.

Once the compression sequence is completed, the bits are organized into a transmission format. Forward error correction (FEC) is also added to allow the receiver or set-top converter to determine where or if a bit error has been made in transmission. The digital signal may also be pre-distorted to compensate for any known problems in the transmission path. The signal is then digitally modulated and upconverted to a carrier frequency. From this point it can be either sent out over microwave, as in an MDS, or wireless cable system, or sent out over the coaxial cable system. It is important to remember that any cabling problem can and will affect the signal, but not to the point where you will see any signal degradation. Digital signals are much different than analog. You may see some sparkles in the picture, but typically when a digital signal goes bad, the signal may freeze or disappear. This is called the digital cliff effect.

Once the signal is received at the customer home, the set-top converter or digital receiver does the exact opposite to the digital signal and processes the different layers of the digital signal. The set-top receiver is capable of concealing transmission problems due to the fact that it provides distortion compensation through an adaptive equalizer circuit. The signal quality may be much better at the output of the digital set-top than at the input to the box.

[Top]

Industry News

Not Across The Border

The FCC made a formal announcement that TCI would not be able to move into the high-power DBS market through a Canadian orbital slot. TCI has two satellites ready for launch at this time and will continue to try and expand its satellite presence.

Digital Set-Top Interoperability

An agreement was made over digital set-top interoperability between General Instrument and Scientific-Atlanta. They have agreed to work toward an arrangement that would include a royalty-free cross-licensing for core encryption, modulation, and forward error correction.

Digital System Guidelines Set

CableLabs and its members have agreed on the major elements of an interoperable digital cable system specification for North America. Under the terms of the specification, digital gear will conform to MPEG-2 main profile/main level parameters, the MPEG-2 transport multiplex, Dolby AC-3 audio, and ATSC service information tables. The downstream modulation will be based on ITU's Annex B, which calls for 64- and 256-QAM with concatenated trellis coded modulation, plus enhancements including variable interleaving depth to reduce latency for delay-sensitive applications.

[Top]

[Sencore Home Page]