Digital Video for AV Integrators
When you transform AV into digital bits and bytes, it opens up a world of possibilities?and a few challenges. Here is everything you need to know about moving digital video around an AV installation.
IT'S ALL IN THE NUMBERS
There are two MPEG standards in wide use today for video compression. The first is MPEG-2, which is the basis for encoding DVDs and cable, satellite, and digital terrestrial TV broadcasts. MPEG-2 has been around for almost 20 years, and has done a good job, but there are practical limits to how much a video signal can be compressed using the MPEG-2 system.
To give you some idea of just how much MPEG compression is typically used, a high-definition television program in the 1920x1080 interlaced HDTV format (1080i) has an uncompressed bit rate of 995 megabits per second (Mbps). That would be impossible to fit into a standard TV channel, let alone on an optical disc. Therefore MPEG-2 compression, which can be done in several different flavors, packs that 1080i HDTV show down to about 18 Mbps so it can be transmitted to your TV in a 6-MHz-wide TV channel. That’s a compression ratio of about 60:1, and yet the picture quality you see at that bit rate is very good.
The same thing applies to a DVD, which plays at the level of standard definition TV (SDTV). Video encoded in the 720x480 interlaced SDTV format (480i) is packed down from its uncompressed bit rate (about 270 Mbps) to around 4 Mbps. Yet again, picture quality is as good or better than anything you’d see on an analog TV set. That’s because MPEG compression is considered a lossless codec. In theory, the original signal can be faithfully reproduced during the decompression stage. In reality, there’s always some loss in the encoding and decoding process, and you may observe odd-looking rectangular picture artifacts known as macroblock artifacts.
While MPEG-2 has served us well, there’s room for improvement. A newer version of the MPEG standard (MPEG-4) has come into wide usage for private video networks, consumer video devices, and Internet video streaming.
The version of MPEG-4 commonly used for video compression is known as H.264, or Advanced Video Codec (AVC), which is an extension of the original MPEG-4 standard. MPEG-4 uses the same I-, P-, and B-frames that MPEG-2 uses, but it adds the ability to compress streams down to the subpixel (SP) level and can look forward and backward at an unlimited number of frames for predictive purposes.
In theory, those features should allow for even greater compression of a video signal. And they do, as proven in tests by the European Broadcast Union (EBU) a few years back. In the EBU tests, HDTV signals could be compressed an additional 50 percent over the acceptable bit rate for MPEG-2, which made it possible for European broadcasters to move ahead with a single HDTV broadcast standard: 1920x1080 pixels with progressive scan, at a picture refresh rate of 50 Hz.
The increased efficiency of MPEG-4 hasn’t escaped content distributors in the U.S. Satellite TV companies are in the midst of a switchover to MPEG-4 from MPEG-2, and cable TV companies will follow soon. The 3D TV channels coming to cable TV will launch with MPEG-2 and move to MPEG-4 by 2011.
Video streaming services such as Netflix and Vudu make exclusive use of MPEG-4 encoding. So does YouTube and just about every camcorder sold these days. Blu-ray discs, which were initially encoded using MPEG-2 compression, have all moved to MPEG-4 coding to pack more data into a program and improve picture quality by delivering movies in a 1920x1080-pixel, progressive-scan format.
BETTER LATE THAN NEVER
You’re probably wondering how it’s possible for codecs to go back and forth in real time to do all of these predictive calculations. The answer is latency, or a delay built in to the video encoder (and your video decoder) that allows for all of these mathematical calculations to take place.
If you were to watch an analog TV program in real time next to one that is being encoded and compressed, you’d see up to a two-second delay between the analog video playback and the digital version. That time interval is required for the codec to do its thing. Some broadcast compression systems that use lots of what is called forward error correction (FEC) to make up for “dropped” bits can add several seconds of latency.
Ordinarily, latency is not an issue when distributing digital video programs, unless there are audio latency issues (lip-sync problems). If the same digital video stream is feeding multiple TVs or monitors, then there will be no time offsets between any of the sets unless their internal video decoders are having problems.
That said, there are wavelet-based, proprietary compression schemes that aim to eliminate virtually all possible latency, particularly where the network is tightly controlled and supporting mission-critical, real-time video applications, such as command and control. The Pure3 codec, developed by Electrosonic and recently sold to Extron Electronics, is one such proprietary format.