By the bye, I think this isn't too hard to understand intuitively (excerpted from the article quoted previously): --------------------------------- http://www.physorg.com/printnews.php?newsid=3D1906 Breakthrough in Coding Theory and Practice [ ... ] i.e., an algorithm that fully utilizes the probabilistic information available at the receiver. The Koetter-Vardy soft-decision decoding algorithm results in substantial coding gains in practice [up to 1.5 decibels on additive white Gaussian noise channels, and much more on Rayleigh-fading channels]. ---------------------------------- DTT receivers of all types to date give poorer performance in echo environments than they do when only random noise is present. This can only be happening because they aren't using the signal energy available to them. It seems obvious that if one has knowledge of the distortion being introduced into a signal, at any signal level, one should be more successful at recovering from the distortion than if one has no knowledge. With random white noise, there is no particular rhyme or reason to the errors. That's why they are called "random." It seems obvious that if there's some sort of pattern to the errors, even with very weak signals, they can be fixed more successfully than signals where errors are random. Ideally, the *true* C/N margin required for successful reception ought to be measurably lower in Rayleigh and Ricean channels than in Gaussian. By "true," of course, I mean that C/N is (C+I)/N, rather than the more optimistic looking C/(N+I) you often see quoted. (I is interference from echo energy.) Bert ---------------------------------------------------------------------- You can UNSUBSCRIBE from the OpenDTV list in two ways: - Using the UNSUBSCRIBE command in your user configuration settings at FreeLists.org - By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word unsubscribe in the subject line.