> further info, consult Ken Pohlmann's "Principles of Digital Audio" I went to school down at UM where Ken Pohlman teaches, although I was not an engineering major, I had the chance to take one of his classes. He explains error correction in a very simply and easy to understand way. Check out that book. When I get some free time, I am going to get it and read it myself. ----- Original Message ----- From: Marvin Humphrey <marvin@...> To: <logic-users@onelist.com> Sent: Saturday, April 01, 2000 5:11 AM Subject: Re: [LUG] Re: mastering > > From: superflo@... > >> And this is the one of the MOST important factors, because EVERY CD you > >> burn will have errors, thats why the Reed Solomon correction process is > >> inside every CD player ! Also Dat has errors ! Thats why they still prefer > >> well done analog tape to begin with. > > > > And what causes the errors you mentioned on every (Audio?)-CD that > > anybody burns at home? > > > > Flo > > It's 1:30 am, and time to go home, but I'll take a brief stab at this > nonetheless. ALL digital storage media designs utilize error detection and > correction. CDR, ADAT, DAT, Exabyte, Hard drives, floppy discs, you name > it, they all use it. It is used primarily because it allows you to greatly > increase the density of the data that you can store. Essentially, you're > writing the data redundantly. Vast sums of R&D money have been spent trying > to determine what the most advantageous error correction algorithms are. > > Error correction kicks in quite often, but you don't hear anything, because > it's possible to reconstruct the sample from the redundant information. > Playback of a pristine CDDA will result in thousands of errors, but all or > almost all will be completely recovered, resulting in perfect or > imperceptibly-close-to-perfect playback. If there's a particular sample > where an "uncorrectable" error occurs, most systems perform "interpolation," > which means: If I know the values for sample A and sample C, but I don't > know the value for B, I can make an educated guess that B is right in > between A and C. Individual interpolations are almost always inaudible. If > a string of several samples in a row come up as "uncorrectable errors," a > phenomenon knows as a "burst error," that's often audible. In order to > minimize the potential for burst errors, information is scattered into > various locations on digital storage media, rather than written in a > continuous stream. Playback of a CDDA with lots of light scratches will > result in tens or even hundreds of individual interpolated samples over the > course of the entire disc, but once again, that's unlikely to be audible. > Playback of a CDDA with heavy scratches can result in skipping, muting, > burst errors ("fzt") or dropouts, all of which even I can hear :-) For > further info, consult Ken Pohlmann's "Principles of Digital Audio" or maybe > try an Altavista search for the phrase "what is parity". (parity is the most > basic error detection algorithm, and is a good place to start). > > As for why analog tape is still groovy, it's simply because a properly > aligned and calibrated analog machine can sound absolutely fabulous. > > Marvin Humphrey .................................... marvin@... > > "...if you're doing an experiment, you should report everything that you > think might make it invalid - not only what you think is right about it: > other causes that could possibly explain your results; and things you > thought of that you've eliminated by some other experiment, and how they > worked - to make sure the other fellow can tell they have been eliminated." > - Richard Feynman > > > > >