Reference Transports: An overall perspective

"Also, even though with computer based audio certain "error free" software can be used, this does not exactly mean that there are no transport device READ errors."

Alex, if one uses (and correctly sets up) Exact Audio Copy (EAC) along with a good CD-ROM drive that supports both C1 and C2 error correction (a Plextor Premium or similar) then you *will* get bit perfect copies, as long as there is nothing inherently wrong with the disk. If there is, EAC will give you an indication each and every time an error is detected. Any scenario which would cause a read error that was not detected by this set up (I'm not really sure that's possible) would neither be detected by a dedicated transport.

Having said all that, I agree with you completely that a bit perfect audio image is only part of the picture. But the tradeoffs between a server based bit perfect image and a fine transport are smaller than the D/A conversion of either of these signals, IMO. I've just recently moved to a server based world (a Squeezebox into my D/A converter), and guess what? That does not sound quite as good as my transport (a Proceed PDT3) into the same converter. Is it comparable to some of the other transports I've tried over the years? Yes. Is it scary close to what I have now? I'm afraid to say, yes, with the deficiency being (I believe) in the Squeezebox, and not in the server based delivery of a bit perfect signal.

"every CD player, even the cheapest one, has C1 and C2 CIRC error correctors"

Alex, I don't know about audio CD players in this regard, but the context of my comments was WRT computer CD drives. And in that environment this is not always the case. If you look here

http://www.daefeatures.co.uk/search.php

(and just hit the Search button to get all results) you will find that there are quite a few drives that do not have C2 correction. You will also note that there are some drives (though few) that do not have Accurate Stream, perhaps even more importatnt than C2 for clean audio extraction.

"So, if there is no error flag, this does not meant that there is no error; it means that C1 and C2 were able to correct the errors."

If an error is corrected, would you then agree that the bit of information now in the system is identical to what was on the disk (minus any jitter component)?

I don't think either Bombaywalla, Lktanx, or I are saying that having an accurate data source is all that is needed for audio to sound good. In my case I am saying that a server based data stream (when done correctly) may in fact be a viable replacement for a transport - from the perspective of assuring an accurate data source as your starting point. Once the data is assured to be accurate then you may be able at least to remove that component from the equation.

I ran a small experiment before posting this reply. I have three different CD drives in my (computer) system. One (the Plex Premium) is used for audio ripping and software program installation. I pulled a few CDs which were ripped from this drive and did a binary compare against what is on my music hard drive, each CD being tested in all three of my CD drives. What would you propose to be the likelihood that one of these undetected errors would be undetected in exactly the same way on each drive, and that an undetected error was matched exactly against all three drives with what had been ripped to my hard drive? All three CD drives are from different manufacturers, BTW. It did not surprise me in the least that the ripped audio files exactly matched what was on each corresponding CD, when read from each of the three drives. My sample is certainly not exhaustive, but I suspect fairly representational.

Again, I don't know about audio CD players, but in the world of computer data if there was any substantive liklihood of that sort of undetected error from a CD read occurring, then how often would software installed from a CD fail to operate becuase what was on the hard drive did not match what was on the CD? Note that I'm not talking about corrected errors. In all the years I've spent installing software from CDs (assuming the CD was not damaged in some way) I have never had that happen. Not once. Ever.

I guess all I'm saying is that there is a way to generate a data source that (if done correcetly) will allow the service of that data to be as accurate as what is on the audio CD, in a way that removes jitter from the serving component. At this point the REST of the chain will determine how that data will sound. And in that respect I agree with you completely.