Formosawest:
Always wonder why folks reply to a 3 year old post, but since your comments are right in line with some current observations I have made, here goes.
I have a Bryston BDA-1 and used it for a number of years, primarily because I had some issues with the analog output stages of a Cambridge Audio CD player (840c) I was using. I love Class A SS amplification and the opamp/IC output stages on the CA just did not cut it. Female vocals and some horns sounded quite strident; micro-phonics and ambiance sounded overemphasized to me with the CA. I then tried outputting the digital output to the Bryston and things sounded much better. But I would assume it is not always the case; I ran some digital out (SPDIF) from a Sony player as an experiment and the opposite was true.
But to the point of my response, I soon substituted the CA 840c with a "memory" transport PS Audio PWT and fed that via AES/EBU to the Bryston and another level of detail and resolution opened up. Much better than any CD transport/deck I had heard up to this point (but I was not comparing Esoteric level gear). I began hearing details on my 80's version CDs that I only remember on the vinyl. All these years I blamed bad pressings and early D/A cd technology, but in reality that info was there, just not getting it via PCM transports.
One small example. On my Caravan album "New Symphonia" (have both vinyl and 80's cd, recent remaster), there is a section where the drummer drops a drumstick just behind the lead singer in a vocal. Have always heard this clearly in vinyl and never in the cd version (not clearly anyway) until I got the PWT. There it was again. And that was not the only case. Clearly the memory transports are better IMHO than all but the very expensive (ie 2-10x) PCM transports. I believe that is simply because the memory transports have time to "attempt" to make a perfect digital copy of the FILE, not just an analog playback of a bit stream with error correction. But since they are still trying to play a CD in real time, they do not have forever to rip a perfect (bit perfect) replica of the sound source.
So if this is true, what about (bit) perfect ripping of a CD music source with software like dBPoweramp and playback through a system like the Bryston BDP-1/2? Well, I just got a BDP-1 and it is even better than my PWT IMHO. Playing FLAC files ripped from a given CD, and A/B'ing with the same CD in the PWT, with both the PWT and the BDP-1 feeding the BDA-1, direct comparison strongly favors the BDP-1 with the digitally ripped files via dBpoweramp. Makes sense to me because the dbpoweramp files are the only real assurance that the digital representation of the CD is bit perfect.
While the PWT uses a CDROM drive, it has a limited period of time (and buffer space) to "rip" a perfect digital copy, whereas dbpoweramp can make as many passes as it needs to of a particular sector in your computer optical drive, comparing results with an international database of prior ripped cds. It also has the advantage of being on most users PC systems having Gigs of RAM and multi-core GHz CPUs for crunching. Not hardware that even a $4k memory player can afford onboard. Once a digital copy is obtained, the Bryston BDP-1 does an excellent job of turning that file into SPDIF bitstream for the DAC.
So although the infrastructure investment/time in ripping all my CDs so they can be played through the BDP-1 is serious, I have found the sonic results to be worth it, as I am re-discovering my music collection yet again.
I am sure many out there will counter that high end PCM transports will sound just as good. But I ask how much further can you go if you have a bit perfect representation of the music? That is the ultimate goal for which all transports ultimately aspire, no matter how beefy the transport, how true the rotation rate and speed, and how low the jitter. It all must end in a bit perfect representation of the music file.