Goes back a long way this thread and I am not sure anyone is still reading it, but here goes.
An interesting little experiment a friend did recently was to take some 192/24 (FLAC) files and load them into Amadeus Pro, he then chopped them up into 10 second chunks and recoded alternate ones 44/16 (AIFF) then stitched the whole lot together and recoded it all to 192/24 (FLAC) and sent the files back to a few of us without clues as to wheat had been done (we agreed not to scrutinise the files before playing them and the sequence wasn't always the same, some started down and up converted and some the other way around). Three things of interest here - Firstly, before we knew what was going on nobody could ascertain the pattern of 10 second chunks but everyone heard significant differences - it as as if we all somehow psycoacoustically integrated these files to make sense of them. The differences between those files and the original were a lot more to do with image width, ambience, instrumental separation etc than other factors though the sensation was of a definitely degraded performance that, had we not heard the original would be heard as not terribly good recordings rather than bad ones.
Once we knew the process it was easier to detect the changes but even then our ears preferred to mask it somehow.
For me what stood out was that the sections that weren't transcoded also sounded different to the original - that is that loading the files into Amadeus made subtle changes. It wasn't dramatic but the two of us with higher end (12k and 30k uk pounds) both heard this independently and unambiguously) the tentative conclusion being that unless files are directly transferred a processing application can alter something within the file to, ultimately change it acoustically.
My own accidental switching of my Naim Unitiserve to 96/24 and subsequently back to 192/24 was clearly noticeable on appropriate tracks and I have since repeated this with friends listening and me, out of their sight, switching the DACs. WAV v FLAC, for the me jury is out as it's quite hard to find two ostensibly identical recordings in both formats so any, what have to be subtle differences, are masked by the recordings simply being different. I have however downloaded some free tracks from Linn in both formats and think I can detect a small difference, but it is extremely subtle and I'd doubt myself rather than stating there was a difference.
An interesting aside to this is that the final size of the 'modified' files showed varying amounts of data compression (due to the FLAC recoding). The conclusion we came to was that whilst the supplied files were 192/24 the original recordings were probably only 44/16!
Anyway, I ramble and I suspect interest in the post has waned, but any comments about this little experiment will be read with interest.
