Home Computer Audio Asylum

Music servers and other computer based digital audio technologies.

RE: In my mind there is only one way to compare format (a) to format (b)

Your comments about synchronization are well placed, but depending on the equipment they may not go far enough. Most modern ADCs use a sigma-delta modulator to capture the analog signal at a high sample rate and low bit depth. This modulator has state and works by feeding back random noise in a chaotic fashion. Therefore even if two identical converters are started at the exact same instant off of synchronized clocks the bits output will probably differ, possibly by a large amount at some points due to cascading converter artifacts.

Presumably a differences in low order bits won't matter much. However, it's not so simple. Some audiophiles have heard differences between two 24 bit files that had identical samples, except that there was 1 low order bit of dither noise added to one of the files. In other cases, audiophiles hear differences between two bit identical CD rips, except that one had a few zero samples at the start, well before any audible music began playing. In both of these cases, similar chaotic effects may have happened in the DAC circuitry, especially if it uses sigma-delta converters as most presently do.

Even something as simple as a file to file sample rate conversion from a low resolution format to a high resolution format and back again may not produce the expected results. When I converted from 16 bits to 24 bits and back to 16 bits I expected and got the dither noise. (The 24 bit example involved similar conversions from 24 bits to 32 bits and back to 24 bits.) When I did a sample rate conversion from 44.1/24 to 88.2/32 and back to 44.1/24 I expected to get a close match, perhaps a -100 dB null. That was not to be. There were various factors in play that prevented getting a null deeper than about -60 dB, in some cases one could easily hear the difference between the two files at a normal volume setting. Was this a violation of the Sampling Theorem? No, the cause turned out to be a time offset of a fraction of a sample due to the particular filters used in the sample rate conversion. I complained to the author at the time. Several years later a new version came out that fixed this offset, or so I am told. I have the new version (iZotope RX4 Advanced) but haven't repeated the tests.

The problem (interesting aspect) with audio is that the most minute and seemingly insignificant detail can make audible differences that are large enough as to overcome "the willing suspension of disbelief" that allows audiophiles to occasionally perceive reproduced music as sounding real.

Tony Lauck

"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar


This post is made possible by the generous support of people like you and our sponsors:
  Crux Audio  


Follow Ups Full Thread
Follow Ups
  • RE: In my mind there is only one way to compare format (a) to format (b) - Tony Lauck 02/18/1510:57:27 02/18/15 (0)

FAQ

Post a Message!

Forgot Password?
Moniker (Username):
Password (Optional):
  Remember my Moniker & Password  (What's this?)    Eat Me
E-Mail (Optional):
Subject:
Message:   (Posts are subject to Content Rules)
Optional Link URL:
Optional Link Title:
Optional Image URL:
Upload Image:
E-mail Replies:  Automagically notify you when someone responds.