Analog Processing vs. digital in mastering.

Do not return to the analog domain once converted to digital:

Earlier in this article, I warned against going back to the analog domain once you've gone digital. Ideally, only one of these conversions is needed, once on the original recording and once on the CD player playback.

…

Loss of transparency in AD/DA conversion:

If a mastering engineer has a favorite analog EQ, or processor they want to use to create a particular sound, they need to carefully weigh the pros and cons. There is always a loss of transparency when passing through analog stages, particularly in A/D and D/A conversion. Anyone who has patched these processors into their mixing consoles is aware of these trade-offs.

In other words, you must carefully weigh what results from adding an analog stage and additional converters with the subjective improvement in processing versus processing the source in the digital domain. There will be some unavoidable slight (or severe) wrapping or loss of transparency due to each conversion. However, perhaps the mastering engineer believes that the music will benefit from the sonic characteristics of a classic compressor or equalizer...perhaps he is looking for a "pumped" quality that cannot be obtained with any of today's digital processors. nowadays (many people complain that digital processing is too "clean"...certainly a topic for another essay). There are many vintage "sounds" and other effects still obtainable with analog processors. And finally, some mastering engineers claim that analog processors sound better than digital processors. I am not one of them. I'm not going to make that blanket statement.

…

Notable digital tools:

Our choice to use analog or digital processing depends on the nature of the source, the music, and the tools available to us. I now have some digital tools that are so remarkable when used correctly that customers with excellent ears can't believe the processing wasn't analog!

What is Jitter?

Jitter is an error in the time base. This is due to the variation of time delays in circuit paths from component to component in the signal path. The two most common causes of jitter are poorly designed phase locked loops (PLLs) and waveform distortion due to mismatched impedances and/or reflections in the signal path.

Shown here is waveform distortion that can cause a timebase error:

A D to A converter can impair the audible dynamic range

A 16-bit recording can have more than 120 dB of dynamic range; a D to A converter with a clock jitter can impair the audible dynamic range to 100 dB or less, depending on the severity of the jitter.

Instruments that lose their sharp outline

I have performed listening experiments on audiophile-quality purist music material recorded with a precise 20-bit A/D converter (interpolated to 16 bits within the A/D). The sonic results of passing this signal through processors that clip the signal at 110, 105, or 96 dB are: increased "grain" in the picture, instruments that lose their sharp outline and focus; reduced sound width; the apparent level loss makes the listener want to turn up the monitor level, even though high-level signals are reproduced at unity gain.

These effects can be heard without having to turn up the listening volume

Contrary to popular belief, you can hear these effects without having to turn up the listening volume any higher than normal (proving that low-level ambience signals are very important to playback quality). Similar degradation has been observed when jitter is present. However, the loss due to jitter is subtle, and mostly audible with higher-grade audiophile D/A converters.