Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Bradley Danyluk]

It recently occured to me that quantization distortion (the nonlinear distortions caused when the audio signal reaches or falls below the lowest dynamic value that a given bit-depth can represent) sounds almost exactly like styrofoam rubbing styrofoam, or nails on a chalkboard, or a large sick bee buzzing nearby.
No wonder people prefer the sound of dither... a relaxing nearby stream of water versus an impending killer bee attack, or falling shipping crates, or a schoolteacher on a rampage. I'm thinking our response to this sound may be based on some evolutionary advantage of really, really disliking very similar sounds.

[Rob Darling]

that is not what quantization distortion is:

Quantization


Quantization is the process of selecting whole numbers to represent the voltage level of each sample. The A/D converter must select a whole number that is closest to the signal level at the instant it?s sampled. This produces small rounding errors that cause distortion.

[Bradley Danyluk]

I thought it strange too but I have heard it referred to before as this several times.

What would you call this type of distortion then?

[David Satz]

Bradley, I don't think that you and robdarling are really disagreeing--you're just talking about the same thing from different angles. Your first message describes the effect of quantization distortion at very low signal amplitudes in the absence of proper dither.

In the popular press this was sometimes called "digital deafness" back in the 1980s. They needed a word for it because it sometimes actually occurred in released recordings--it wasn't recognized as a problem yet by some record producers, who were unaccustomed to the wide dynamic range.

Most of the first generation of CD tape masters were produced on the Sony PCM-1600 professional digital processor. It didn't have proper dither, though some "accidental dither" was generated by the input noise of its audio circuitry. About two years later the successor model PCM-1610 had proper dither available, but units were delivered from the factory with the switch set to "off," many studios had no idea what it was all about and simply left the switch set that way, and Sony treated the entire topic like something that only a few kooks would ever be interested in.

That may have contributed to the mixed reviews which the earliest CDs got at the time--though I could list other factors that were probably more important, such as the lousy nth-generation equalized/compressed/limited master tapes that the major U.S. record companies all used for their initial CD releases, before they caught on to the fact that CD players were letting the public hear how awful those tapes really sounded!

But it was soon recognized (especially due to the work of Stanley Lipshitz and John Vanderkooy in the AES) that proper dither is an intrinsic requirement for A/D conversion in audio. Since then, hardly any hardware or software has been sold without some usable dither implementation, and some systems have even done well with special implementations of it that reduce the subjectively audible noise floor even further. Of course that doesn't mean that every end user understands it or uses it properly.

--best regards

[dcollins]

David Satz wrote on Fri, 22 July 2005 05:05

About two years later the successor model PCM-1610 had proper dither available, but units were delivered from the factory with the switch set to "off," many studios had no idea what it was all about and simply left the switch set that way, and Sony treated the entire topic like something that only a few kooks would ever be interested in.

Kooks that turned the switch on also remember that the poor low-level linearity of those SAR A/D's made dither at the 16 bit level mostly irrelevant!

DC