Quick idea about the dithering matter, without suggesting to shed a lot of light sending myself in such subjects: making sure the bit depth is properly used is understandable, even though it may well be the difference between a straight AD-converted signal of 16 bits, coming from a natural source or mix/production, and a dithered signal brought down to 16 bits is not all too much, and also the risk exists, especially with large and uncomely dither noise, that the resulting noise adds up to a negative improvement.

Audio isn't the same as visual, even though of course it's a nice picture in the video, the equivalent in audio would be to dither a signal of full (0dBS) amplitude, and surely that can't be the objective!

Also clear from the sine wave with blocky-roundings (isn't math wonderful ?) example is that there may be confusion about vertical dithering may have time -rounding effects, which would be a wrong suggestion, and isn't mentioned as such. But it is true that the example dither signal makes clear there's a need to bandlimit also the dither signal, and to hope/make sure there's no correlation with the real" signals of similar nature, or you might get phasing dither signal interference.

Thinking about the result of dither on certain signal properties, like the usually pre-equalized mid range on CDs: it may be important to dither at the right point in your production path, or you will emphasize certain predictable signal filter properties, like the impulse response of the equalizer used. Futhermore, there are signal correlation computations (like with a (averaged FFT)) which may or may not be influenced by dither, and also which may give pretty low level (<-50dBS) pulsing sub-bands very important for the perceptual quality of audio that could get messed up with wrongly tuned dither, and in pro-audio studio production, they, as well as (natural or well done artificial) reverb give natural dither patterns!

T.

--
dupswapdrop -- the music-dsp mailing list and website:
subscription info, FAQ, source code archive, list archive, book reviews, dsp 
links
http://music.columbia.edu/cmc/music-dsp
http://music.columbia.edu/mailman/listinfo/music-dsp

Reply via email to