If you’re going to dither more than once then “coloured” (UV22 or any other shaped dither) should not be used, acording to that article. Coloured dither should only ever be applied once and that would be the very last process done.
So basically if I were to ever mixdown to 44.1/16 for burning to CD I should be using it but “don’t bother otherwise”…?
When the sources are already in 44.1khz then I wouldn’t use dithering, only when I have a diffrent samplerate and bitdepth I use UV22HR as last insert on the masterbus when exporting the mix.
I should test this out since I have the frequency analyzer VST.
Does the test tone plug-in generate a sine wave? If so, I’ll use it to generate an MP3 and do a frequency analysis on the result to see whether we get the same high noise floor as was illustrated in the article (and, hence, would require the use of dithering).