Unavoidable dither?

Greetings, people of audacity!
I’ve made a post some weeks ago asking how to avoid cumulative dither when working with an already dithered 16-bit input. From what I understood, it seemed that I could simply import a 16-bit file, process it, export it as 32-bit float to then export that output file to 16-bit in order not to add dither twice. However, though this trick works for exporting non-processed mixes, audacity still automatically adds dither when exporting those processed 32-bit float files to 16-bit. Does this occur because every samples can’t be perfectly represented as 16-bit values? If no, can I safely set dither to none when exporting or will I get harsh harmonics effects even though the source material was previously dithered ?

I am kind of lost here…any help would be much appreciated :melting_face:
In any case, thank you all for your time! May everyone have a magnificent day ! :slight_smile:

Yes. “None” should be OK, and you don’t “lose” the existing dither." At 16-bits, you can’t hear dither, the effects of dither, or the lack of dither, under any normal-reasonable listening conditions. You’re not going to hear “harsh harmonics” So, don’t loose sleep over it!

If you work with 8-bit audio can try it to see which you prefer. At 8-bits you can hear quantization noise and dither noise is supposed to sound better.

Making an 8-bit file is a good way to hear quantization noise. It’s similar to regular analog noise and it’s a like a “fuzz”. Like analog noise it’s most noticeable with quite sounds but unlike analog noise it goes-away completely with digital silence.

Tricky question. If you make a slight volume or EQ change, or something like that, the existing dither should be OK. But with a fade-out, where the audio (and dither) get truncated to silence, theoretically additional dither may be better.

Thanks for your response DVDdoug!

Don’t worry, I am aware that cumulative dither is in no way a real threat to perceivable audio quality. As a matter of fact, since I could spot absolutely no discernible audio differences between processed input and final exported output, it took me years before actually learning that audacity automatically adds dither! Honestly, why I’m being so conscious about it is just to be absolutely certain I’m maintaining as much audio “quality” as I can for my mixes, even if in the end such “degradations” would be of no importance.

Anyway, Thank you so much for your time! I shall wish you a marvellous day :smile: