question about dithering

2016-03-23 09:02:25

Hi. I was experimenting with dithering and there is something I'm wondering so someone who understands please answer. I was converting 96 kHz 24 bit file to 44.1 kHz 16 bit file without dithering option in foobar and when I canceled them out in audacity to see the difference I could see / hear the difference as white noise but only if I amplify the result a lot. Does that mean foobar still uses dithering although I didn't select it?

Re: question about dithering

Reply #1 – 2016-03-23 09:18:31

My guess is that foobar uses shaped dithering when the option is selected and normal white noise when it's not.

Re: question about dithering

Reply #2 – 2016-03-23 10:10:15

When dither isn't enabled there is no dithering used. If foobar2000 forced dithering you would never get bit-perfect output from it.

Re: question about dithering

Reply #3 – 2016-03-23 10:29:14

Thank you for your quick replay. So why is the difference white noise? I was expecting pure quantization error noise only.

Re: question about dithering

Reply #4 – 2016-03-23 15:36:52

Why should it be something else?

Re: question about dithering

Reply #5 – 2016-03-23 21:15:32

I don't know. That's why I'm testing this, to find out. I thought it would be some more specific noise. BTW why can't I cancel original and converted file when I use dBpoweramp/SSRC resampler like with other one?

Re: question about dithering

Reply #6 – 2016-03-23 21:30:28

Apparently 8 least significant bits of the signal contain only some random noise, so all you can see is just white noise.

About dBpoweramp/SSRC resampler: IIRC it adds some very small delay, so you cannot perfectly align original and converted signals.

Re: question about dithering

Reply #7 – 2016-03-25 13:02:50

Thank you very much!
One more question: Is there any option in foobar for other types of dither so I can test them too?