When I compose music, I work entirely with MIDI/virtual instruments. When I've finished writing my parts, I bounce the MIDI tracks to individual audio tracks and proceed to mix. I just recently noticed for the first time that there is a dithering option on the bounce to tracks dialogue that is defaulted to 'on'. Since I never noticed this before, I've obviously been applying it and haven't noticed it causing any harm. But theoretically, is there any reason for me to apply dithering at this stage? Shouldn't I just turn off dithering until I export the final mastered audio track? 

Any advice on this would be really helpful? 

No need for dithering while mixing

Only the final track for CD (44.1/16)
 
 

So, have I been degrading the quality of my music (even if barely perceptively) by dithering while bouncing to tracks?

stxx
No need for dithering while mixing

 
When you render a synth track it will follow the bitrate setting under options>file>audio data>render bit depth. I am not sure whether the internal processing of Sonar is 32bit, or whether it follows the overall bitrate of the project setting. I would think that if you are rendering to anything less than whatever that internal processing bitrate is you probably should have dithering checked.
 
I render synth tracks @ 24bit and have dithering checked.
 
Hopefully someone who knows more will chime in:)

skitch_84
So, have I been degrading the quality of my music (even if barely perceptively) by dithering while bouncing to tracks?

Yes

KingsMix

skitch_84
So, have I been degrading the quality of my music (even if barely perceptively) by dithering while bouncing to tracks?

Yes

I wouldn't lose any sleep over this one Chris. I would wager that if you were to conduct a blind study there probably is not a sound engineer on earth, who when confronted with a mix-down of two versions of your project, one in which the individual source tracks had dithering applied at bounce, and one where they did not, could pick the difference beyond sheer chance. Conversely, I am sure there are plenty who think they could
 
It would still be great to hear from someone across the details of this though to confirm one way or the other. I am yet to be convinced that you shouldn't apply dither when you bounce a track to 24 or 16 bit, for example. Surely there is a reason dither is there, and it defaults to on? The answers so far have been a little light on evidence:)

skitch_84
So, have I been degrading the quality of my music (even if barely perceptively) by dithering while bouncing to tracks?

Yes and no. Yes, it is technically degrading quality, but imperceptively. You'll never hear the difference and it'll be irrelevant when you ultimately truncate your data to 16 bits or encode it to an MP3.
 
The purpose of dither is to hide imperfections inherent in wordlength reduction, by making quantization errors more random. But when you bounce a track, the result is (usually) a 32-bit floating-point file, which preserves all the resolution the DAW is capable of, and that's a level of fine detail that goes way beyond anything you can hear. Dither would just add noise unnecessarily, even if it's noise you can't hear.

SONAR is smart enough not to apply dither except when the bit depth is being reduced.
 
Here's a post from Noel on the topic from some time back:
 
     http://forum.cakewalk.com/FindPost/1204822

bitflipper

skitch_84
So, have I been degrading the quality of my music (even if barely perceptively) by dithering while bouncing to tracks?

Yes and no. Yes, it is technically degrading quality, but imperceptively. You'll never hear the difference and it'll be irrelevant when you ultimately truncate your data to 16 bits or encode it to an MP3.
 
The purpose of dither is to hide imperfections inherent in wordlength reduction, by making quantization errors more random. But when you bounce a track, the result is (usually) a 32-bit floating-point file, which preserves all the resolution the DAW is capable of, and that's a level of fine detail that goes way beyond anything you can hear. Dither would just add noise unnecessarily, even if it's noise you can't hear.

EDIT: ANSWER FOUND IN LINKED POST THANKS:)
 
Hey Bitflipper. Thanks for the info. I get what you are saying with regard to 32bit renders. When the render bit depth in preferences is set to 24 and you bounce or freeze a track, the resulting file is 24bit. In this case, if Sonar is processing 'globally' at 32bit,  wouldn't dithering be required, or would it not matter?
Sorry to quiz you, but I bounce a lot tracks in projects at different bitrates and am I keen to understand the ins and outs!