Dithering

Sonar dithering is great.

The difference is in noise shaping. This means pushing the noise into the less audible spectrum. Make the children listen to it who can hear those frequencies.

When you are applying dither before the final version, triangular makes sense since shaped noise will stack into the audible range when applied over and over. I use powr3 on final mixes.

Try bouncing silence with dither then normalize to -3dB and observe it in a spectral analyzer like in the EQ. That will show you the shape of the noise in the audible spectrum. Also you can listen to it.

In general, dither properly applied might get you an extra dB down to the noise floor... this means your mix sounds 1 dB better ish. For CDs this is out of 96 theoretical dB. For 24 bit it is over 140.

This means that the improvement for dithering a 24 bit file is unlikely to be appreciated by a human. I do it as a best practice.

One last thing... lots of people will probably jump on this thread with strongly held opinions like only dither when reducing bit depth. I have a Sweetwater sound book that says dithering can reduce the noise floor when altering sample rate too. The fact that so many smart people have so many different opinions points to the fact that it isn't easy to notice which is better. This should emphasize how very little it matters.