TremoJem
I might be doing it wrong, but I only dither with Ozone 5.

That's what I plan to do from now on.

I dither in Ozone as well, but I have used PowR3 plenty of times.  I'm not going to pretend I can hear a difference, and it is probably true that you won't either.  My advice:  pick one and don't waste more time thinking about it.  If you want your recordings to be better, get better computer fans or room treatment or something.  Don't think about dithering.

Craig, Very interesting.  I agree on dithered v non-dithered.  I am going to try to dig up the data I saw on the sonic differences among dithering types and make sure I am not confused.  I know I personally cannot hear the differences in the types but I am not in a pristine laboratory.

Another good resource to check out is the educational guides on the iZotope website. Although some portions apply specifically to their products, they are written at a more universal level for the most part. There is a 38-page pdf on dithering in the "Ozone 6" section. (The top 3 listed pdf's are all worth a look).

Rectangular and triangular are computationally cheap, which isn't important in the modern world.
 
Pretty much all the other ones made by different companies are dither+noise shaping, where the noise shaping is essentially an EQ curve that moves much of the noise and quantization error to frequencies where your ears are less sensitive and/or it will likely be masked by your signal.
 
Despite all the fancy technical names designed to impress you, they're basically the same but just have different EQ curves to shape the noise.
 
The short answer is to use one with noise shaping once when dithering down to 16bit where it might make a difference in borderline cases. At higher bit depths it doesn't make any difference in the real world.

Anyone who can hear a difference between the different dithering schemes must clearly be an android.. This could be a good test for the state department to use..

Try the test I outlined above.
 
To me, the difference was clear and repeatable.

Dither adds noise. The difference between dither algorithms is in the frequency distribution of that noise. The simplest methods (square, triangle) pretty much don't care how the noise is distributed, so they can skip the extra steps and run more efficiently. Proprietary algorithms (Pwr-X, MBIT+, UV-22) reduce noise frequencies that human ears are most sensitive to and move that energy up into the stratosphere where nobody cares.
 
In theory, that makes the more sophisticated algorithms more transparent. In practice, nobody can hear the difference. Sure, you can artificially create a scenario in which dither noise can be heard, but not in any real-world applications, where it's irrelevant.
 
For that matter, in most pop/rock recordings you could skip dither altogether and nobody would notice. I wouldn't do that in a harpsichord piece, especially with a golden-eared harpsichordist checking my work. But for a typical mix, even a dynamic one, few listeners would be able to distinguish between dithered vs. non-dithered version. Don't take my word for it - try it yourself.

If you raise the volume during quiet parts, it's not a valid test. You're really testing what a lower bit depth sounds like.
 
At best, with 16 bit audio dither type is only going to make a difference in unusual cases - some combination of: average signal level far below 0dBfs, loud listening level, quiet listening environment, type of dither used and the signal itself.

thanks everyone ....