Just wondering if any of you guys have had a chance to evaluate the result from this plugin yet, and what you think? My quick initial impression after playing with a couple of images was that I get better results with FocusMagic. The Topaz plugin seems more prone to artifacts. Maybe I just need to get to know the controls better in the Topaz plugin, but so far I'm disappointed.
I really wish FocusMagic would get a 64-bit update...
I'm busy finding out how to get the best results with different scenarios. My initial impression is that it takes more effort to achieve good results than with FocusMagic, but good results are possible. The options with FocusMagic are more limited and less fine grained, so it is faster to reach an optimum. However, besides the seemingly dead end for 64-bit, there is not much (user) choice about which type of issue is being addressed (e.g. defocus or diffraction, which require different approaches).
There are more options with InFocus to choose from. For general use, I assume it's based on a Gaussian type of PSF, the "Generic" blur type setting works best, but one has to avoid a too large radius setting. What is too large? Well that's the trick to find out, but for a well focused shot with a good lens values around 0.80 or 1.00 seem to be good choices. Modest "Suppress Artifacts" settings (<25) seem enough to remove small ringing artifacts. InFocus does have a larger tendency to develop such artifacts than FocusMagic, but FM probably does it without user intervention and InFocus allows the user to decide.
The "Out-of-Focus" blur type, which probably uses more of a disc shaped PSF, is intended for exactly that, OOF images, but the already in-focus areas can develop artifacts, so it's probably best used on a masked layer. Again, artifacts usually mean that too large a radius is chosen.
The "Straight Motion" blur type is obviously for motion blur, and again artifacts are easy to generate, perhaps too easy.
The "Unknown/Estimate" type of blur is very interesting, because it's intended for difficult/mixed blur scenarios.
So overall, good results can be achieved, but it is also (too) easy to generate artifacts as a by-product. One has to exercise restraint, and it's always beneficial to use a sharpening layer because that gives a lot of additional control. Another possibility is to first resize the image (Bicubic Smoother, e.g. to 200%), then do the InFocus Sharpening, and then resize back to the original size (Bicubic). This seems less artifact prone, but does require more memory and it's slower because more data will be processed.