Helicon Focus has some options for the sample size, the smoothness, and the downsampling. Downsampling seems to just reduce the size of the image--not of interest to a photographer, but useful for an image analyst looking for battlefield targets, for example. The sample size is just how many pixels the software looks at each time. The default is 8. I've gone downwards to 1, and seen no detectable change. On a smaller computer, going down would probably make the process longer; my Mac G5 just eats this type of image processing, so all I can say is maybe. Likewise, the smoothness control, default 4, doesn't seem to do all that much. I'm probably not giving it much of a challenge with a 3008 x 2000 image. Smaller images might benefit more. There are two ways Helicon could accomplish it results. First, it can create opacity masks for each slice, which are weird looking black and white images. Without going into all the math details, these masks identify what is sharp and what is not. Again oversimplifying, a sharp 8 pixel sample is brighter than the same blurred sample. Since it creates the masks, which allow all sorts of photo trickery in "polishing" the image, that's my guess. Or, second, it could just directly compare each sample without creating masks. Pretty much the same math, just with a variation. All the information is there from the camera; it's just number crunching. Anyone else been experimenting with Helicon?
