Fred Miranda wrote:
Yes, perhaps Adam could add this option automatically for the script.
For example: in a 256 image, you could get resulting images at 32, 64, 128 and 256. (so 4 DNG files). Then we would choose which one is best for the scene and delete the other ones.
Feature has been checked in. I made it flexible by using fractional factors for the image counts instead of absolute values, that way it scales to the number of images per collection of files detected. For example, you can specify '1/4 1/2 1', which on a 128 image stack will translate to 32, 64, and 128 images but on a 256 image stack will translate to 64, 128, and 256. You can also specify as many fractional values as you want, for example '.1 .2 .3 .4 .5 .6 .7 .8 .9 1', so you'll get 10 renditions, each with 10% more images vs the last.
Here's the documentation for the new option.
'partialstacks', '<list of fractions>' - (mean-stacking only) - Creates multiple stack renditions from each set of files, with the number of images per rendition equal to a fractional amount applied to the total image count for that collection of files. For example, on a detected stack of 256 images you can create a stack using the first 64, 128, and then all 256 images by specifying 'partialstacks', '1/4, 1/2, 1'. You can also use decimal notation: 'partialstacks', '.25, .5, 1'. This feature is useful when using stacking for ND filter effects, allowing you to select from multiple renditions to select the rendition that best matches your creative vision. Be sure to include the value of '1' in the list if you want a rendition that includes all images. Note that the number of images for any fractional value will be rounded down. For example, if you specify 'partialstacks', '1/4, 1/2, 1' on a set with 7 images, the 1/4 value will be 1 image (1/4 * 7 = 1.75, which is rounded down to 1.0). Also note that any duplicate image counts resulting from similar fractional values will be discarded. For example, 'partialstacks', '0.25, 0.35' on a set with 8 images will produce one stacked output with 2 images, since 0.25*8 and 0.35*8 both round down to a count of 2 images.