On Wed, Apr 22, 2015 at 12:46:59PM -0700, Larry Colen wrote: > > What would be really nifty would be the option of getting all three or four > low resolution raw files in addition to the superresolution final file, on > the off chance the merge doesn't work, you'd still have unmerged files that > would be usable.
I don't see a lot of point to that. The logic that combines the four original images into a a single multi-channel image is pretty straigntforward; with the exception of the green component, which is the sum (or average) of values from two of the original images, each component value in the merged image is just a copy of the value in one of the four input images (possibly offset by one pixel horizontally and/or vertically). I consider it extremely unlikely that errors would occur in doing this combining (and suspect there may, in fact, be custom hardware such as a one-scanline shift register to support doing this rapidly). Where I think there is far more possibility for error is in the accuracy of the sensor shift mechanism in shifting by exactly one pixel. If such an error does occur the final image may not have quite the full spatial resolution that would be possible, but it's still going to be better than a Bayer-transformed result derived from just a single one of the four input images. And, in any case, it's pretty simple to pick apart the combined image to get the four original images (except for some slight uncertainty in the green channel). -- PDML Pentax-Discuss Mail List [email protected] http://pdml.net/mailman/listinfo/pdml_pdml.net to UNSUBSCRIBE from the PDML, please visit the link directly above and follow the directions.
