There's been some talk about the 'new' and improved output resolution engine in V9. When up-rezzing output does V9 work on the RAW data or does it first convert into TIFF/JPEG format and up-rez from there (and ditto for downsize) ?
Hi Manoli,
I'm not privy to the internals of C1, but it is clear to me sofar that the downsampling quality is much better, and that it's probably due to the concepts that Nicolas Robidoux and I tested/discussed in
another thread. Since part of the magic is in resampling in a linear gamma space, or something close to that, it seems logical to assume that rescaling is partly based on the demosaiced Raw-ish data before being gamma pre-compensated for output, but also partly in gamma adjusted space.
So I guess it's hard to label it in simple terms, but the result is much better than it used to be. I have not tested it yet for upsampling and other distortions (e.g. keystone correction and lens distortion corrections), so I'm not sure, yet, what has changed there.
One thing I am missing (a lot) though, is a good
preview of how the sharpening settings will affect the rescaled result, and a more advanced sharpening option for improved sharpening when processing time is less important, but best output sharpness quality is of the essence.
As far as I can see now (athough I could be mistaken), the sharpening settings are applied at 100% and then resampled, while the reverse order would be required. Sharpening as the last step is almost 'always' the best method, but without being able to preview it it becomes a guessing game.
Such an accurate sharpening preview at the final output size is also needed in the print module.
Cheers,
Bart