On Tue, Apr 26, 2016 at 11:53 AM, johannes hanika <hana...@gmail.com> wrote:
> heya, > > On Wed, Apr 27, 2016 at 6:04 AM, J. Liles <malnour...@gmail.com> wrote: > > Group, > > > > I've recently acquired a camera with an X-Trans sensor, and have had > little > > trouble generating images that contain areas prone to demosaicing > artifacts. > > hehe, i know what you mean. > > > Is anyone interested in these? If so, is there a certain place I should > > host/send them? I have camera generated JPG for comparison along with the > > RAF files. > > nice. very interested. if you could put them up somewhere where we > could access them for a few days at least that would be great. i know > for instance the rawtherapee guys would probably be very interested in > this, too. > OK, I'll put some of my nastiest images on my website and post a link. I also have some with problematic colors too (neon lights!), which I find missing from many studio tests. > > > In playing around I've discovered a few things: > > > > 1) Color smoothing is mandatory, as fine lines and especially specular > > highlights will always generate nasty color artifacts when using the > baisc > > demosaicing. Seems like a smarter highlight recovery algorithm could help > > identify many of these problem spots. > > i find removing the highest two frequency bands of colour information > using the equalizer is often times a good idea. > > > 2) None of the 'denoise' modules have sufficient parameter ranges to deal > > well with the X-Trans files. > > what do you mean? i think the profiled denoising works very well on > fuji cameras. the original paper was written with test data from fuji > cameras and i think the noise model matches them better than most > other cameras. > Profiled denoise simply obliterates detail at the default settings and when you turn down the strength it leaves behind some ugly artifacts... Everything will be smooth except for sparse contrasty speckles. This is not just true for X-Trans but also for my Bayer images. I've found profiled denoise to be the least useful module for denoising. > > > 3) When sufficient denoising is performed to remove maze and color > > artifacts, the difference between VNG and Markesteijn demosaicing is > > indiscernible. > > agreed. > > > I have a style preset which can produce images that match the camera > JPGs as > > far as noise/detail goes (utilizing the equalizer module) at ISO 3200. > > nice! > I'll post a link to that too. > > > Color is another matter, but I've ordered one of the Wolf Faust IT8 > charts > > to try and make an ICC profile. > > i used an it8 and the darktable-lut tool (only in git master > currently) to reproduce the provia/velvia/astia/classic > chrome/monochrome colour settings quite faithfully, using the > colorchecker lut module (can share if you're interested, want to > publish only once i tested them some more). i also have a reference > lut to produce correctly calibrated output, but tested that even less. > Excellent. I would be very interested in taking a look at those. > > > (I tried making one using the imaging-resource.com multi target studio > > shots, and while the results were OK, they weren't great). > > > > P.S. Is there any plan to support Fuji's RawExposureBias exif tag? It is > > necessary to display high ISO RAF images at the correct brightnesss. > > is that just a software exposure correction? never heard of it. > Yes. Basically the Fuji cameras don't apply digital gain to the high ISO raw images like every other camera on the planet, they just put this tag in the exif which tells raw converters how many EV underexposed the raw data is. The converter is supposed to boost the exposure by the specified amount. > cheers, > -jo > > > ___________________________________________________________________________ > > darktable developer mailing list to unsubscribe send a mail to > > darktable-dev+unsubscr...@lists.darktable.org > ___________________________________________________________________________ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
