Here's my attempt. The base curve gets the histograms pretty close, and the velvia module boosts the midrange saturation to approximately that of the preview jpeg.
As for whether there's an "easy way"... maybe somebody will write a nice histogram solver some day, but my method is to snap a screenshot of the sample jpeg's histogram and pin it up next to the "live" one for reference while tweaking the curve, it's a lot easier than flipping back and forth. For whatever that's worth... Cheers! On 2013-05-27 [email protected] wrote: > I've been using darktable for a month or two and in general I feel I > can find my way around it and get good results for my photos, but every > so often I get stymied in processing a particular picture and can't get > what I feel are good results. This is especially frustrating to me when > the out of camera JPG rendering seems to be a decent starting point and > I can't even duplicate that. > > Consider the following NEF and embedded preview image: > http://www.cs.toronto.edu/~cks/tmp/darktable/DSC_1026.NEF > http://www.cs.toronto.edu/~cks/tmp/darktable/DSC_1026-preview4.jpg > > (The preview is generated using a version of the Nikon 'neutral' picture > style that increases the sharpening and reduces the contrast slightly.) > > I was unable to get a result that didn't blow sections of the sky > out while still keeping the foreground light enough (and maintaining > contrast). Several experiments wound up with the sky too dark, too > contrasty, or both. I could probably have fixed this with sufficient > use of masks, but I think that I'm missing something in basic curve > processing. And in general I'd like to work out a relatively easy and > fast way to get to a starting point similar to the preview JPEG; the > preview image still needs work and so if it'd take me a lot of work to > get even there, the photo probably isn't worth it at all. > > (In general I've found that the darktable base curve(s) are often much > more aggressive about blowing highlights than I'd like. My attempts to > fiddle with the normal Nikon base curve to fix this have generally been > unsuccessful. I've had more luck resetting the base curve to linear and > then applying the 'med contrast' tone curve preset, but even that's not > fully there.) > > Thanks in advance to anyone who feels like taking a shot at this. > > (I should mention that I'm currently using the git tip version of > darktable.) > > - cks > PS: I know that this is not a great picture in general but I think it > makes a good illustration of the disconnect between the camera > preview image and what I can achieve in darktable with my current > understanding (probably better than a better picture, honestly).
D7100-jys-attempt.dtstyle
Description: Binary data
<<attachment: screenshot-1369707055.png>>
------------------------------------------------------------------------------ Try New Relic Now & We'll Send You this Cool Shirt New Relic is the only SaaS-based application performance monitoring service that delivers powerful full stack analytics. Optimize and monitor your browser, app, & servers with just a few lines of code. Try New Relic and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_may
_______________________________________________ Darktable-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/darktable-users
