Pages: [1]   Go Down

Author Topic: Sigma DP Merrill custom icc profiles, need some ideas  (Read 2625 times)

Ligament

  • Full Member
  • ***
  • Offline Offline
  • Posts: 212
Sigma DP Merrill custom icc profiles, need some ideas
« on: July 06, 2014, 10:55:23 pm »

I've been trying to emulate Martin Zimelka's excellent idea of creating a custom icc profile for the DP Merrill series using Argyll

http://www.martinzimelka.com/homepage/Blog/Entries/2013/7/28_Sigma_DP2m_workflow.html

I've installed Coca via wine on my mac, have shot images of the QP Card 203 book.

I opened this image in SPP, keeping all settings at 0, white balance set to sunlight, color mode neutral.

question: should I set custom white balance in SPP using the neutral grey patch on the QP Card?

Save as TIFF 16-bit, ProPhoto color space.

import this TIFF into Coca

generate a icc profile

Place this icc profile into library/colorsync/profiles

I then open colorsync and get the most absurd looking Lab plots (see attached).

1. Why are the plots coming out so odd?
2. Assuming I can figure #1 out, what do I then do in ACR or photoshop to apply the icc profiles?


Logged

PhilipCummins

  • Full Member
  • ***
  • Offline Offline
  • Posts: 133
Re: Sigma DP Merrill custom icc profiles, need some ideas
« Reply #1 on: July 08, 2014, 07:19:51 am »

Hazarding a guess:

Reading from Martin's website it looks like he uses the Sunlight WB preset to create the profile to avoid ambiguity over a custom WB which would basically bring an extra variable into the profile generation phase. In theory the profile should be able to pick up the correct adjustments from the patches in the QPCard (18% grey or similar).

I would try saving as a 8 or 16-bit Adobe RGB or sRGB profile vs ProPhoto to see if the Lab plots work out better. If you can generate a correct ICC input profile you would then apply it to the image when opening in Photoshop and compare with some of the same images without the profile to find out which one you prefer. You could then try adjusting the WB using the QPCard at this point.

If you want you can try comparing it to some sample profiles I made with my DP2M to see what you think. Unfortunately capturing the correct output on a ColorChecker SG (or similar) is pretty difficult but it would give you an idea of the subtle differences the input profile can make. (For myself, it managed to increase some blue sky saturation and smooth out some odd colours for foliage). If I had some more time I'd been interested in trying to find a better way of handling it.
Logged
Pages: [1]   Go Up