Ulysseita
Offline
• •
Upload & Sell: On
|
Mitch Alland wrote:
I find the marketing message of the cobalt-image site confusing. Looking at it from the perspective of the two cameras I use, Leica M10 and Ricoh GRIII, I don't see what it will give me compared to other profile/preset packages: better matching between the two cameras as a starting point for further processing, or better film simulation?
For example, I've looked at The Archetype Process ("TAP" f.k.a. C1ick Match) and the RNI packages. While RNI has profiles linked to the camera profiles in CameraRaw/Lightroom (including M10 and GRIII), the RNI profile/preset combination don't seem particularly useful.
On the other hand, the TAP profiles have profiles based on Noritsu and Frontier scanners, which have different color rendering; but have no linking to the camera you are using. While some of the TAP profiles are interesting, they all have excessive reds and oranges when used on M10 files: they may have been tuned for Canon and Nikon files. I’ve asked a couple of times on the TAP FaceBook group about the issue of linking to camera profiles, but the developer has not responded to these questions. In my view this is a major deficiency of TAP, particularly considering the price of the packages. Paradoxically, I usually find the TAP profiles a better starting point than the allegedly "more accurate" RNI profiles/presets.
How do the cobolt-image profiles compare to TAP and RNI packages?
I take it that cobolt-image has no Kodak film simulations except for Kodachrome, although this thread has some examples of Portra and E100G examples, if I remember correctly, What is the status of Kodak film situations?
____________________
Frog Leaping photobook:
https://www.frogleaping.org ...Show more →
We are in a different league.
Is not an overestimation, other products are just artistic, free interpretations of film types, or digital types, nothing scientific.
We start from a base, with output IDEALLY comparable between cameras, and that is the first step, the most important from a photographer, a professional photographer perspective, having many bodies and THE SAME color output with this level of reproduction:
-600 samples tested (sg has 140 )
-two main illuminants d65 and StdA (DNG standard)
-different transformation than the Bradford model used by Adobe for the d50 calculation.
-for DNG the internal forward matrix is limited in the Prophoto/ Human Locus in order to avoid posterizations.
-Looktables and ICC (C1) are adjusted to be relative to any different contrast curve.
-CIECAM02 is the transformation model (better tonal separation)
-main colorimetric stadium is proprietary, to build profiles we use DcamProf.
here is a sample of delta E between Adobe STD and Cobalt STD on A7r2 for CC24 of our former Cobalt 2.0 version (we are in 2.1)
Is just a basic test because of the simple cc24.
On CG the difference is way bigger.
Adobe D65 : patch match average DE 1.63, DE LCh 0.64 1.10 0.78
median DE 0.76, DE LCh 0.40 0.45 0.21
p90 DE 4.37, DE LCh 1.79 3.68 2.27
max DE 8.96, DE LCh 3.54 6.10 8.40
Cobalt D65 : patch match average DE 1.37, DE LCh 0.67 0.96 0.52
median DE 0.70, DE LCh 0.41 0.38 0.14
p90 DE 3.63, DE LCh 1.71 2.77 1.57
max DE 6.31, DE LCh 3.37 5.17 4.98
ADOBE
https://cobalt-image.com/wp-content/uploads/2019/07/pgamut-adobe-r5.png
COBALT
https://cobalt-image.com/wp-content/uploads/2019/07/p-gamut-cobal-r5.png
After this base, we build emulations.
As a sample, the Monochrom Leica CCD is made using a math model over the above profile to MATCH the original DNG with a Delta E of 0.8.
That means on the field to be able to have this incredible accuracy on this emulation over ANY camera supported by our base profile.
Same for Fuji and every digital emulation ( we are fixing the R) and for FILM, I already wrote our point of view and our vision.
So, as already said, is not an overestimation, it's a fact.
And.
About Capture One and Cobalt.
We can even forget to remind you about the better color reproduction of Cobalt over the PRO by C1.
Just a single thing makes a huge difference; Capture One uses Single-illuminant profiles, so if the profile is made for D65, it's a fact, will be off on stda, or other illuminants.
That's why Phase One makes profiles for every illuminant on their cameras and the cameras cost many thousands of dollars; you are buying THE color reproduction after resolution and IQ.
That's why we sell TWO profiles for two different illuminants over any camera we support for C1.
So, a better color reproduction compared to the standard and for different light situation; you can love even a bad color profiles for your taste, we are not selling NICE color profiles for your cameras, we are selling PERFECT color profiles for your cameras, the artistic side is still yours, we only offer the ideal starting point for your post-production, as Phase One or Hasselblad do with their customers after the huge investment and nobody does with mainstream cameras.
Is it still crazy to have a color profile for every camera to buy after that clarification?
|
Archive 2021 · Cobalt Profiles
