So I'm just curious. Why would the Adobe supplied profiles vary so much for different camera models. I understand each model is different, and thus profiles are different. But I would expect the "delta" from the Adobe Standard profile to be roughly the same across models.
The camera-specific profiles are intended to mimic Canon's own picture styles. If Canon made 5D's Faithful different from 7D's — so did Adobe. At least that's the intention, as I understand. There's no reason to expect that one camera's Faithful DNG profile will match one from other camera.
How do the Faithul profiles for bot cameras behave in DPP?
I was told, a few weeks ago, that "Camera Standard" was the profile Adobe makes like the camera manufacturer's interpretation at default settings. Now you're implying it's "Camera Faithful".
I'm curious: Which is it supposed to be?
I must have missed the memo.
In any case, neither is particularly close for my Canon 40D, though "Camera Standard" seems to be closer.
Camera Standard (and the various other Camera-prefixed profiles) for Canon models will roughly match what you'd get out of DPP (as opposed to the camera JPEG) at defaults, with the various auto corrections disabled (e.g., peripheral illumination correction, auto lighting optimizer, etc.).
Camera Faithful uses a tone curve with less contrast than the default Adobe Standard curve, so you'll likely see some lightening of the shadows if you compare the two directly.
Noel, "Camera Standard" profile is the profile that should give similar results as DPP with "Standard" picture style selected with saturation, contrast etc at default
"Camera Faithful" profile is the profile that should give similar results as DPP with "Faithful" picture style selected with saturation, contrast etc at default
They slightly differ from both DPP and in camera jpeg in case of 400D also, so I'm currently working, as experiment only, on my own program to make better calibration - it's not an easy task ... but first results are quite promissing ...
Thanks to all, especially Eric, for the responses. Took me a while to work the Canon DPP and DNG Profile Editor programs in some more experiments. I use these so rarely I always have to re-do the learning curve.
Like most, I suspect, I've learned that using ACR to match in-camera jpegs or DPP processing is doable, but a sometimes difficult task. Personally, I don't care. I like what ACR does and use it exclusively. But I remained curious why I was getting such different, and opposite results on 5D2 images vs. 7D images.
By looking at the tone curves in both the DNG Profile editor and DPP, the answer was clear. DPP has fairly consistent curves for the various profiles, while the Adobe profiles vary considerably.
The following are composite charts I made from screen caps, plus a little layering in Photoshop to superimpose the curves. In the DPP curves it's easy to see that the "Faithful" and "Neutral" curves are slightly lower than the curves for the "Standard", "Portrait", and "Landscape" profiles. And the DPP curves are consistent across the 3 camera models I tested.
The Adobe profiles, however, vary considerably between camera models. For the 1Ds, the Adobe Faithful/Neutral curve is slightly higher than the Standard curve. It raises tones fairly evenly across the entire tonal range.
For the 5D2, the Adobe Faithful/Neutral curve is lower than the Standard, and more so in the midtones than in highlights and shadows.
For the 7D, the Adobe Faithful/Neutral curve is lower than the Standard, mostly in the highlights, but only slightly.
No real point here, I guess, other than understanding that Adobe profiles are not consistent across camera models. I had come to prefer the Camera Faithful profile for my portrait shots with the 5D2 because I felt it gave more pleasing skin tones by lowering the hot spots. When I tried to demonstrate that to my "student" with his 7D I got the opposite effect, which led to my confusion.
Thanks for the follow-up. Very interesting information!
redcrown on guard wrote:
the Adobe profiles vary considerably.
You have once again discovered, as many have, that Adobe is a curious combination of art and science.
From what I found out so far, approach to calibration in Adobe also seem to undergo some changes during time, so it's understandable that not all profiles are equal
So Adobe standard profiles for recent cameras are quite different than Adobe standard for older cameras. Older profiles don't seem to be close to colorimetric rendition, while new profiles seem to be. Also, older profiles are de facto single-illuminant, while new profiles are dual illuminant (a sort of)
But radical change is in Camera profiles. So far, there are profiles for two cameras (Canon 550D and Nikon D3S) where different approach to color conversion is used, and this approach is better for several reasons (one of them was discussed several months ago).
However, looking at profile for 550D, I feel there is still room for improvement ...
Regarding room for improvement:
One of the things I found when I was fooling with the profile editor (back a few years now) was that at the time it was not hard to make a profile that worked with a default EV setting of something less than 0 (e.g., -0.5 EV) that recovered more highlights than the default profiles did at 0 EV. The profile brightened everything else up so that the exposure came out about right overall, but rolled off the highlights instead of clipping them. Voila, instant dynamic range increase, and very handy for experimenting with "exposing to the right".
Though I now use the Camera Standard profile (with lots of color tweaks) I still convert with a default of a slightly negative EV value by default so as not to lose highlights that are really there in my 40D raw images.
Yes, baseline exposure value is a kind of rudiment from the days when ACR didn't use profiles, so it was necessary to adjust brightnes to similar level as photos from camera. So when used with raws from my 400D (and 40D probably the same), top 0.25 EV is cut off, which is not the case in DPP or photos from the camera. If I apply recover slider, this headrom is recovered, but colors are changed (they are not most precise with recover 0 anyway)