Is there an estimate for when the final profiles will be out?
What happens if I then remove those beta profiles? Will LR fall back to the official profiles or will it not find any profile?
Thomas, images that use the new beta profiles will fall back to using the Adobe Standard profile if you remove the beta profiles. (This is generally true whenever you remove a profile that an image depends on.)
Am I assuming (hoping) correctly that the upgrade installation of Lightroom 3 doesn't update these camera profiles? I've been noticing a lot of posterizing as well as magenta and green colour artifacts in neutral tones since upgrading to my D700. I'm really hoping that updating my profiles when I get home will fix things.
Any word on when these will be out of beta? Will they be replacing the D3 and D700 profiles that are currently included in the Camera Raw 6.2 and LR 3.2 releases or are they going to be added as permanent alternate profiles alongside the current "V1" versions?
I tried searching Adobe Labs for updated information and can't even find any references to the beta profiles ... is there any place to keep tabs on their status? It's been a long time since the betas came out.
I may be doing something wrong, but I'm getting drastically different results when I open an NEF file in Capture NX2 versus Lightroom 3 (using both the boxed camera profiles and these beta v2 profiles) from images shot on a D700.
In NX2, I ensured that I was using the "Standard" profile, and in the case of LR3, "Camera Standard v2".
Here are the two different results:
The LR3 image looks much more color saturated, among other things, but from what I remember, the NX2 image was much closer to how the scene looked when I was doing the shooting.
Can someone help me understand why LR3 produces such a different default image from the NEF file than NX2?
Thanks in advance!
There were many posts about this in the past. So I took a look into Adobe's profiles for D700 (original and beta v2). I have some experience with this, because I made my own program for making dng profiles for some Canon cameras, that are not supported by Adobe. I noticed several things:
First, in the lookup table of both versions of the profile, values for Value segment 0 (= total black) are a simple copy of segment 1 (=dark midtones). My opinion is that this approach is a bit oversimplified, resulting with big differences in dark tones between ACR and NX (oversaturated blue and orange/brown on your picture). Midtones and highlights are much better
Obviously, points for total black can't be calibrated at all. What I did in my program was that I made calibration on very dark tones, and did some recalculation to aproximatelly match camera jpegs. It's not possible to get total match, but I came much closer
Second, there is a difference in tone curve between original D700 profile and beta v2. Tone curve in the original profile has a "notch" in the lowest part, caused by the fact that blacks slider default value should be 5 (according to Adobe). On the beta v2 profile, bottom part of the tone curve is smooth and looks similar to what I got on my profiles with blacks slider set to 0. So I suppose that for this profile (and some other latest profiles), blacks should be set to 0.
It would be nice to get some comment about this from Eric or Jeff
I attached a printscreen from my program that can display a lookup table in the profile, and this is how Hue segment 0 (= red) looks like.
On the left pane, there are value and saturation remapping points (in Photo Pro color space), and on the right, there are hue shifts
Also, there is a tone curve on the left pane (displayed in 3 different scales)
I'm not familiar with Nikon software, but in the case of Canon, there can also be some differences between sRGB and AdobeRGB. Canon has separate profiles for these color spaces and a kind of perceptual rendering is used. In ACR, there is only one profile for both color spaces. So in case you choose Adobe RGB, you will get aproximation of Canon's perceptual profile, bacause calibration is made on this one I presume, bacause it has bigger gamut. If you choose sRGB, out of gamut colors will be simply clamped into sRGB, instead of compressing into range, which would be done by the camera or DPP software
I don't know how things work technically in these two converters, but, from my experience with D700 camera which I own, NX almost completely desaturate dark tones, probably to mask shadow color noise or I don't know why but... I had really hard time to get any saturation for example in people faces when I do shooting in backlight situation, using flash fill, but flash was not enough powerful so it left them relatively dark. These areas in NX are simply gray, no matter what I do to boost saturation there. This looks like "feature" from NX, but I don't really like it, and ACR does much better job for these dark areas - at least for my eyes if not technically.
Canon compacts behave exactly the same way - very dark tones on in-camera jpegs are desaturated or completely gray, for the reason you mentioned. It's usually ok for the photo as is, but if you try to brighten it, in case it's too dark, you get strange results. Fortunately, my 400D doesn't seem to have this 'feature'
So, it's quiestionable whether Adobe camera profiles should precisely copy shadow parts of the camera profile (or other similar features) or do it in more acceptable way. But it was stated by Adobe that intention was to copy the profile as is, no mater how good or bad it is, because that's what people asked for, which turns out not to be the case here
We try to match the Canon/Nikon colors using the widest color space available in their respective software (Wide Gamut for the former, ProPhoto for the latter). This is done to minimize chances of clipping.
The problem with choosing sRGB / Adobe RGB and trying to match that, is that then we artificially (and needlessly) constrain the set of working colors. We did not find this to be acceptable.
It is also a reason why we don't match camera JPEGs directly. Camera JPEGs are either in sRGB or Adobe RGB. (This is a good choice, actually. ProPhoto RGB + JPEG exhibit significant problems.)
There are some limitations to our ability to match exactly. For example, we generally do not take into account color noise suppression (and its potential effect on dark tones), because this is strongly ISO dependent.
BTW, nice visualization a few posts ago, Vit -- I like it!
Can you comment the difference between the tone curve of original Camera standard profile and v2 profile (I attached similar diagram for v2 profile below). On original profile, tone curve has a "step" on the beginning (visible on middle curve - the brown one), while v2 profile tone curve doesn't. This "step" is compensation for Blacks=5. So, is v2 profile supposed to be used with Blacks=0, or there is some other reason
It's also evident that lookup table is different (especially hue shifts in this Hue segment = 0) , but this is because table in v2 profile maps from raw color space to PhotoPro, while in original profile it just remaps values in PhotoPro color space (I used the same way for CHDK profiles, because it has some advantages we allready discussed before)
The v2 profile is still intended to be used with Blacks = 5 (i.e., the default setting in ACR). The D3/D700 were more difficult for me to tone match because our default shadow/highlight clip points are defined differently than Nikon's default endpoints. The original profiles were designed to match as closely as possible through the entire tonal scale, but could lead to some posterization issues in both the deepest shadows & top stop or 1/2 stop of highlights. For the v2 profiles, I've chosen to use a smoother curve which reduces these artifacts.
You are right about the LUT differences. It helps to avoid clipping issues.
Thanks. I was just courious about the way you calibrated this profile, because I was dealing with similar problematic for my cameras
I also tried smoothing the bottom part of tone curve when making a profile for 400D, in combination with Blacks=5, but this resulted in deep shadows being darker than on jpeg (like in the case of your v2 profile for D700). After looking at the SDK code and some experimenting, it turned out that Blacks=5 clips bottom part of histogram, and that's why strange bottom part of TC needed. So i decided to use Blacks=0 in my case, which gave accurate result and simplified the calculation
Any chance of eliminating BaselineExposure (and blacks=5) in the future ? However, I'm aware that baseline exposure isn't fixed for the camera. For instance, it's usually -0.25 on 400D, but could be different in some occasions (longer exposure, usage of flash etc)
Are you still planning on releasing a finalized version of these profiles? (It'd be nice to get a version with similar-to-Nikon colors, but also without a dramatically steeper tone curve - even if it had to have blacks=0).
If not I'm wondering if it's possible to just adjust the tone curve of an existing profile (but I'm guessing that the color-matching is probably interdependent on the tone curve).
Hi photo81, yes, I would like to get a finalized version released. As you say, it may rely on some extra adjustments (which could be easily rolled into a preset). I will follow-up with you offline.
I tried running the DNG Profile Editor (chart tab) on a D700 image of the ColorChecker using Camera Standard version 2 profile and I get the following error message:
A fatal runtime error has occurred.
Description: invalid hue
Source filename: pe_recipe.cpp
Source line: 2734
The DNG Profile Editor works fine with the D700 Camera Standard regular profile. I only have a problem with version 2 profiles. Any comments?
Nice work, i realy like people getting this deep into color.
A lot better, indeed the burned out highlight are better.But stil i need to export to PS to make the image "pop"
In dark area's reds and orange are too saturated still. I'd like to sow you samples but i would not put them on public display.
In these files you can easily see what happens when push he blacks slider up. In dark tones a yellow blanket turns first orange and then red.
Same happens with faces. Which is still the biggest problem: The over saturated reds in bronzed (older) faces when a bit underexposed.
Second, i wonder, is it possible to develop an evan better tahn Nikon profile? With real natural colors? Natural greens and yellows for example?
i know, it's not what was the original request but wouldn't that be an awsome selling point?
But i remain very thankful for your work.
So, how do you define natural colors ?
If you for example tako a photo through the window, then display it in ACR or any other similar software and expect that you will see on the monitor exactly the same as through the window in the nature, I'm affraid that's not gonna happen with any profile
You are absolutly right. Natural is hard to define and non existing. On the other hand, unnatural is pretty easy to define. Like if yellow turns orange, blue is too purple. We are all trying to calibrate printers and monitors and i have the feeling that nikon colors could be improved. Maybe one could closer to natural or high end camera's. Or maybe just different, say Canon like..
I think the new v2 profiles are a lot better in that perspective already. Some odd purple green cast in shadows is now gone.
I noticed you are into improving the colors, interesting, I owned a 400d as well and liked it. Would share your profiles?
Well, I wouldn't say I wanted to improve colors. I just happened that a group of programmers wrote a software that enabled getting the raw from some Canon compacts. Unfortunately, there were no color specialists among them, so finally I made a software to make camera profiles for them, after I found out that I couldn't come reasonably close with dng profile editor (and also, I didn't have a color checker chart). As a byproduct, I also made a camera standard profile for 400D (with another program), as I wasn't entirely satisfied with the one from Adobe. It is closer to jpeg than the one from Adobe (at least in sRGB space), but to be ownest, on most real life photos, there is no big difference. I will publish it in CHDK forum in a day or two, so check there (I already had intention to do it, but since there was almost nobody interested in profiles for compacts, I didn't do it so far)
Some months ago I installed the Beta Profiles (D3) on my Mac (OS X 10.6.5).
Till yesterday I was able to choose the V2 Profiles during the import or later.
Now they are in LR3 (3.3) not available any more.
In PS CS4 I can see them.
I copy the beta profiles to /Library/Application/Support/Adobe/CameraRaw/CameraProfiles twice without success.
Does any one know what's the fault?
It is indeed possible to build a better profile, Menno. But we would need to do it for every user individually, because every user has a different notion of "better" or "more natural". And sometimes, better is not necessarily more natural. For instance, Velvia was (and is) popular for landscape photography, and I understand how it makes the colors pop, but I don't consider it very natural ...
That said, there are still areas in color where we can improve, and I would not be surprised to see additional advancements made in ACR/LR as time goes on.
safo, that is the correct place to which to install the profiles. Make sure in LR when you're trying to select the profile, that you are processing raw files (they will not show up in the Profile popup menu for JPEGs, TIFFs, etc.).