Your posted screen shot shows the expected, normal difference between the camera manufacturer's typically over-saturated, over-sharpened and over-contrasty profiles and what Adobe considers a more natural rendering of the image. The Camera Standard profile provided by Adobe is designed to show you a rendering along the lines of the former.
Here's the "camera standard rendering you posted with some heavy manipulation;
Joe Photo wrote:
How do I know what is the 'correct' color ?
Please allow me to suggest an alternative way to think about color... IMHO there is no such thing as "correct" color. Only "pleasing" color. Keep in mind that "pleasing" could include "that looks about the way I remember it."
Until we get displays not unlike the view screen in Star Trek, no display device or print can represent reality in all the dynamicism of reality itself, therefore everything you see in a photograph is just a recognizable facsimile of the original.
That there is a difference between the results from using different profiles sometimes bothers people. I believe that if Adobe were to default Camera Raw to produce color that's very close to the original in-camera / camera manufacturer-supplied conversion this would come up less often, even though the camera manufacturers don't always make the best choices.
Adobe does offer some capability to make your very own camera profile, by the way, via the DNG Profile Editor available from Adobe Labs.
1 person found this helpful
There are instances where "CORRECT" colors are a requirement, e.g. when photographing paintings of the Old Masters. In such a case, you'll need total lighting control, a professional-level scanning camera back and a full color checker chart in the picture.
Other than such specialized instances, Noel is right. You are the judge of the colors in your work.
What has to be understood is that ACR does not mimic the results of the camera manufacturer's conversion software. The "Camera Standard" is a recent concession to those users who want the closest approximation to the camera software, without the extremes in saturation and contrast chosen by the camera manufacturers software designed to appeal to a majority of mass consumers.
station_two accurately points out that the 'Camera Standard' profile is one that acr imports from Canon but it seems definetly closer to the true color that I remember. But memory, especially mine, is a unrealible choice for catalog work.
This IS a situation that color HAS to be correct as I am shooting for a flower breeder catalog and why I shoot with gray card and studio lights. I probably need to inlude a color chart. Unfortunately the shoot was out of town and I no longer have the flowers to compare for "correct" color
You will note there is no appreciable color difference in the gray card or background color of the two renderings of this photo. The most likely explanation for the pink/magenta color shift is infrared color pigment in the flower that the human eye does not see, but the camera does see (and, to the delight of biologist, so do birds and bees also see) . This is a well known phenomenon in flower photography.
If I were to include a color chart as startion_two suggests, or perhaps a pantone swatch that was close to true color, it would not have the infrared flower color shift, and thus be a better target for post production. Still I wish there were a better way to be assured of correct color using a profile.
I see that you have manipulated the Camera Standard photo to get close to the Adobe Standard and it is a Photoshop chore to be done outside of ACR that I am hoping to avoid. I normally set my Bridge default preferences to Adobe Standard but started messing around with the Camera profiles in acr when I felt the color was off
Thanks Noel. This is a situation where color is not subjective but I no longer have the real color (of the flower next to me) to compare. I have looked at the adobe labs dng profile editor to create profiles and must confess it is way beyond my understanding
1 person found this helpful
Getting the correct color has been age old challenge, long before the advent of digital cameras. With film, I used a color meter, made upteen clip tests getting the correct filter for each batch of film I used; one of the many reason pro's bought bricks of film and kept them in the freezer so they had a stash of calibrated film from the same batch. Each new batch required a new found of testing. Same went for some lens combinations as well, and using strobes a must to test, some strobes their color temperature would change over a matter of a few hours. But shooting test targets through the shoot should take care of that.
To meet this challenge today I shoot a calibrated test target, that allows be to set the white balance as well as create a custom camera profile for each shoot. I use both the X-Rite (location work) and the Datacolor Checkr in the studio. They work quite well, but you do need to have a calibrated work flow from capture (camera), display (monitor) and printer, and recalibration on a regular basis is a must; I recalibrate on a weekly basis, one of my Friday chores.
Since upgrading to the D800 it has become my mainstay, before that I shot with both Nikon (D3) and Canon (5D MKII) and the color differences from the same shoot were amazing to say the least. So for accurate color you do need a calibrated system, both the camera profile products I use are reasonably priced X-Rite ~$99 and the Datacolor ~$140.
Thanks Mike - I well remember testing bricks of film and keping gel filters to correct, though I admit to only using them on the most demanding commercial clients who would actually knew enough or had time enough to care and actually follow through with the color separations and color corrections with the printer.
Anyway, I admit I have not used a calibrated test target for my camera. I do calibrate the display regularly but have only been using a gray card for camera and white balance, thinking that woudl zero out the color as well. I will investigate the camera custom profile.
I wonder though, why the neutral gray in my test stays nearly the same in both profile renderings of the flower ? In the old days of film testing we would correct to the gray card. As I write this I am realizing the answer is each digital color probaby needs to be calibrated ? This is all a separate issue from infrared shift in certain flower colors.
Probably the response of the sensor and how the hardware/firmware and software interpret and output the image. I have found that straight out the camera the Nikon sotware gives a better color rendition than the Adobe Standard, that is probably because the developers have access to sensor data that Adobe would not have. But I have found creating a custom profile for my camera for differing shooting conditions, can match and often beat the output from ViewNX2.
Also I have noticed the newer generations of cameras the color fidelity from the sensors is getting better. A custom profile for the D800 shows subtle changes, whereas the a profile made from the same shoot with my Nikon D3 shows very noticeable changes, the same with the 5D MKII'
My biggest challenge was getting the same image fidelity in print. I print with Epson printers, and while the out of the box color profiles for their paper and other paper suppliers are good, not good enough when you need to match corporate colors for a client. Some corporations not only trade mark their logo but also the colors; and art directors can be very picky. I have found the X-Rite Colormunki Photo to be pretty accurate for monitor and print calibration, not eactly cheap, but being able to build printer profiles is a blessing. The other nice thing about Colormunki Photo is they throw in Passport for free, you only get a smaller version of the color swatches, and they are not housed in a nice plastic case like the standalone product - but it works
Thanks again. It is great when you can work in a closed system for calibration all the way to making your own prints and custom paper profiles. Most of my work is editorial and I must depend on increasingly untrained junior photo editors or copy editors who depend upon the pro photographer to provide final files. Well and good, and I like being able to present my work as professional quality but it is pretty obvious in the editorial world anyway, that the concept of calibration is not well implemented. There is a growing disconnect between those who pick a photo for publication and those who prepare the files for press.
While I have you here and a bit off subject, but what color temp do you use for monitor calibration ? There seem to be two standards, either 5000 for print or 6500 (or even 7200) default for web. Won't a beautifully calibrated file for print look quite different when they put that file on-line ?
I calibrate my monitor to D65 (6500), and that matches the lighting conditions in my studio office which is dim with virtually no natural light. I do have the option of going to D50 (5000) which I would use if there was more natural lighting. But testing showed D65 worked best for me in my work environment.
The majority of my work ends up in print, but I do prepare images for the web and that can be a crap shoot depending on the end user browser. For web displayed images I convert to sRGB and test on the major browsers. The web based images I put out are typically pushing the product so color fidelity while important is not paramount. But images going into catalogs, booklets, corporate communications and the like color fidelity is paramount. But slowly browsers are becoming color aware and a few of them can use the embedded profile.
With the companies I work for I have an agreement that I produce the final images for the web and no one else touches them. Had some problems in the past where some one in the marketing department tried editing some of the jpegs and made a real pigs ear out of them, color was off, some had artifacts, poor cropping and more. So after the dust settled and I showed the CEO what I had delivered he saw who the real culprit was, and we agreed that I had control over delivery, and have used that model ever since and have a disclaimer on my purchase order.
I do prepare images for the web and that can be a crap shoot depending on the end user browser.
Amen. And the end user monitors of course, most of which are too cheap to be able to be profiled (e.g., because of shifts due to viewing angle).
I managed to get Microsoft to admit once (privately) that they've programmed the color-management in the Internet Explorer browser in a half-baked fashion - to always assume the monitor is sRGB regardless of profile. It's done on purpose but they wouldn't say why - and it's not going to change any time soon. I figure there could be one or a number of reasons:
1. More folks might complain about problems, raising their support cost, if IE were to embrace monitor profiles. It's an element outside their control, though one wonders why they couldn't just create a truly robust color-management module that would either work or fall back to sRGB if it failed. Perhaps Microsoft doesn't have that kind of talent.
2. Microsoft's priorities and thus their direction have gone a bit wacky, with their recent all-eggs-in-one-basket pursuit of the tablet market. Have you noticed that color accuracy seems to be a non-priority for portable electronics? As long as red things look sort of reddish, good enough?
3. Most non-technical / non-professioal users don't really care about color-management, and in fact may run away from it screaming. Laypeople tend to like the "more vibrant" color they see when an sRGB image is displayed on a wide-gamut monitor. Where's the incentive to desaturate those colors? Back to item 1 above.
Microsoft stopped developing a Mac version of IE years ago, yet the last version they released still runs on Macs, poorly, but it's fully color managed. Other, previous versions were too.
1 person found this helpful
Final folow up.
I sent a selection photos of side by side comparision to my flower breeder client, and in every case the Adobe Standard calibration was better than the Camera Standard. Lesson learned.
I earned some brownie points with them for this exercise but it should NOT have gotten to this point.
I am glad to hear this result (we do have lots of flowers in our training data, after all), but keep in mind that no single color profile can get the colors "correct" for all materials. This is because profiles are necessarily optimized for some criteria, and it's impossible to satisfy all criteria simultaneously. For example, blue hydrangea and morning glory petals are notoriously hard to get right without messing up other common colors. This is one reason why we have the DNG Profile Editor which can be used to tweak profiles to suit your specific needs & preferences.