The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

Xrite Camera Profiling Software

RLB

Member
While I don't doubt the ability of the Xrite software to make a profile I'm skeptical about the quality of that profile based on my past expereice with other systems that are far more advanced.

My concern with the Xrite set up are regarding the aparent lack of sophistication of the process and software:

- Small patch set with only 24 reference tiles. Previously I've profiled cameras using a 200-300 patch set reference target.

- No redudant patches in the target to account to verify the quality of consistancy

- No direction about the type of light source to use for the target, how to light target, etc. which will have a significant effect on the
quality and accuracy of the profile

I too wish that there were a good, inexpensive and easy to use system for profiling a camera, however thats fantasy according to my 20+ years of color management consulting and building profiles for both cameras and printers. Honestly, a $40-$50k DB should absolute come from the manufacturer with such a custom profile, and it should not be the end users task to make such. JMHO.

As a start, if you are looking for accurate color I'd suggest an Eizo monitor calibrated correctly, use Linear for the processing curve, and have a quality spectrally neutral gray card in the image set you are shooting. I've tested many greay cards and sadly most are not as spectally neutral as they claim. If just one link in your color management chain is broken, your end result will suffer.

Robert
 

etrump

Well-known member
Hope I’m not wrong Robert, but it is my understanding the icc profiles are layered over the more extensive profiles created by P1. I’ve routinely created profiles with just a few specific colors for colormatching purposes. I would consider the color passport target as something similar to do more of a normalization process.

While I don't doubt the ability of the Xrite software to make a profile I'm skeptical about the quality of that profile based on my past expereice with other systems that are far more advanced.

My concern with the Xrite set up are regarding the aparent lack of sophistication of the process and software:

- Small patch set with only 24 reference tiles. Previously I've profiled cameras using a 200-300 patch set reference target.

- No redudant patches in the target to account to verify the quality of consistancy

- No direction about the type of light source to use for the target, how to light target, etc. which will have a significant effect on the
quality and accuracy of the profile

I too wish that there were a good, inexpensive and easy to use system for profiling a camera, however thats fantasy according to my 20+ years of color management consulting and building profiles for both cameras and printers. Honestly, a $40-$50k DB should absolute come from the manufacturer with such a custom profile, and it should not be the end users task to make such. JMHO.

As a start, if you are looking for accurate color I'd suggest an Eizo monitor calibrated correctly, use Linear for the processing curve, and have a quality spectrally neutral gray card in the image set you are shooting. I've tested many greay cards and sadly most are not as spectally neutral as they claim. If just one link in your color management chain is broken, your end result will suffer.

Robert
 

dougpeterson

Workshop Member
I too wish that there were a good, inexpensive and easy to use system for profiling a camera, however thats fantasy according to my 20+ years of color management consulting and building profiles for both cameras and printers. Honestly, a $40-$50k DB should absolute come from the manufacturer with such a custom profile, and it should not be the end users task to make such. JMHO.

As a start, if you are looking for accurate color I'd suggest an Eizo monitor calibrated correctly, use Linear for the processing curve, and have a quality spectrally neutral gray card in the image set you are shooting. I've tested many greay cards and sadly most are not as spectally neutral as they claim. If just one link in your color management chain is broken, your end result will suffer.

Robert
Capture One Cultural Heritage comes with bespoke profiles for every modern P1 back. They are incredibly accurate and require no fine tuning by the end user.
 

dchew

Well-known member
Robert,
Many good points here. I've had limited experience with camera calibration and have several questions about your comments:

While I don't doubt the ability of the Xrite software to make a profile I'm skeptical about the quality of that profile based on my past expereice with other systems that are far more advanced.

My concern with the Xrite set up are regarding the aparent lack of sophistication of the process and software:
- Small patch set with only 24 reference tiles. Previously I've profiled cameras using a 200-300 patch set reference target.
- No redudant patches in the target to account to verify the quality of consistancy
- No direction about the type of light source to use for the target, how to light target, etc. which will have a significant effect on the
quality and accuracy of the profile
I also wonder about what Anders brought up in a similar thread on Lula: The lack of ability to adjust to something different than the linear curve. I have no way of making a profile specifically for the film curve in C1. That may or may not be a big deal. So far, my one test of the x-rite / C1 profile process was reasonably successful. I took hundreds of images of leather samples in late Dec, just prior to this release. At that point I tried three different profiles: IQ 4150 Flash, IQ Flat reproduction and Leaf Product (for the 4150). Of those three "canned" profiles, I found the Leaf Product profile / Film Curve to be the best for most colors. In some cases, it rendered the sample a bit warm, even though the white balance was neutral and consistent across all samples. After this custom profile release, I created a profile and applied it to all the images (I have a Colorchecker image with each sample). It is too soon to tell since I have yet to compare the actual samples, but it "appears" as though the custom profile does a much better job with at least greens and blues, and perhaps almost all of the images. And, that was even with applying the Film Curve. Because the images were shot assuming I was using the Leaf curve, they were too underexposed for Linear.

I too wish that there were a good, inexpensive and easy to use system for profiling a camera, however thats fantasy according to my 20+ years of color management consulting and building profiles for both cameras and printers. Honestly, a $40-$50k DB should absolute come from the manufacturer with such a custom profile, and it should not be the end users task to make such. JMHO.
But I always thought the main point about making these custom profiles was to handle specific lighting situations. For example, my setup included the relatively new Fiilex LED lights. How would any of Phase One's profiles, even if custom for a specific back, matter in that case?

As a start, if you are looking for accurate color I'd suggest an Eizo monitor calibrated correctly, use Linear for the processing curve, and have a quality spectrally neutral gray card in the image set you are shooting. I've tested many greay cards and sadly most are not as spectally neutral as they claim. If just one link in your color management chain is broken, your end result will suffer.

Robert
Interesting. I too have noticed this, especially in the “tint” or green/magenta direction. I have several white-balance cards, ranging from several colorchecker classics, a Passport, whi-bal’s and some QPCards. They all seem to have slightly different white balances that can be noticeable. Here is an image white balanced on the colorchecker. The Whi-Bal has a definite magenta cast.

 

RLB

Member
Hope I’m not wrong Robert, but it is my understanding the icc profiles are layered over the more extensive profiles created by P1. I’ve routinely created profiles with just a few specific colors for colormatching purposes. I would consider the color passport target as something similar to do more of a normalization process.
Ed, you bring up a good point as to if the Xrite system is actually creating a new profile or simply an adjustment set to an existing one. If anyone else knows, maybe they can chime in. I'm hoping to test the Xrite software soon and evaluate the profiles in Colorthink.

R
 

RLB

Member
As you know, Linear will typically be the most color accurate, but requires about 1 stop more light when capturing the images. Therefore it can be difficult to obtain the best quality if one shoots with Film curve in mind and then later tries to process with Linear.

There are a couple of issues here regarding light source. First, light source and spectral response. Second, how certain pigments and dyes respond to various light source spectrums. Ideally, one would create a custom profile with the specific DB and the specific light source in order to obtain the most accurate profile. IE: Daylight, strobe (specific brand and head), tungsten, LED (specific brand and model), HMI (anyone use these anymore?), etc.

I've tested quite a few "gray" card for spectral neutrality and while most are decent there is definitely a range in the results. One would guess that the manufacturer has some sort of QC going on as well for consistent quality throughout various production runs. One of our cats (gray of course) actually has remarkable spectrally neutral fur.
This is handy when he demands to get on set.
 

RLB

Member
Robert,
Many good points here. I've had limited experience with camera calibration and have several questions about your comments:

I also wonder about what Anders brought up in a similar thread on Lula: The lack of ability to adjust to something different than the linear curve. I have no way of making a profile specifically for the film curve in C1. That may or may not be a big deal. So far, my one test of the x-rite / C1 profile process was reasonably successful. I took hundreds of images of leather samples in late Dec, just prior to this release. At that point I tried three different profiles: IQ 4150 Flash, IQ Flat reproduction and Leaf Product (for the 4150). Of those three "canned" profiles, I found the Leaf Product profile / Film Curve to be the best for most colors. In some cases, it rendered the sample a bit warm, even though the white balance was neutral and consistent across all samples. After this custom profile release, I created a profile and applied it to all the images (I have a Colorchecker image with each sample). It is too soon to tell since I have yet to compare the actual samples, but it "appears" as though the custom profile does a much better job with at least greens and blues, and perhaps almost all of the images. And, that was even with applying the Film Curve. Because the images were shot assuming I was using the Leaf curve, they were too underexposed for Linear.



But I always thought the main point about making these custom profiles was to handle specific lighting situations. For example, my setup included the relatively new Fiilex LED lights. How would any of Phase One's profiles, even if custom for a specific back, matter in that case?



Interesting. I too have noticed this, especially in the “tint” or green/magenta direction. I have several white-balance cards, ranging from several colorchecker classics, a Passport, whi-bal’s and some QPCards. They all seem to have slightly different white balances that can be noticeable. Here is an image white balanced on the colorchecker. The Whi-Bal has a definite magenta cast.

Ideally one would make a profile using the same DB and light source that they plan to shoot under. That said its based on the spectral response of the light source...so sunlight/daylight could be one profile no matter what time it was shot. LED pose the biggest challenge IMHO since there are so many different model and manufacturers.

Gray cards. We've found a substantial difference in quality in true spectral neutrality from brand to brand. One wonders if there is any QC within each brands production runs, one would certainly hope so.

While P1 does offer the Cultural Edition which has more profiles to choose from, its quite expensive, and if one is already shelling out $47k for a IQ4 one would think that those profiles should be included in C1 standard, of course, without the extra functionality that the CHP version offers. JMHO.
 

RLB

Member
Ideally one would make a profile using the same DB and light source that they plan to shoot under. That said its based on the spectral response of the light source...so sunlight/daylight could be one profile no matter what time it was shot. LED pose the biggest challenge IMHO since there are so many different model and manufacturers.

Gray cards. We've found a substantial difference in quality in true spectral neutrality from brand to brand. One wonders if there is any QC within each brands production runs, one would certainly hope so.

While P1 does offer the Cultural Edition which has more profiles to choose from, its quite expensive, and if one is already shelling out $47k for a IQ4 one would think that those profiles should be included in C1 standard, of course, without the extra functionality that the CHP version offers. JMHO.
 

Steve Hendrix

Well-known member
Ed, you bring up a good point as to if the Xrite system is actually creating a new profile or simply an adjustment set to an existing one. If anyone else knows, maybe they can chime in. I'm hoping to test the Xrite software soon and evaluate the profiles in Colorthink.

R

Since the Profiles are settled in the "Other" folder in Capture One Base Characteristics, then if they are layered on top, this would apply to any profile selected in the "Other" folder. Actually you choose an alternate profile no matter what non Phase One profile you're choosing, so not sure how it would determine which Phase One profile to dedicate as the base profile layer if you're not specifying it. Not sure if it is the case or not, regardless of which profile is selected. I suspect not, but don't know for sure.

Semi-related sidebar, I consider the profiles in Capture One a great source of "Color Styles", just run your mouse up and down and stop where you like (for esthetically pleasing color, not for technically accurate color). How about a Canon EOS 4000D Generic Profile on an IQ4 150 for an upper midwest landscape shot - not bad!

There are some very serious color profiling tools out there, some are quite intensive, expensive, and versatile. As Robert points out, there are many nodes to optimal color management, the more of them you can check off correctly, the better your results. On that note, I have been told that the Xrite 24 patch color checker is not the end game for this application, just a starting point. There are plans for charts with many more color patches available. What the Xrite development does is bring a very user friendly - not to mention Free - profiling tool to the masses using Capture One.



Steve Hendrix/CI
 

dougpeterson

Workshop Member
Gray cards. We've found a substantial difference in quality in true spectral neutrality from brand to brand. One wonders if there is any QC within each brands production runs, one would certainly hope so.
There is no such thing as a truly spectrally neutral target, only varying levels of close-to-neutral (in the same way that nothing in the universe can reach absolute zero, or that you can never have a totally silent room). But in most cases the good targets are close enough that there is no practical effect of their slight deviations from neutral.

Image Science Associates targets are the targets we recommend for those doing hyper critical work.
 

dougpeterson

Workshop Member
Ideally one would make a profile using the same DB and light source that they plan to shoot under. That said its based on the spectral response of the light source...so sunlight/daylight could be one profile no matter what time it was shot. LED pose the biggest challenge IMHO since there are so many different model and manufacturers.
Correct.

That's why DT ended up having to make our own LED light source for Cultural Heritage work. We couldn't find anyone making one with a high enough CRI/CQS and a flat enough spectrum for hyper-critical work, and when we saw manufacturers that came sort-of close, they didn't have the production consistency and QC required.

Every DT Photon is hand-balanced at the time of creation to a homogenous spectrum (most manufacturers barely QC the white balance; nobody else that I'm aware of does 100% QC of the spectrum) and the spectrum is recorded at our factory during burn-in testing before shipping. Because every DT Photon is made to the same spectral spec (one that is very flat, and therefore easy to profile) and every P1 back is calibrated at the factory, the "factory" (aka "canned" or "bespoke") profile can apply with extreme accuracy to any combination of P1 back and DT Photon. The bespoke color profiles in Capture One CH work better out of the box than any profile I've seen a user custom make, including some extremely color savvy clients, and they lack the fragility that custom-made profiles typically exhibit. It just works.

One of the initially-hidden problems with using general-purpose light brands for hyper-critical color work is that those brands frequently change their models, with each new model having its own unique spectral output. Even worse sometimes they make "silent changes" to their product (same SKU, same model name/number, but with changes to the LEDs themselves or to the drivers) that change the spectral output. So even if you use LEDs from those brands that are known for "high CRI" (which is a pretty poor way to gauge spectral quality at the high-end) you cannot count on a given custom profile working on two of the same make and model.

Of course, in most settings none of this matters; we are talking here about the kind of color accuracy called for by the reproduction of rare and valuable cultural heritage materials. It doesn't apply to 99% of the readers here.
 
Top