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!

Capture One or LR6?

ErikKaffehr

Well-known member
Hi Wayne,

I would say that is exactly what "perceptional rendering intent" is expected to do. Relative colorimetric is expected to leave colours unaffected.

I think there can be a disconnect between rendition of what we can see on screen and in print, but soft proofing should display a decent estimate of the print, although it cannot show colours within the printer gamut but outside screen gamut.

Best regards
Erik




The entire concept of color management is allowing the device profiles to manipulate the colors into the visible output space, so even though we can't "see" a color, the way it is rendered allows us to see the colors as they relate to each other, which is more important in the scheme of human vision that the actual color that might be there scientifically.
 

ErikKaffehr

Well-known member
Purple rendition, beating a dead horse?

Hi,

It may be beating a dead horse, but we have some very knowledgeable gentlemen here so it may be an interesting discussion. Check the screen dump below:

Screen Shot 2016-07-28 at 07.50.12.jpg

The crops show (from left to right)

Sony Alpha 99, C1, WB on third grey patch from the left, C1 "standard" profile
P45+, according to Jack's advice, using FLASH profile and FLASH WB
Sony Alpha 99, LR6, DNG Profile Editor dual illuminant, WB on third grey patch
P45+, LR6, DNG Profile Editor dual illuminant, WB on third grey patch
Sony Alpha 99, C1, WB on third grey patch from the left, C1 "generic" profile

The measured colours are here:


What I think Jack has demonstrated on the previous image is that the profile is very sensitive to changes in white balance, check the screen dump below. From left, WB = FLASH, centre third grey field from the left on a ColorChecker Passport, increasing "Kelvins" until the petals are clearly purple. Increasing "Kelvins" makes greens yellows. Also transitions on the petals are not clean, it sort of flops between purple and blue.
Screen Shot 2016-07-28 at 08.07.54.jpg

So it seems that Phase One has a decent quality profile for the Sony Alpha (Standard) and a bad one (Generic). The Generic profile for the Sony Alpha has similar issues to the P45+ profile.

Now, this is clearly an exception, unless you are shooting deep bluish purple flowers.

For me, the conclusion is that I can get a good quality profile using a simple to use tool from Adobe (the Adobe DNG Profile Editor). Or, I can use a tool that is well documented, flexible but needs some more work. But, I am an engineer and not an artist, so I am expected to be able to read manuals and use some command line stuff :)

The raw images are here:
https://www.dropbox.com/s/dy76rntvcoq94w0/20150107-_DSC6397.arw?dl=0
https://www.dropbox.com/s/3jd3aa9fj3n143e/20150107-CF046070.iiq?dl=0

The spectral data for the petals is here:
https://www.dropbox.com/s/up2gfxfveyrd1qj/Violet.txt?dl=0

Best regards
Erik
 
M

mjr

Guest
Morning

I am sure we all see colours in different ways and we obviously prefer different things, I also have no doubt that some people will not like the look from C1 as some won't like the look from LR, it is subjective after all. I stumbled across this shot from years ago, taken in bright sunlight, opened in C1 and left at default settings, white balance as shot, profile being P25+ daylight and I can't really see anything wrong with it, colour separation is pretty good to my eye.

The biggest issue for me personally with this thread is that now I want a P25+ again!

 

torger

Active member
The eye is very good at adapting to various conditions, so when you have no reference to compare to you need to have pretty large color discrepancies before it looks like colors have been adjusted by the profile or the photographer/artist. When several renders are A/B compared (preferably layered on top) the differences stand out, and when you have the real scene in fresh memory it's also easier to spot the subjective adjustments.

And some just have a color memory and color vision that is better than others, and I've come across a bunch of those when working with DCamProf. I think I have myself above average color vision and score fine in tests, but wow some are just incredible. It's the same as with tasting foods, where I'm the opposite average or below average. I have a friend of mine that is a "super taster" and can break down tastes in lots of different names, and I just say "ahhh salty, I like this, can I have some more?". The drawback of being sensitive is that you get more picky though.

Being picky about color does not necessarily mean that you desire realism, but that you're more likely to not just accept any look put on the table. And even if you do like realism (it really is a good starting point for post-processing) there is a quite wide span of subjectivity to interpret that too. The subjective looks provided by manufacturers are varying a lot, just compare Leaf and Phase One for example. If you are a color "super taster" it's almost certain that you can by yourself design your own profile/look that suits you better than what the manufacturer gives to you. This also means that if you happen to like the price of a Leaf better than of a Phase, but you're not so into Leaf's profiles you can still get the Leaf and just design your own look, which I think is pretty neat. Unfortunately for super tasters the software support to do this is currently weak. There more I talk about this the more I feel I must do something about that ;)
 

torger

Active member
Regarding profiles' sensitivity to white balance; all non-linear corrections is made relative to neutral, so if the neutral is moved around the non-linear corrections move accordingly, so it's important that the chosen white balance represents a neutral color. If you want a creative cast it's better not applied with white balance but some other tools.

However most well-made profiles are not that non-linear and thus not too sensitive when white balance is shifted, so generally it's overkill to not make creative casts with the white balance. Strong non-linearities result in poor gradients, as demonstrated by Erik where the flower flaps between purple and blue.

This is one of the presumably few cases when Capture One actually contains a badly designed profile as it 1) is too non-linear and 2) even if white balance is correct the corrections move purple to blue. But the thing is when you have released a profile you can't really fix it, as people have already started to use it in their images and don't want rendering to change when getting an upgrade, the only you can do is to add more profiles.

(That you can't make a profile too non-linear to avoid broken gradients means that hardware color response will shine through to a larger extent than if you make a full-corrected profile.)
 

Paul2660

Well-known member
I believe C1 also keeps all the data in a "large gamut" space which is probably a derivative of ppRGB like LR. The recipes control the rendering into the output file. But I've new

I've never really had a problem letting the color management system map the full gamut ppRGB 16bit file into the Epson output color space. I also feel the gamut of the Epson printers in the blue regions exceed the gamut of AdobeRGB, so forcing the output tiff into AdobeRGB clips some of the available color to the printer.

The entire concept of color management is allowing the device profiles to manipulate the colors into the visible output space, so even though we can't "see" a color, the way it is rendered allows us to see the colors as they relate to each other, which is more important in the scheme of human vision that the actual color that might be there scientifically.

Hi Wayne,

It's obviously a personal preference. I tend to use a bit of saturation in my work, (which will very quickly tend to push the blues out of gamut for a Epson 9900 even a glossy profile. Sure the Glossy profile will hold a lot more of the blues within gamut. However for one image, I may print it on matte canvas (least possible amount of gamut), or glossy canvas or matte paper etc. Same image depending on the paper profile will hold less or more of the blue gamut, but most times I will still see a bit of loss in the blues. Sometimes the loss of gamut is not noticeable other times it can be very noticeable.

Profiles, I have my own, generated in iOne or profiles generated by the paper companies.

Each person should see what works for themselves. I use Chromix software to view profiles, and as I recall even a Premium glossy PP profile by Epson or myself can't full hold to the full gamut of the Adobe 1998 color space ( for the 9900), they can get close and the newer generation of Epson's may get there. But I won't be upgrading to the 9000 or any new Epson, instead I have to make the 9900 work. The best case for me is the point that Jack mentioned, soft proofing. For my work I consistently see colors out of gamut but my image workflow more than likely takes the files there. Prophoto will usually show more colors out of gamut for me. And attempting to fix the gamut issue tend to take more time than it's worth for me.

I have recently started using Prophoto again, for the IQ100 files and so far no problems.

Paul C
 

Jack

Sr. Administrator
Staff member
Paul,

You need to look at a gamut map again. Wayne is correct that Adobe RGB won't hold as much in the high Blues as the Epson X900 ink --- as an easy example, compare maps of Adobe RGB to Epson Premium Luster 250 X900...

Regardless, I agree that it's personal choice, and especially if you prefer heavier saturations, then OOG colors can start to look bad or become distracting in the way they saturate. Just curious though, given your preference for saturated output, is there a reason you don't edit directly in your proof (output) space?

~~~

Re rendering intents, remember that simply speaking:

Perceptual compresses colors outside the gamut to look as normal as possible inside the target gamut;

Relative compresses colors outside the gamut to the last available and nearest color inside the gamut (thus can create outer limit saturation issues);

Absolute trims out any colors outside the gamut using only colors that map directly into the target gamut;

BPC on attempts to keep the gray line linear inside the target gamut during any compressions;

BPC off allows gray to shift off axis relatively with any compressions that occur (allowing some grays to hue shift to non-gray).
 

Paul2660

Well-known member
Hi Jack,

I will have to look at the maps again, as it's be a while.

I used to edit in the proof space all the time, when I did most most the heavy lifting in CC, now most of that is done in C1 or LR. LR as far I know only allows you to softproof, i.e. virtual copy etc. I tend to use a lot of adjustment brushes and adjustment layers in C1. LR will bog down enough that I just quit doing it. I have never tried editing in the proof space in C1, will have to look at that in the future.



Paul C
 

Jack

Sr. Administrator
Staff member
Paul,

To be clear here is my basic workflow -- and I realize it may seem a bit of a cluge to many, but it works very well for me:

I do my initial edit in C1 using 16-bit PP. This becomes my working tiff. While I do use the layer tools in C1, I still find some of them less satisfying than comparable layers in CS, so still use CS for final edit -- though admittedly, I am doing more and more in C1 and less and less in CS... Anyway, I output the 16-bit PP tiff from C1 and do the heavy lifting layer work in CS, still in 16-bit PP. When complete, I have a full layered base file as my working file and name it XYZ_working. Now based on my desired output, I will final edit the file for the output space inside CS, and save it with an appropriate name like XYZ_print_20x24_EFA. FWIW I also still print from CS simply because I like the workflow. (Virtually all of my printing is done one-at-a-time, and rarely do I need to batch print.)
 

Bill Caulfeild-Browne

Well-known member
I too prefer to print from PS, but like you Jack, I find I'm doing more and more post in C1. It's come a long way since C17 where so many of my past files were processed. I invariably re-process them in C19 now if I'm going to print them again but still use the Canon 16 bit module in CS
 

Wayne Fox

Workshop Member
Paul,

Perceptual compresses colors outside the gamut to look as normal as possible inside the target gamut;
and to accomplish that perceptual has to move colors that are within gamut, to make room for the out of gamut colors. To me perceptual is mostly about trying to maintain the relationship between tones and colors.

personally I've found that either intent satisfies me 70%-80% of the time with not much visible difference, where as about 20% of the time one does appear better when printed, and about it seems about 10% of the time the difference is somewhat obvious and dramatic.
 

GrahamWelland

Subscriber & Workshop Member
What Jack and Wayne said. I process in prophotoRGB within C1 and I absolutely agree with Wayne that ultimately it's the perception of the relative colours to reality vs the absolute technical RGB, CMYK, LAB values. I have no problem at all with editing with a fully colour managed workflow, display, and print pipeline and then ultimately tweaking my output file to get the print to render EXACTLY as I want it to as opposed to be colormetrically accurate.

I've always wondered about the profiles that the various manufacturers and image processing software vendors use vs the variability of our own eyesight and perception. For example, perhaps I'm different than others in that I see the world of natural green with a more blue vs yellow rendering so that I've typically found 'Nikon/Leaf' greens closer to my view of the world vs the more yellow 'Canon' or out of box Phase One renderings.
 

torger

Active member
There are some variabilities in color perception, but as it's about color matching, eg same color on screen as in print, it doesn't really matter if we see a color as coral or pink as we can still match colors.

A camera rendering is more complex though as it needs to compensate for the medium by increasing contrast which affects color perception, so a realistic render, if you desire that, can't be colorimetric. It's still about color matching though. Look at the real flower, then the flower on screen or print and see that they match in color.

If I get to choose if the default render matches colors as well as possible or if it doesn't, I choose the former as I see no point in that my personal look should be a modulation of some other's look rather than a modulation of the real thing. If the manufacturer's subjective look is reasonable it's not a big thing though, but I can't see any drawback of a realistic render.

The greatest advantage of making own camera profiles is not that though, but it is that camera hardware and raw converter choice gets more flexible. Like Phase One's digital back user interface but not its yellowish default render? No problem, your own profile your own look (realistic or not, your choice). Like Lightroom's functions but not the look from the bundled profiles? No problem, make your own profile. You shoot with several cameras and want all images to have a consistent look? No problem, make custom profiles for them all.
 

tjv

Active member
I had a bit of a play with some images today that I'd previously converted in C1 with the Leaf ProPhoto camera profile. I found that the Leaf Product HS5 profile gave me great results too, and the greens seemed better in that they were less intense / nuclear. I think part of my problem has been that I'm learning many new things at once, or at least extending knowledge in many areas at once. The ProPhoto camera profile is more saturated by default it seems, so in a sense it's easier to see the effect of small changes as the colour shifts are magnified. Practice makes perfect, I guess. I'm so used to editing film scans with a set workflow I've forgotten the versatility of good digital captures and processing. And C1 is certainly more advanced than LR...
 
Top