Hmm…I’m pretty sure Joe didn’t say that. It doesn’t even make sense. What does it mean to say printer inks are too saturated for Dcam 3 profiles? If you mean - “UCK3 inks on some glossy media can attain a colour gamut wider than Dcam 3”, that would make sense.
That makes no sense to me also. Assigning a chroma variant in Photoshop modifies the colourfulness of the image uniformly. This is based on aesthetic decisions. On the other hand, to “handle the cheeky Epson INK K3 upstart” requires good gamut mapping from your source profile (document space) to destination profile (printer space). Assigning chroma variants won’t help you there.
The Dcam space’s TRC was designed with the intent of being “more perceptually linear” (than gamma-based TRCs or even L*) in Joe’s opinion, not with the primary intent of matching the printer’s tone curve. In fact, different printers have different tone response so it would be impossible to try and match it with a single TRC. A well-behaved printer aspires towards perceptually linear output, so ideally its tone curve is a reasonably close match to the one in Joe’s profiles. When Epson improved the driver tables for the Pro Stylus models back in the day, it happily brought their native tone response much closer to Joe’s proprietary tone curve and thus this design had this as an added bonus. Having the curves match closely is a big deal to minimise quantisation if you are working with 8 bit images and 8 bit conversions. It is not longer an issue with high bit processing that is recommended today. That’s why I mentioned earlier that having this special TRC is not so relevant today anymore.
Anyway, @Keentolearn did you manage to figure out how to add your own “working profile”? Guess what, in a different thread I was dealing with a different problem which required me to also do this, which I have never done before, and thanks to your asking here, I was able to successfully do this after some mild struggle.
I tried the tidying up of the DCam profiles so that the filenames matched in MacOs/ pictures and in Json…(copy and paste)
I’m not there yet but right now my situation is to stay backstage on what is merely a technical issue and soak up the tremendous flow of knowledge that is exploding around me…
You have proved that Custom profiles work in RT and thats good enough for me to continue until I find out what the problem is.
One point …the Json file never becomes a Colorsync controlled profile? …does RT become the application to open the Json profile …if its not in the Colorsync stable …Right? ps. I’ve always used Mac systems and rarely use Terminal in Mac unless I’m forced to…
Hmm…I’m pretty sure Joe didn’t say that. It doesn’t even make sense. What does it mean to say printer inks are too saturated for Dcam 3 profiles? If you mean - “UCK3 inks on some glossy media can attain a colour gamut wider than Dcam 3”, that would make sense.
i was’nt specific enough…
what Holmes actually said and I quote from his website " when a chroma variant set is used to control image color, the issue of the working space not being large enough to encompass all the colors that every printer can print ( as for example Chrome Space 100 and DCam 3 do not encompass the entire gamut possible with EPSON’s new UltraChrome K3 inkset ) largely vanishes because the space expands as you push colors outwards to the point where you could print those out-of-working space gamut colors (which did’nt exist in the image to begin with) anyway "
My comment :I suppose alternatively, we could skip across to DCam 4/5 or Prophoto in 16 bit…Holmes has variants too for Pro Photo and no Im not his agent!
It’s all on his website and much better described than I can manage.
One other thought… Holmes suggests that we think about what colors we want to photograph first and then think about the smoothest way to get those colors down on paper.
I also agree with you about an earlier comment you made about having respect for the input process first …not the output printed product being the most important issue…I like landscape color and I like K3 inks for printing those colors and I also keep the master file intact waiting for the post digital era. optimism
That is, a color space designed to process the image with it.
Which is, indeed a «working profile» designed to process the image with it. Do you see the analogy? The difference is that RT works in linear space, that is, without TRCs
I’m sorry, but when I was a Photoshop user, I understood by the many posts, articles and tutorials I read, that a working profile, with its TRC, was meant to edit the image, and it could also be the color space (or profile) of the output image, or not. My very last step right before exporting the image was always to convert the color space to a better suited one (usually to sRGB).
If they are designed to increase saturation of an image by just assigning them, then they do have a look (or an immediate result) that other spaces doesn’t offer.
And by the way, the fact that you’ve already said that doesn’t mean everybody MUST understand or agree with it. You should have the patience to explain it again, preferably with other words or other facts that made it clearer to the people that still has doubts. You’re not necessarily in possession of the truth, and if you are, then explain it to us 3-4 times, or as many times as necessary, until we understand it and agree with you.
Sorry, I don’t see the analogy, what is the contrasting example you are comparing to as an analogy?
Certainly one can insist on the use of Dcam spaces as the “Working Profile” for RawTherapee to process the image with, which will then ignore any TRCs embedded in them. We have no disagreement here at all.
Yes, we do not disagree at all on this also. The difference is Thanatomanic is talking in the context of RawTherapee, not Photoshop. When you talk about working space profile in Photoshop, it is always using the TRC of the working space profile, is it not? But when you talk about the “Working Profile” (in the words of the RT user interface), it ignores the TRC of the profile, even if you set it to your own custom profile which has an embedded TRC, is this not the case?
So Thanatomanic was insisting that the -
How is it complete nonsense in the context of Photoshop, or when the profile is used as the “Output Profile” in the context of RawTherapee? But of course it (meaning the TRC) is “nonsense” when used as the “Working Profile” in RawTherapee since it (TRC) is ignored.
Perhaps now @XavAL can see that we have no disagreement between me and you.
The Dcam spaces are not designed to do that. This is a gross misunderstanding. The Chroma Variants are what is used to increase saturation. Two very different things. The Dcam spaces are like e.g. sRGB, Adobe RGB etc. just normal RGB working spaces.
Let me try then to break it down for you. Say you are working on an image in sRGB. Then you assign Adobe RGB to it in Photoshop and of course the colours change. Can you say that Adobe RGB is a profile that “adds a look”? Can we also say that since Adobe RGB has a TRC of Gamma 2.2, this is complete nonsense and thus cannot be used as a working profile [for RawTherapee]? Of course, this is wrong.
Some folks are confusing the regular Dcam spaces with the special chroma variants of these spaces. To put it another way, the Dcam spaces are what you convert to just like any other regular RGB working space. The Chroma Variants are what you assign to modify the colour.
Joe has made chroma variants of Adobe and Prophoto as well, which modify saturation when you assign them to your images that are in those respective colour spaces. The chroma variants are essentially AdobeRGB/Prophoto/Dcam X spaces with their primaries slightly modified, shifted outwards or inwards if you like, such that the RGB space is then uniformly larger or smaller (in reference to Lab space). As such, the effect of assigning these chroma variants allows one to increase or decrease chroma with exactly the same visual effect as the Lab chromaticity slider in RawTherapee. That’s all there is to them. And you don’t have to use them at all if you don’t want you, you can just use the regular working spaces alone as normal working spaces which certainly do not add any looks and do not modify saturation one iota.
I assume the confusion arouse out of Matt’s writing which was initially very convoluted with lots of misleading/wrong points, which I’ve tried to tease apart as you can see. It seemed only to fuel others to drive the conversation further away from his main issue, and even raised some ire about Joe Holmes’s writing style and pricing strategy which seems misplaced in the context of the thread, along with repeated false accusations about how the profiles work. I had wanted to refrain from fanning the flames even more, since I have no intention to convince anyone to switch to using Dcam spaces. Joe’s articles about them and how they work are there for anyone to read, but continuously insisting on a false statement after someone tells you it is incorrect is not helpful towards learning.
It goes both ways. I cannot force anyone to understand something they don’t want to understand. I’ll leave it at this for the time being, hopefully this clears the air once and for all.
I have never used Photoshop extensively, but after re-installing it, playing with it and reading this and this, I understand things better. @XavAl’s explanation seems correct. Apparently Photoshop intends for you to only pick one color space and assumes you use soft-proofing to ensure your colors will look right on your medium. Then if you don’t show the advanced options on export and pick an different color space there (the output color space), Adobe assumes your working space is your output space. This space will be embedded in your exported file.
So the key difference here is that the choices for working color space and output color space (or profile, whatever you like to call it) is more visible in RawTherapee than in Photoshop.
If it does, that would be very bad for many editing operations. See here for more information: https://ninedegreesbelow.com/photography/test-for-linear-processing.html
Elle’s conclusion is that for Photoshop “Some editing operations and blend modes are automatically linearized, and some are not, and some editing operations are linearized at 32-bit floating point, but not at 16-bit integer precision.”
The key difference here is that in RawTherapee the TRC is ignored so you always operate in linear RGB mode.
Well, agreed on the regular DCam spaces. But Holmes doesn’t advertise his chroma variants any different than his regular DCam spaces. That means he implies they are working profiles too, right? That means a look is applied, which is not what a working profile should do. But since Adobe muddles working and output profiles together, I guess you can say that things are all right in the end?
This is still something I don’t understand. Can you explain what you do in Photoshop that makes these operations uniquely different?
Edit: I must correct myself from earlier statements where I strongly advocate to work in linear RGB space. While this is sometimes true, that is not always the case. Coincidentally, there is a recent nice post here that gives a few examples of operations that you want to do outside linear space. I think that Adobe may have the philosophy “use the TRC from the working profile and work in non-linear space by default, but internally convert to linear when necessary”, and RawTherapee is the opposite “ignore the TRC, work in linear, except in some cases”.
Unfamiliarity with the GUI of an app you don’t typically use does not mean its functionality is any less visible objectively, though subjectively you are right. It is less visible to you.
We need to break this down properly. Firstly, let’s look at RawTherapee. You can only pick one colour space for each of input, working and output. Only one. It is a raw converter and image processor, so the data has to necessarily move through several transforms, but only one colour profile can be used at each stage. Photoshop is no different in this regard.
Photoshop is just a raster image editor that works on non-raw images, raw images have to be processed by its plug-in, Adobe Camera Raw, or some other software and sent to Photoshop in a format it can parse. The image file can have only one profile associated with it at a time.
When you open your document, Photohop honours the embedded profile by default. If there is no embedded profile, you can assign a profile to your document. This is similar to the “input” profile of RawTherapee, though for RT it is usually a camera device specific profile for raw images, while usually in Photoshop you are assigning a device-independent working space profile. Photoshop can even use CMYK profiles and more which RawTherapee can’t.
The document profile is also the “working profile” (using RT speak) in Photoshop. This should be easy to understand.
The user can save the document as is with the current working profile. There is also a function to “Convert to profile” which is similar to the “Output Profile” in RawTherapee. You can choose to convert to any profile and choose the desired rendering intent and save that as another copy of your image or override the current document.
assumes you use soft-proofing to ensure your colors will look right on your medium.
Softproofing your image will not make it somehow look right in your chosen output medium! This is no different in RawTherapee. In softproofing, there is no actual change to the document’s profile as it stands. It is merely an on-the-fly simulation.
You still have to make the deliberate choice of converting to the correct output profile, just as you would do in RawTherapee.
Edit: I see you changed your post on this point as I was making my own. Please read on further below to see I acknowledged it.
I was generalising a bit there. Indeed certain functions in Photoshop can operate in linear mode even if the working space profile is non-linear. Photoshop also doesn’t have true 16 bit support, it is more like 15 bits + one signed integer. Some editing operations are forcibly 8 bit only too. It is too complex a program to describe in few words, but generally speaking, Photoshop honours the TRC of your working document.
That is not true in all situations also. If you set the working space TRC to a custom one, RGB operations in RT will not operate in linear mode. Also, you can choose log for resampling in geometric corrections. There are possibly others which are not coming to mind just now…
Hmm, honestly I have no idea what you really mean by this.
Yes, the chroma variants can be used as working profiles. If you use them as such, they cannot (read: totally impossible) modify saturation whatsoever and will not impose any look whatsoever. On their own, they behave normally just like any well-behaved RGB working space profile.
You must understand that Photoshop is not like RawTherapee where you can make edits, close the program and somehow those parametric edits are saved and associated with the correct document. Photoshop is old fashioned. You open a document, working on it, save it. Certain operations are destructive, after you save the file, and then open it, you can’t go back to a previous state. So this file can only have one color profile associated with it at any given point in time, period. This doesn’t mean “working” and “output” profiles are muddled. Photoshop has what is called the “Document Profile”, which is the current profile that is selected. It is up to the user to change it depending on their needs, whenever.
I recommend you to try this. Start with an image, JPEG or TIFF, which you have saved with the sRGB profile with its standard TRC (not gamma=1.0). Open it in RawTherapee’s editor. In the “Input Profile”, select Adobe RGB Gamma 2.2. Or maybe ProPhoto RGB Gamma 1.8. What happens? Your images colours are shifted right? The look has changed, right? They are more saturated, right? Do you say “a look as been applied”? So can we say Adobe RGB and Prophoto RGB are not doing what a working profile should do? No, of course not. Now why is that? Is it because they are badly designed? No, it is because you are using them the wrong way which is causing you to have this thinking. I am asking you to assign AdobeRGB/Prophoto RGB to an sRGB image, not convert. That is why the “look” changes. It is not due to the profiles but rather how you used them.
The chroma variants work exactly the same way. It is an unconventional way of using profiles to modify saturation, by assigning them in this way. It does not mean that they don’t behave like normal working profiles when used as such. If you instead have converted to the chroma variants, you will see that the image’s look does not change at all. Zero change whatsoever.
There is nothing special about the way Photoshop works with profiles that is unique. It is based on the same colour management principles as RawTherapee. Maybe you need to have a deeper understanding of what it means to assign vs convert to a profile. This is independent of Photoshop and RT. If you don’t understand this well, there is no point in continuing this discussion.
Beautiful! Yes! :)) Linear does not always been better. The main difference is Adobe is working on non-raw images which typically have gamma-encoded profiles (i.e. what you output from RawTherapee or other raw processing software), while RawTherapee typically works on raw images. Hence the observations you have made.
@samuelchia I think I know what the difference between assigning and converting is. I also think my earlier post here explains this well enough. There is no inherent difference between raw images or non-raw images: for each type you must assign the right input profile (either a device space or device-independent space or use a previously embedded space), then convert to the working profile in which you do all your edits and then either keep this working profile upon export or convert to a different output profile. This profile is embedded in the exported file so that other software can again correctly use it upon opening.
There may also be no inherent difference between how Photoshop and RT handle things. Both use the input > working > output paradigm, but from what I see RT is more explicit in its GUI about these options. In RT you choose your output profile directly in the GUI and you are presented with a preview that is already based on this profile. In Photoshop you must use soft-proofing to simulate how your image will look like for a certain output profile and then you can convert to this profile when you export. Maybe different sides of the same coin?
Well @Thanatomanic, I take you at your word. But your continued insistence on this point clearly demonstrates you don’t fully understand what is going on. Or perhaps you are not ready to consider a different perspective that there are different ways to make use of profiles that are not conventional (the ways you have just listed), which may be why you cannot accept how they work and therefore just assume Joe’s profiles to be faulty.
No, you don’t need soft proofing to do this. In the “convert to profile” function, you can see an immediate preview of the effects before you hit the OK button. It sounds to me like you are merely unfamiliar with Photoshop, period. I could just as easily say the GUI of RawTherapee is unintuitive and awful if I am a first time user, doesn’t seem fair yeah?
Well done…not yet…but I’ll get it…Xavier have proved it works and now you too…I’m sorry that I didnt say at the start exactly what my end goal was… spent a lot of time in Rawpedia at first…I’m a newbie as you know now…and in the end it was you guys that are dong the real helping out…and the manual is well assembled I have to admit.
A lightbulb moment for me…“HT does not use TRC in RAW”…Holmes doesn’t either…his TRC is part of the working space in Photoshop itself ( post ACR Raw ) and at this stage he is trying to improve the TRC in the darker tones of the DCam 1/2/3/4/5 working spaces to adjust to and prepare for meeting up with the tone curves of the “better built " printers. Incidentally, he comments that the first “proprietary tone” curve built wasn’t his …the first” proprietary" TRC was in fact developed by the Srgb color space where an effort was made to brighten darker tones .
I totally accept at the same time (as you wisely pointed out) that in a 16 bit situation the TRC matchup between Photoshop working space and printer space becomes a non issue!
Could anyone tell me the difference between a Gamma Curve of say 2.2/1.8 and a Tone Response Curve (TRC) ? …i have tried everywhere to get a definition but with no luck.
In ICC-land, the TRC can be a number of constructs, of which gamma is one. “TRC” refers to whatever tone modification mechanism is included in the profile.
S [quote=“samuelchia, post:67, topic:18022”]
Anyway, @Keentolearn did you manage to figure out how to add your own “working profile”? Guess what, in a different thread I was dealing with a different problem which required me to also do this, which I have never done before, and thanks to your asking here, I was able to successfully do this after some mild struggle.
[/quote]
As a courtesy to all of you kind people who took an interest in my wish to add a custom working space to RT the result is as follows:
i finally got the .json shell working smoothly as workingspaces.json
2 I instructed RT preferences as to where i placed the workingspaces.json file for inclusion as a working space in the Combobox. There was no problem…i even varied the destination from RT’s choice of profile (.json) location ( 334/338 Rawpedia) to creating my own location (eg…Mac OS PICTURES) for workingspaces.json…again no problem…RT preferences teamed up and registered the location alterations seamlessly.
When I double clicked the workingspaces.json file the RT program was immediately and successfully opened up…no matter where on the directory tree i had chosen to locate workingspaces.json
4 we know that custom working spaces can be added to combobox…already proved
5 i however have not been able to place my custom working space in my combobox
6 the question is ;
How can i have a defective workingspaces.json file if it is successfully able to open the RT application ?
Anyway thanks to all…I can now be absolutely certain that I can ,at least and at last, avail of RT’s superb sizing tool and return to Photoshop without having to convert or assign profiles (.icc land) or working spaces at any time(s) between RT and CC.
I also learnt a vast amount from this thread and to you kind people I offer my gratitude and best wishes.
I am having the same trouble (I use Windows 10). I created a workingspaces.json file in notepad and saved it as ANSI text.
My colour profiles are all in the system directory which is what is pointed to in RT’s colour management preferences. The relevant profile appears in RTs list of output profiles, but despite the below json file I can not get it to appear in the list of working profiles. The .icc is in the sysetm folder and not in RT’s list of output profiles so the program does go to the right place to
find it.
The system directory is C:\Windows\System32\spool\drivers\color\ and that is where I have placed workingspaces.json. I even put a copy of the json in C:\Program Files\RawTherapee\iccprofiles\ but that didn’t help either.
to all…I can now be absolutely certain that I can ,at least and at last, avail of RT’s superb sizing tool and return to Photoshop without having to convert or assign profiles (.icc land) or working spaces at any time(s) between RT and CC.