FAQs and Tips

(Go back to main Profiles page.)

August 16, 2006 (Revised October 22, 2018)

  1) System Requirements?

  2) Installation?

  3) How does one choose which complete sets to use?

  4) See two-dimensional gamut plots here
  5) How do you use the RGB spaces and their Chroma Variants?

  6) Is it OK to convert directly into the Chroma Variants instead of into the master space? No.
  7) Is there a way to apply the effect of a Chroma Variant to just one part of an image? Yes.
  8) Can you use the variants to push the chroma by more than 100%? Yes.
  9) Any special things to avoid?
10) Check for clipping
11) In Photoshop, to see what you can't see,...

12) It's all simulations in Photoshop

13) In Photoshop, to avoid the default soft proof setup...

14) What is quantization error?

15) How do these spaces fit my camera's gamut?
16) Questions and Tech Support

 

1) System Requirements? Mac OS X, Windows 98 or later, Adobe Photoshop 6 or later, or any other application which can both convert from profile to profile and assign profiles to images and which supports color management simulations to show you correct color on screen. If you don't have proper color management and a calibrated and profiled monitor, you can't get to square one with digital imaging.

2) Installation: See Installing Profiles

3) How do you choose which complete sets to use?

For working with new captures of a wide range of subjects with any digital camera, the ideal setup is to have the DCam 3, DCam 2 and DCam 1 sets. For a more economical solution, the DCam 3 set is the best all around tool. I use DCam 2 the most, as most of my images fit comfortably into it but not into DCam 1. Some require DCam 3's larger gamut and some fit nicely into DCam 1. So DCam 3 first, then consider adding DCam 2 then DCam 1.

If you already have work in ProPhoto RGB and/or Adobe RGB (1998) which is in progress or finished, particularly if the file has adjustment layers, that work will be best served by the Chroma Variant sets built for those two spaces.

If you are making scans of transparency film, the Chrome Space 100 set is the clear choice. For scanning color negatives, Chrome Space 100 will do well, but the DCam 3 set is optimal.

4) See 2D gamut plots here (and a few QuickTime 3D movies)

5) How do you use the RGB spaces and their Chroma Variants?

  1.  Convert into the master space before you begin editing the image.
  2. Assign any chroma variant for that master space until you like the effect and as often as you like.
  3. When you're done editing the image, convert into your output profile, usually a printer profile.

6) Is it OK to convert into the Chroma Variants directly instead of into the master space? Not recommended, even though in some cases it appears to work the same as assigning. Photoshop treats a conversion into some of the chroma variants as though it were an assign, and assignment is what the control is all about, so just assign it to be sure everything works right. Convert into the master space, then assign the variants. Converting from the variants is entirely fine, and also entirely necessary in order for them to function.

Converting from one profile to another keeps the colors the same, while changing all the numbers in the file. Assigning a new profile to an image does the opposite.

7) Is there a way to apply the effect of a Chroma Variant to just one part of an image? Yes. You will have to copy the main image layer into a new document and assign a variant that has the color shift you want, relative to the master profile. Then convert the image from the variant to the master profile, then reinsert the image as a layer right on top of the main image layer, keeping the file numbers constant (not allowing Photoshop to convert from the source profile to the main image profile when you paste). Then use the mask of the upper image layer to control where the image has the local adjustment. For example, fill the layer mask with black and then paint white into the mask with the paintbrush tool and a large, very soft brush to make any region of the upper image layer visible. I have never needed to do this, but it's possible, should you want it.

8) Can you use the variants to push the chroma by more than 100%? Yes. You basically do the same thing as #7) above. Say you've got one image layer in your multi-layer PSD of TIFF file with adjustment layers, and the image was originally converted into DCam 2. It's a very low color image perhaps, but for whatever reason, adding 99% chroma, but assigning the strongest variant in the set isn't enough (this is highly unusual but it can happen). What to do, to go further without using the much more damaging Hue/Sat layer tool (fine for small adjustments but usually that's about it)? It's pretty easy: say you want a total of about 120% chroma boost. Do it in two passes on a duplicate layer. If the boost needed is from 100% (the original is at 100% and you want to send it to ~220%). Just duplicate the image layer and paste it into a new document with DCam 2 assigned. Then assign the DCam 2 +12% variant. Next convert from the +12 variant back to the master space, but watch for clipping, as it's possible. Now the image is in DCam 2 but it's got 12% more chroma. Next copy it back to the main file, above the original image layer without a conversion (if the main file isn't in the master space). With the +99 profile assigned, the top image layer will now be at a little over 220% of normal chroma, or just over +120%. And so on.

9) Any special things to avoid? There is one thing to watch out for: Some plug-ins operate incorrectly on image layers when they have a chroma variant assigned. If a positive variant is assigned, making the image more colorful, and you run such a plug-in on the image, it will wind up with a double hit of the chroma boost, basically. These plug-ins that I've noticed seem to improperly convert the image from the variant to the master space, in effect, as part of their processing. To avoid this unwanted extra boost in chroma, briefly assign the master space before you run the plug-in, then re-assign the variant of choice afterwards.

10) Check for clipping in Photoshop by opening the Levels tool or a Levels adjustment layer and holding down the Option/Alt key while moving the end sliders of the Input Levels part of the tool. The shadow slider should make the whole image turn white, except for any values that are already as low or lower than the value of the current slider position. The highlight slider should make the whole image turn black, again except for any values that are already as high or higher than the value of the current slider position. The clipped areas are color coded so you can tell which channel(s) are clipping. If you zoom to 100%, you get absolutely precise, pixel for pixel (non-averaged) warnings.

It is particularly important that this clipping warning, unlike some, shows you clipping in any individual channel.

To check for clipping in Adobe Camera Raw 3, keep the Shadows and Highlights checkboxes checked and look for the colored warning for 255s or black that will be happening as the file is mapped into the currently selected RGB working space. This great feature should be more widely used in Photoshop and elsewhere, because checking for inadvertent clipping is a necessary task as many points in any quality imaging process. Unfortunately though, the shadow warning is designed to only show you pixels where all three channels have gone to zero!

11) In Photoshop, to see what you can't see right, because the source colors are outside of your monitor's gamut: set up a soft proof of your own monitor, and then enable the Gamut Warning feature! Colors from image values which have been altered to make them fit into your screen's gamut will be marked as out of gamut. Go to View > Proof Setup > Custom and select your current monitor profile as the "Device to Simulate". Click OK, then enable the Gamut Warning option in the View menu. This can only trigger a warning if your source profile for an image has a bigger gamut than your monitor profile does.

12) It's all simulations in Photoshop: To avoid terminal confusion about what you see in Photoshop, it's vital to understand that what you are seeing, whenever you view an image in Photoshop, is one of two kinds of simulations (as I tried to explain in #7 above), either:

A) What I call a source simulation, where the image's numerical values are being converted to new values for screen display only, using either the embedded source profile (or if none is embedded using the current default RGB profile which is set in the Color Settings dialog), and your monitor profile, so that colors will ostensibly appear correctly on screen. The exception is when the current source profile is your monitor profile, in which case there is no simulation when a soft proof is not enabled.

Or

B) A soft proof, i.e. Proof Colors, is enabled in the View Menu (Command-Y Mac/Control-Y Windows). This causes a fancier conversion, from the source profile to the selected output profile (usually a printer profile) and then back to the monitor profile. This let's you see how image colors will be scrunched to fit into printer space. It also lets you enable the Gamut Warning (Shift-Command-Y Mac/Shift-Control-Y Windows) which will overlay a color of your choosing onto all source colors that are outside the gamut of the destination profile. This is a very useful learning tool. For it to be of value when editing an image, you must set up the correct simulation in View > Proof Setup > Custom...

13) In Photoshop, to avoid the default soft proof setup being "Working CMYK..." for every image you open, enter the setup you like in View > Proof Setup > Custom, and then hold down the Option/Alt key (the typical hiding place in Photoshop for important, unlabeled functions) and click the "—> Default" button.

14) What is quantization error anyway? When image colors exist or are mapped into any RGB space, the space functions like a giant 3D grid of finite, tiny locations where a color can be placed. An image in 8-bit per channel form can sit in a 3D array of 16.7 million locations, or 256 planes in each of the three dimensions of the cube. Colors can be at the precise locations in the grid, but not in-between. If the RGB space is defined so as to cover a wider range of colors, the spaces between the steps increase proportionally. Since most images cover only a few percent of the volume of the RGB space, an image may exist in just a couple of dozen or fewer planes in vertical orientations. If such a space is enlarged to twice the radius, the number of planes of image info will be cut in half in each lateral dimension. The number of vertical steps (horizontal planes) doesn't change as RGB spaces get wider or narrower, rather the preservation of vertical steps is best optimized by keeping the tone curves of spaces being used as the source and destination for conversions similar. Preserving the vertical steps during image processing is more critical to image appearance than preserving horizontal steps, to a point. When one looks at a precise 3D model of the colors present in an image, as thousands of points hanging in space, one can see what look like the rows in an orchard viewed from a passing car. The gaps between the trees run out in several directions, but in 3D instead of 2D. When comparing two converted files, one into a small DCam profile and one into a larger DCam profile, the difference is that the gaps between all the rows have grown wider, and the number of locations where originally distinct colors can be placed is no longer enough to keep all of the colors distinct, or to keep differences between colors as subtle.

15) George K. writes: "Do these three sets [DCam1, DCam 2 and DCam 3] fully cover a medium-format camera, such as the Pentax 645Z?"  Here is my reply:

I use a 645Z and rely on these three sets for everything. The thing is though that it's not the gamut of the camera that's the issue, it's the gamut of the subject matter, as processed by the camera profile and the raw conversion image rendering controls, and only rarely as limited by the gamut of the camera.

Overwhelmingly, the subject itself plus the way the tonality and saturation of the subject are adjusted in both the exposure and in the raw conversion app is what determines how much gamut is needed in the working space to contain the colors that you're trying to store in it without clipping.

For example, one could photograph a subject with almost no color in it, resulting in a gamut like a super skinny football (American type football, not a soccer ball, i.e. pointy at the ends and fatter in the middle), and have that gamut very easily fit within one of the smaller/narrower working spaces, and yet the ends of that gamut can be almost as prone to clipping as a wider gamut just by the image being rendered as too light or too dark, or by the white balance being set so as to turn a very light gray into a very light blue-gray, for example. A wider gamut gives you a little more headroom the more so the further out you get away from the gray axis.

Picture the roof and the floor of any given RGB working space as being like the top of a circus tent or the same inverted. You can hit the cloth with your tall pole or other shape either by being too tall or by getting more off to the side (more colorful) or both.  A wider tent gives you a little more room to work when you're up near the top (white), but it's still easy to hit the canvas.

The ultimate gamut limitations of the monitors we have and may have in the future, as well as the printing systems we have or may have, do play a role in the sense that at some point, it becomes just senseless to push colors outward to regions of color space where, for example, by adding either chroma or saturation to the image, or by adding too much contrast, or by over- or underexposing in the first place, because all that's accomplished is that colors get clipped badly on the displays and colors must be mapped strongly back toward gray by the tables inside of a printer profile (in the case of being too colorful as opposed to being too light or too dark).

If a printer profile isn't perfectly behaved, this inward mapping can come with a price, e.g. apparent hue shifting used to be a common problem when someone used Velvia to capture blue scenes e.g. sky reflecting in water, and have a too-blue white point of the scene for the native white point of the film (typically 5300K for daylight transparency film) and then also no UV filtration also making the film bluer, as well as the film itself boosting the chroma by exaggerating contrast of the lower middletones tremendously, and sometimes on top of that the photographer would boost the color even more during raw conversion and wind up with some of the colors in the image being so far out away from gray that they were way beyond what even the monitor could display or even what was an actual, real color, and thus need to be mapped way, way in to reach the printer gamut. (Printer blue gamuts are way smaller than monitor blue gamuts, even today, with much improved printer blues.) Inward gamut mapping by printer profile tables also often causes undesirable shifts in lightness, which get worse when the colors have to be mapped a longer distance inward.

It's hard to explain without 3D illustrations of colors as dots in space but that's why there's really no point to using a larger space than DCam 3 to start with for almost any picture unless it's for some special purpose, e.g. to help assure that a bulk process of many images results in the fewest clipped colors. In normal working, I would recommend respecting the boundaries of whichever space you pick (the choice of space being informed by how much color is in the image at the bright and dark ends usually) and taking care to not clip the colors upon conversion into the RGB working space from the camera profile (during raw conversion), by being aware of and looking for clipping and making adjustments with the raw converter's tools accordingly, using the Levels, Exposure, Curves, Saturation and/or Chroma controls.

I have yet to see any issue with one camera's gamut being bigger than another's and that leading to an issue where a picture of a given subject required a larger working space on account of that. I'm thinking of the blue-green LEDs used in traffic lights to mean Green — Go. They are gamut busters, as are other LEDs. Also, for the record certain cyan colors are the colors most likely to be within the gamut of recent inkjet printers which we are apt to never capture in an image (so it's OK for DCam 3 not to include some printable cyans). I've surveyed almost all of the colors in the world (I never examined tropical fish, which might be gamut busters in very clear water, i.e. up close, in really good light) and I used the actual colors of the world to inform the design of the color spaces and I still think it was the right thing to do. Laser beams and other relatively monochromatic, ultra-saturated colors are the kind of thing you have to start trying to render to encounter limitations of DCam 3. For such cases, DCam 5 is best, but then you still won't be able to see those actual colors on any display or in any kind of print, rather you'll only see a clipped approximation of them.

So, yes, they're great for the wonderful 645Z (what a beautiful sensor!), and I think for every digital camera.

16) Questions and Tech Support: In theory they're all answered here somewhere, but if you want to ask me about my spaces or chroma variants send me a note and I'll get back to you when I can. If you are confused about how color management works (as everyone is while learning it) it is not my responsibility to unconfuse you about it! Please keep that in mind. Profile sets do not come with a free education in color management, beyond what you can glean here. For that, you may want to arrange for some one-on-one tutoring here, if I have time (see my workshops page for more info).