Conceptual doubt.

I have the following situation where I need clarification to understand.

I converted an RGB image to cmyk and assigned the Blurb Color profile with a perceptual rendering intent.

When I open the both the image and the Blurb ICC profile in ColorThink, the image has a lot of points outside the Blurb ICC color space. It would be my assumption that once I convert an image and assign profile all point would be inside the space of the color profile? Is this not the case?


Hi Martin,

When you convert an image to a profile, all the points in the image should be within the gamut of that profile. So your assumption is correct.

Here are a couple of ideas on why you might not be seeing this with your setup:

  • There are a few different ways to convert an image to cmyk in Photoshop, so depending on the exact method you used, you might have gotten different results. Did you use the “Image > Mode > cmyk color” method, or the “Edit > Convert to profile” method?

  • In the ColorThink Grapher, you might be seeing a difference due to one object being displayed using absolute colorimetric and the other using relative colorimetric - especially if your profile has a white point other than D50.

Thanks for your response.

-I used the “Edit > Convert to profile” method.

-My monitor profile it’s at D65. Yet in this particular case I evaluating images using other profiles, which where the Blurb ICC profile which I assume ti’s also at D65?

-Perhaps you can also clarify to me, is there any difference in results in doing convert to profile, versus assign profile?


Just thought I’d answer your query about the difference between converting to & assigning a profile.

There is definitely a huge & extremely important difference between the two operations. When converting, the actual pixel values (RGB, CMYK, etc) are altered so that the visual appearance of the image does not change too much. So if you have an image in anyRGB and convert to a printer profile, ideally the image won’t change its appearance much (lets say the image is completely within the gamut of the printer profile you’re converting to, and you convert using relative colorimetric). If the image had a pixel with RGB values of 100r146g69b, it might end up being 88r179g54b (for example) once converted to the printer profile.

When a profile is assigned to an image, what you are actually doing is changing how you define your color space. An easy way to picture this is to think of the difference between Adobe RGB and ProPhoto RGB - the RGB primaries in ProPhoto are much more saturated than those of Adobe RGB. But, here’s the difference from converting, the pixel values themselves are not altered whatsoever when assigning. No RGB values are changed. It is just the actual definition of what shade/tone/chroma/hue you want for your red, green & blue primaries, which is why if you go to assign a profile is Photoshop & scroll through the list you will find your image getting some wild color changes when you get to some of the rarely used profiles, like some of the scanner or camera profiles.

An easy way to remember the difference is that assigning a profile keeps the (RGB/CMYK/etc) pixel values but changes the image’s appearance. Converting to a profile keeps the image’s appearance (as much as is possible) but changes the pixel values.

Hope that’s not too confusing. :stuck_out_tongue:

I could say that I follow. I had to read it a few times. Thanks a lot. Appreciate it.