Re: Rendering Intent for 3D plot in ColorSync Utility
Re: Rendering Intent for 3D plot in ColorSync Utility
- Subject: Re: Rendering Intent for 3D plot in ColorSync Utility
- From: Marco Ugolini <email@hidden>
- Date: Wed, 24 Mar 2010 19:54:47 -0700
- Thread-topic: Rendering Intent for 3D plot in ColorSync Utility
In a message dated 3/24/10 1:04 PM, david wollmann wrote:
>>> For example if I look at US Web Coated SWOP v2 and sRGB in ColorSync
>>> Utility one can see that there are many CMYK colors that extend
>>> beyond the sRGB boundary.
>>
>> That's the kind of distortion caused by scaling the white and black
>> point, resulting in a projection that makes the profile look way out of
>> proportion to its actual size.
>
> Well, help me understand this, doesn't Perceptual always scale the end
> points when used as the rendering intent when doing a conversion from
> one color space to another?
Perceptual scales both white and black points of the source to the white and
black points of the destination, respectively.
But that is not the same as what the ColorSync Utility is graphing, since it
scales the white and black point of the profile to the white and black
points *of the CIELAB color space* (in which the 3D projection is shown) --
which, obviously, produces a wildly inaccurate representation of the
device's ACTUAL gamut, since the white and black points in THAT Lab
environment are placed at the very limits of human perceptibility.
> Let me use this example, I use a RIP to print to an Epson wide format
> printer. I have the RIP set to use Perceptual Rendering from my
> incoming RGB to the target printer profile. The work being printed is
> photography and art, watercolors and illustration, no press proofing.
> Perceptual gives me the best output as the RIP does not support
> Relative with Black Point Compensation (BPC).
Well, not necessarily. It all depends on whether or not a specific image
actually contains colors that would be clipped upon output to your device
using the Relative Colorimetric intent. If it doesn't, then you'd probably
produce a more vivid print using RelCol, without clipping any detail in the
more highly saturated areas of the image.
As an example, an image may be in ProPhoto RGB, and still NOT contain any
colors in it that go beyond the gamut of your inkjet printer. In other
words, just because an image is in a profile whose gamut is larger than that
of your output device, it doesn't necessarily mean that said image MUST be
printed using a Perceptual intent.
That holds truer and truer as the gamut of inkjet printers keeps expanding,
affording the user more latitude in the use of rendering intent.
One must always use a grain of salt in deciding image-specific strategies.
This is where tools like ColorThink Pro are very useful in assessing an
image versus your printer's capability to reproduce it optimally.
> Relative only scales the black point if you have BPC and Absolute does
> not have BPC but it does map or print the paper white point?
Absolute does not have BPC because it ALWAYS maps both the white and black
points in the source to their exact colorimetric equivalents in the
destination. Of course, if the destination's black point is LIGHTER than the
black point in the source, clipping of shadow detail is likely to occur.
The Absolute intent is used to create color proofs. Therefore, by
definition, in order to make a valid proof, the inkjet device/paper
combination used for proofing MUST be able to produce blacks whose darkness
SURPASSES that of the blacks in the source (so that one can effectively
print blacks that look LIGHTER), and whites that are at least as bright as,
or brighter than, those in the source (so that one can successfully simulate
substrates that appear DARKER).
> So for this scenario the Perceptual 3D plot would apply, yes?
Yes, for art prints, but not for color proofing. Also, RelCol could work
quite well for art prints too, on a per-image basis. There is no one answer
that covers all printing instances.
> If I was doing press proofing then Perceptual is completely useless,
> because I would want Relative and Absolute, as a better comparison. Is
> this correct?
Only Absolute for color proofing, not RelCol. Plus those caveats that I
indicated in my earlier paragraph.
Marco Ugolini
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Colorsync-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden