Re: The color of gray
Re: The color of gray
- Subject: Re: The color of gray
- From: Steve Upton <email@hidden>
- Date: Tue, 02 Sep 2014 13:03:10 -0700
On Sep 1, 2014, at 11:07 AM, Ben Goren <email@hidden> wrote:
> Rather, the whole profiling process is built such that grays are as I described. Reflective grays are assumed to be spectrally flat, and emissive grays are assumed to lie on the chromaticity coordinates.
I realize that this thread seems to be winding down with collective responses of understanding, but I just can’t let this statement go without clarification.
Ben’s answers are great *physics* answers and seem correct from that perspective.
But color management, as it stands today using ICC profiles, is not based on spectral neutrality but rather, colorimetric neutrality. Also, the neutrality provided by a flat spectral response may not be what the profiling software or calibration mechanism is aiming for.
So, part one: Colorimetry:
To achieve gray balance, ICC profiles need only to balance the Lab values of the output (or estimations of the output) so the L* value is at the correct tone and the a* and b* values are at the gray aim point (which is not typically 0,0 as one might think).
What this means is the system is nominally using 3 spectral bands (corresponding to XYZ eye response) to determine gray via human perception rather than 36 some-odd bands to determine it spectrally. In many cases this could result in very similar output but metamerism can certainly mess things up.
Also, parenthetically, it’s worth mentioning that some profiling software may use spectral information when calculating gray balance or other color recipes. But that still doesn’t solve the problem (see below)
Part two: Gray Aim:
Having a fair amount of experience with the process of gray balancing printing systems (I co-write the IDEALink, Curve2, & Curve3 G7 calibration software), I can tell you that the aim used for gray can vary a fair amount depending on the color of the substrate, the color of the darkest black and the intention of the user.
Think of the gray aim as a line (or curve) stretching from the blackest point of print (which may not be entirely black) up to the media / paper white. Gray balancing software, or the calculation of neutrals in an ICC profile, will attempt to determine the best CMY or CMYK recipes to hit that aim all the way from max black to paper white.
If you choose a spectrally neutral line, then as the ‘neutrals’ lighten they will first go from the color of the darkest black (brownish warm in many cases) over to the absolute neutral gray axis. Then they will continue to lighten while being absolutely neutral (increasingly fighting against the paper color that shows through more and more). Finally, with no more ink available, the ‘neutrals’ will fling over to the paper color. As you might imagine, this doesn’t often result in pleasing neutrals. When the black print point and the paper white are not very neutral it’s not a pleasing or expected result.
A more natural gray aim is more like a line from the black print point straight up to the paper white. This tends to be smoother, avoids color changes / fringing and matches the expectations for neutrals of the viewer. Importantly, this doesn’t tend to be spectrally or absolutely neutral at all.
Using spectral data to choose inking recipes that reduce metamerism is a great idea and some systems may do so, as I mentioned. But I’m pretty confident that spectral neutrality is not the basis for neutral calculations in the ICC architecture today.
regards,
Steve
_______________________________________________
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