Re: purpose of Granger rainbow chart?
Re: purpose of Granger rainbow chart?
- Subject: Re: purpose of Granger rainbow chart?
- From: Jacob Rus <email@hidden>
- Date: Mon, 8 Mar 2010 12:00:01 -0800
Marco Ugolini wrote:
> Jacob Rus wrote:
>> What I meant to say is
>> that the function is not smooth (continuously differentiable), and in
>> particularly noticeably has a nasty diagonally slanted edge in
>> lightness (L*), where the derivative changes abruptly.
>>
>> This visual artifact is clearly visible
>> <http://www.luminous-landscape.com/images31/Granger_Chart.jpg>.
>
> "Derivative"? I can't follow you. Is that calculus?
>
> Plain English, please, if this is to be a conversation, rather than a
> monologue. Speaking for myself, I'm not a Bachelor of Science.
Sorry. I didn't mean to be confusing.
Basically, what I meant to say is that we can see what look like sharp
edges in a funny diagonal pattern in that image. (I hope I'm not alone
in seeing this!)
This funny diagonal pattern is caused by the perceptual lightness
(usually expressed as something like Munsell Value or CIELAB L*) is
not smooth. More on that shortly.
> Incidentally, as far as I know, isn't "luma" a term specifically meant for
> video signals?
>
> My understanding is that, in the case of digital images, we use LUMINOSITY
> (or LIGHTNESS), not luma, independently of whether they are shown on a
> self-luminous display or output as a reflective print.
Well, yes, it is.
Basically, Photoshop's "luminosity" blend mode is calculated by taking (roughly)
0.3 * R + 0.6 * G + 0.1 * B, which is just the formulation for
NTSC/Rec. 601 "Luma". Given that "luminosity" doesn't really have any
scientific meaning, and is just a term made up for this use by
Photoshop, but has been adopted by various other software applications
to mean things other than this, I think "luma" is a less ambiguous
term. (If someone looks up "luma", thay'll land on a bunch of pages
which explain it in clear direct language).
Anyhow, the thing is that luma (or "luminosity" if you like) is
computed from gamma-corrected R, G, and B components. This is
unfortunately not the way that the human visual system works, which
means that ordering the luma of a set of colors does not imply the
same ordering in perceived lightness. Basically, the use of luma, both
in video applications, and in Photoshop, is a way of approximating
lightness without going through a computationally expensive (at least
for mid-1990s computers) gamma conversion step.
So when we look at this picture, the parts near the chromatic colors
end up at a higher lightness than their luma (if used as a proxy for
lightness) implies, and therefore look like a bright zig-zag band
along the edges stretching from pure red to pure yellow to pure green
&c.
Anyway, the artifacts caused by the use of luma as a vertical
direction are likely to have a similar effect on prints after
converting to CMYK, making avoiding visual artifacts with this test
image sort of impossible (as far as I can tell).
Does that make any sense? I dunno, maybe I'm just confused about the
purpose and use of this rainbow chart.
Cheers,
Jacob
_______________________________________________
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