Re: 1 billion colors
Re: 1 billion colors
- Subject: Re: 1 billion colors
- From: Wire ~ via colorsync-users <email@hidden>
- Date: Mon, 6 Jan 2020 04:57:21 -0800
This convo is on the edge of nonsense.
OMG Do some color numbers in some color space produce indistinguishable
stimulus? Consider a color-blind viewer. Or close your eyes.
The point of the extra bits in context could not be more simple: 8 bpc
color rendering can have problems with quantization noise. As dynamic range
is expanding in newer gear pertaining to UHD TV HDR, this limit is
encroached. A couple more bits per channel were added to standard display
data formats. Just like high bpc support was added to Photoshop.
Marketeers looked at old-school Windows / Mac color control panel
"hundreds, thousands, true (millions)" which always just counted the number
of unique numerical combinations and disregarded important details like
indexed color (the original LUT) and kept rolling with it. So what?
To know anything about the CIE tristimulus model, is to know it's about a
math that connects a physics of light to an idealized observer using
statistical sampling of perceptible difference across a human population.
The subjectivity of the qualia is subsumed by the model. It works because
our visual systems work a lot alike, just like most of us have arms and
legs with fingers and toes, but not wings.
To note that Pro Photo RGB has some values that don't lie within the CIE
spectrum locus is fine. In integer Lab lots of values don't fall into the
spectrum locus.
However an RGB / YUV display that processes 1 billion colors on its
interface may reasonably be expected to produce 1 billion quantifiably
different stimuli on its face that lie within its CIE gamut volume, because
of the way the data format works. We all know (I hope) that how many of
these are needed for any given application is a subjective matter. If you
only present grayscale images, maybe only 0.00001% of them aren needed. If
you prefer sepia, maybe more, etc.
The marketing claim is fair: there are applications where the full range of
input values are known to be useful and applicable.
Don't get lost in the noodling over semantics of the term color. If you
want to freak out about it, consider the word "time" and linguistic tense.
Then observe a Zen moment of silent contemplation.
On Mon, Jan 6, 2020 at 4:04 AM Roger Breton via colorsync-users <
email@hidden> wrote:
> ... Took the liberty to change the title of the thread.
>
> Out of 16.7 million or 1.07 billion colors, I suspect there are "redundant"
> combinations? Such that RGB = A,B,C or G,H,K come out to be the "same"
> color, visually? I think it is NOT unintuitive to think that this is in
> fact
> very likely, David? The same could be said of 1 billion colors, there ought
> to be combinations of colors that come out to be the same visually?
> Although
> I don't know anyone who ever tested this, extensively. And if I may, I
> don't
> know to what extent can a person with normal color vision be expected to be
> able to discriminate among 16.7 million colors?
>
> / Roger
>
_______________________________________________
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