Re: On monochrome images and metamerism, or not.
Re: On monochrome images and metamerism, or not.
- Subject: Re: On monochrome images and metamerism, or not.
- From: Simon ONeill <email@hidden>
- Date: Sun, 20 Jun 2010 18:41:23 +0100
Hi Ernst,
I have observed differences in a typical (good colour vision) observer's ability to distinguish between colour shifts near the neutral grey axis and between more saturated colours. I did some tests which showed people unable to distinguish easily between DeltaE(76) of 5.0 for saturated colours (near the gamut edge of an inkjet printer) whereas a DeltaEab of 0.5 is typically visible when near the neutral (grey) axis. This was a very small sample, but was intended to demonstrate (as it did) that the Lab space is not truly "perceptually uniform".
In a colour print therefore, we might be asking an observer to remark on differences in colours in a region where they can easily fail to detect differences of 5.0 deltaEab whereas the whole (mildly defective) monochrome print limits its colour gamut to a region of maximum sensitivity (i.e. all almost grey and very near to the neutral grey axis).
I suspect that you are correct in the suggestion that the whole image provides a neutral "reference", so small (ab) deviations are easily detected. I suspect that an overall cast is unlikely to be detected if it applies uniformly to the image - like a paper-base shift would. (e.g. Papers with optical brighteners appear white until compared with a truly neutral white.) This is probably not too surprising because the eye & brain's ability to adapt to different illuminants is impressive, and a small "tilt" in the spectrum from a blue-ish paper is easily ignored when there is no reference to compare with. Perhaps the eye&brain adapt to the "majority view" of the image.
Simon
On 19 Jun 2010, at 16:07, Ernst Dinkla wrote:
> What is it that makes the unexpected color shift in B&W images between a
> D50 viewing light and daylight more distinct compared to the color print
> from the same printer model? Specifically the hue shifts over the tone
> range. Assuming that the B&W print gives a very even hue impression
> under D50 light over the tone range.
>
> Is it based on a "virtual" expectation of what a greyscale image should
> be or do we have some reference in the image itself that marks the
> shifts? Is the explanation that in a color print small color shifts go
> unnoticed not related to the fact that there are too many variables in
> the print to create a reference for unexpected shifts in other parts of
> the image? The adaption of the eye and brain to different light sources
> covering up unexpected shifts in the color print but not in the
> monochrome print as there is some reference left in the other monochrome
> tone?
>
> I think it is that reference to the rest of the print, a kind of mean
> values of hue and saturation, that let us decide whether we think the
> tone ranges match or fail on hue and saturation. A shift in tone/value
> in the same light chance probably going unnoticed as a reference for that
> isn't present. That is different to the metameric match/failure with
> two or more samples. In the metameric match the two samples give equal
> stimuli on the 3 values and they can differ on the 3 values in a
> metameric failure. The reference to one another makes it possible to say
> on what of the 3 values they differ. At the same time making that
> condition of two samples necessary.
>
> I wouldn't mind if someone says that the tone range of a single B&W
> print under D50 makes a good metameric match or metameric failure and
> adds "I mean on hue and saturation". It may not be written in the books
> like that but I guess that the researchers were not unaware of the
> relation between the phenomena either.
>
> I'm using the HSV model here as I find that the easiest to describe it
> with. I'm sure there is a better color perception model possible but
> that doesn't change the idea behind it.
>
>
> --
> Met vriendelijke groeten, Ernst
>
>
> Dinkla Gallery Canvas Wrap Actions
>
> | Dinkla Grafische Techniek |
> | www.pigment-print.com |
> | ( unvollendet ) |
>
> _______________________________________________
> 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
_______________________________________________
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