Re: Colorimeter vs. spectrophotometer in display profiling
Re: Colorimeter vs. spectrophotometer in display profiling
- Subject: Re: Colorimeter vs. spectrophotometer in display profiling
- From: email@hidden
- Date: Wed, 1 Oct 2008 00:27:49 EDT
Graeme,
<<Eric is using "bands" in a looser sense, indicating that the XYZ weighting
filters have their peaks at different wavelengths, and can therefore
be regarded as 3 (rather poorly separated and overlapping) "bands".>>
That's right. As a former instrument designer, I'm using the term bands in
the general sense that designers do: a spectral integral. The term makes few
assumptions about shape, size, or other important properties as you correctly
note.
<<Given the way it's technical specifications read, I'm struggling
to see how such optimization can be slipped into the equation,
apart from the particular wavelength re-sampling co-efficients,
and they show no signs of being optimized at different wavelengths
(ie. the filter shape is constant).>>
As it should be. AFAIK, when well-done, there is no way to detect this other
than by reverse engineering the firmware or object code. However, there are
some "trick" samples that you can use to probe the optimization and instrument
design approach. But you won't get the specifics on that from me!
<<I would suspect the vast majority of patents aren't actually
being used in any real products>>
I don't know about that in general. But in this case, it is a fact that the
optimzation methods described earlier and that are in the literature, are
indeed being widely used. That is in no small part why we can get such terrific
performance from inexpensive spectros and colorimeters that depart substantially
from ideal beahvior.
<<Sorry, I still don't see the logic of this argument, unless one introduces
quantum and/or quantization error effects>>
That indeed is a good part of it at these low signal levels. As I tried to
put it more simply earlier, sampling low light levels, amplifiying, quantizing,
mulitplying and subtracting to re-assemble XYZ is a huge challenge in making
the sum of the parts approach the whole. See Edmund's description - he
basicallly has it right.
This is getting to a level of detail that's probably not of much interest to
this list. But I would be happy to go over the vector space math, optimization
strategies, and references with you off-line, by Skype, or at CIC. These
kinds of optimizations is one of my favorite subjects...
Eric Walowit
Koln
**************
Looking for simple solutions to your real-life financial
challenges? Check out WalletPop for the latest news and information, tips and
calculators.
(http://www.walletpop.com/?NCID=emlcntuswall00000001)
_______________________________________________
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