Even more about monitors and gun control
Even more about monitors and gun control
- Subject: Even more about monitors and gun control
- From: "Nick Milley" <email@hidden>
- Date: Fri, 15 Jun 2001 11:06:41 -0400
In a message dated Wed, 13 Jun 2001 17:54:59 EDT email@hidden writes:
>
No apology necessary; this is interesting Geek stuff that is not often
>
discussed in public.
Thanks - you probably don't want to encourage me too much.
>
Your count is a bit off; the Spyder has eight sensors... and you fail to
>
mention how this algorithm goes about improving your dark readings, or how
3
>
sensors can be more accurate then 8 for color readings. I'd love to hear
more
>
on any of that...
I believe the Spyder has seven sensors for measurement and one sensor for
detection of monitor refresh. Similarly, the Chroma4 has three sensors for
measurement and one sensor refresh detection.
As far as three versus seven, consider the nature of the problem. The whole
purpose of a color sensor for
monitor calibration is to measure monitor color the way the eye sees it. The
eye has detectors for red, green, and blue light and a monitor calibrator
tries to match the eye's RGB sensitivities by placing tuned color filters
over the light sensing elements. The color filters are supposed to pass
light the way the eye responds, the so-called "color matching function."
Anyhow, the Chroma4 uses three sensors (R,G,B) with tuned filters which
closely match the eye's response - very straightforward and the technique
used by most monitor measuring instruments including LMT of Berlin, Germany,
the highest of the high-end instruments ($35,000+).
Anyhow, the Spyder uses seven filters. Well, it is very hard to make seven
filters pass light like three filters do - you have to have very sharp edges
and very tight alignment. Imagine taking one sheet of paper with three
curves
and cutting them up into seven pieces and trying to tape them back together
again - you can come close but perfect is very hard to achieve. If you were
trying to measure a reflective color, for paint formulation, for instance,
more sensors (with different color filters or a grating, e.g.) would be
important
to give spectral data. But, when trying to replicate the color-matching
function,
more sensors creates a curve-matching problem.
Then there is the question of electronics - each sensor's output must be
timed and counted exactly in sync with the monitor's refresh cycle. With
three sensors, you only have to synchronize and take data from three
sources. Obviously with seven, you have over twice the problem. Also, the
more
sensors, the more sources for counting errors especially at dark luminances
which I have already talked about. Maybe this is why the Spyder is not
spec'ed for color accuracy below 10 cd/m2, the approximate luminance of a
90,90,90 gray patch (on a monitor of 80cd/m2 at full white). Kind of
troublesome if you are serious about color.
Anyhow, I believe the whole object of a color sensor is to accurately
measure color on a monitor - the number of sensors in the device is
irrelevant.
Tell me your specs and I'll tell you how good you are. Looking
at the specs of the Chroma4 (x,y +/-.002) and the Spyder (x,y +/-.005), it
should be frightening that I could have two devices that give me an
interunit error of .01. This is horrible and the visual error would be
overwhelming. I submit that large numbers of sensors do not provide a viable
solution for the measurement of color on a monitor.
Hope I didn't bore most of you out there.
Nick Milley
Sequel Imaging