Re: monitor calibraters
Re: monitor calibraters
- Subject: Re: monitor calibraters
- From: Steve Upton <email@hidden>
- Date: Thu, 19 Apr 2001 23:07:13 -0700
At 8:48 AM -0400 4/19/01, email@hidden wrote:
On to part two: what about other devices? We've heard a lot lately about the
advantages of using real spectrophotometers to calibrate monitors, from
companies that have supported the SpectroCam, and now from Gretag Macbeth
with thier EyeOne moniotor calibrator. "True 64 band spectral data, capable
of reading the quirks and spikes of LCD screens!"... but lets consider the
balance of this capability carefully. If three sensors is unsufficient to
determine the color of monitor phosphors, and only capable of accurately
deermining luminance, what about the other extreme? What value can we place
on 64 bands of spectral data, versus the three in a Chroma4, or the 8 or 9 in
a Monitor Spyder or a DTP92? Considering that these monitor profiles need to
be applied using current technology, there is such a thing as overkill.
I'm going to have to disagree here...below
How
does one apply all that spectral data to a video card that offers no method
of calibrating the monitor except for a single number gamma curve for each
channel?
The spectral data are in no way applied to the gamma curve. Also,
while the monitor profile may represent the gamma using a single
value, the actual calibration of the monitor in the graphics card (in
conjunction with the pre-calibration of the monitor itself) is
performed with curve information in the vcgt tag that can contain
many values to represent the curve. It can use one curve for all 3
channels or a separate curve per channel. More correctly one result
of the spectral data is used in calculating the curves - but they do
not, of course, store any spectral data.
And even when applying the profile in Photoshop, is excess data of
this type offering any significant improvement to the R, G & B curves in the
profile?
There is no "excess" of data in this situation. The color device, be
it a colorimeter or spectrophotometer, is doing its best to simulate
human vision. The result of this attempt is typically XYZ data (for
monitor purposes). Colorimeters break the spectrum into defined bands
using optical filters. The information is then used to try to
simulate the response of the human eye. In spectral readings, the
light is diffracted into many more bands (31-36 in most of the
instruments us mortals use). This spectral data is then combined with
human eye response curves to obtain the desired colorimetric
information. No matter how well constructed a colorimeter is, it will
suffer more from physical limitations (filter colorants, etc) and
manufacturing/mechanical limitations (consistency, fading, aging)
than a spectrophotometer will. Spectrophotometers gather more data
and use them more accurately. Add to that their flexibility - like
the ability to use different math for LCD purposes - and you have a
better instrument.
Not in my experience.
This I understand. I appreciate your experience. I have found, in the
field, that colorimeters will often do a very acceptable job. There
are instances, however, when our Spectrolino has built a considerably
better profile that several different colorimeters from different
manufacturers on hand.
Until we are doing something significant with
true 3d look up tables for monitors, then the 8 or 9 sensor level is an
optimal balance for monitor profiling. 64 bands it a bit analogous to carving
it with a scalpel, then applying it with a sledgehammer... but until
ColorVision releases an LCD optimized version of the Monitor Spyder, we won't
know what the practical comparison really is.
I'm not sure this is true, that practical part that is. With the
Eye-one we now have an instrument for under $600 which builds a
high-quality profile of CRT's and LCD's. That strikes me as well
within the limits of practicality - it did a little over a year ago
when we where all recommending the DTP-92 in that price range.
This all said. I should emphasize that a fair amount of my side of
the discussion is on the academic side. An 8-band colorimeter is
probably one of the better compromises between the detailed gathering
of spectral information and an inexpensive and rugged instrument. As
with most things in this industry, I think each device has its place
in the market and this is why we carry, recommend and sell both types.
Regards,
Steve Upton
+--------------------------------------------------+
CHROMiX / Profile Central
www.chromix.com www.profilecentral.com
+--------------------------------------------------+
--