[argyllcms] Re: Argyll CMS as a spectroradiometer
- From: Graeme Gill <graeme@xxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Thu, 21 Oct 2010 14:05:28 +1100
Jeremy Pronk wrote:
lightbox with considerable dynamic range between them. As far as I can see
Argyll is suitable for this but I'm not sure which consumer probes are
suitable. The colorimeters seem to handle more dynamic range than the
spectrometers. Do I require an emissive chart reading probe or would a
display probe be suitable? Is it possible to calculate absolute luminance?
Generally display probes are not suitable, because their filters and
calibration matrices are setup for typical display color spectra,
and the color accuracy will be poor for other sources.
Additionally, the lowest cost instruments tend to be based
on commodity light to frequency converters, that (I think) don't
work very well at low light levels.
So you do need a spectrometer. The entry level spectrometers are not
temperature stabilized, and this can be the biggest issue with measuring
low light levels. The instrument drivers can make a difference too,
depending on how they adjust the instrument integration times to
suite the light level being measured.
The most likely candidate at this level, is the EyeOne Pro
(although a Spectrolino would be just as good if not better,
of you can get one.)
A step up from that is probably a spectrometer module from
StellarNet or Ocean Optics, but once again they are not temperature
stabilized. The next step up from that are the semi-laboratory
grade instruments at about 10 x the cost.
The paper can be found here,
http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.151.6187
Interesting. What's not explained is why measuring a HDR test chart
gives any better results than creating an RGB to XYZ matrix from
a more limited DR measurement. If the assumption is that you are working
with linear light values (ie. sensor non-linearity has been removed),
and the filters and sensors do not have differing spectral behaviour
at different light levels (usually a pretty good assumption), then
it's not clear what is gained.
Graeme Gill.
Other related posts:
