[argyllcms] Re: Spyder4: Difference in using argyllcms and original software

  • From: Ben Goren <ben@xxxxxxxxxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Wed, 11 Nov 2015 10:22:30 -0700

On Nov 11, 2015, at 7:49 AM, Robert Hansen <robertdeanhansen@xxxxxxxxx> wrote:

How long does it take to calibrate with a colorimeter? My i1pro takes forever.

As a generalization, colorimeters are superior to spectrometers for profiling
displays -- with some caveats.

Conceptually, you can think of a colorimeter as a three-channel spectrometer.
And simple physics / math / optics / geometry says that, for a given form
factor, you get more photons for a wide slice of the spectrum than for a narrow
one. And more photons means more signal and thus a better signal-to-noise
ratio, so you have more options for either shorter integration times (faster
operation) or better reads of dark sample or both. That's especially the case
with the i1 Display Pro, which has an unprecedentedly large aperture for that
class of device.

The cost, of course, is -- obviously -- spectral resolution. But, when
profiling a display...well, if you know the spectral shape of the primaries --
the spectral power distributions of the red, green, and blue light sources --
then all you really need to know for profiling is the relative intensities of
the primaries for any given RGB value sent to the display.

A colorimeter isn't going to be able to determine the spectral shapes of the
primaries, but this isn't as much of a problem as one might suspect. First, for
"good enough" work, there's not all that much variation between different
display technologies. Start with some basic assumptions about the display type
(CRT, LCD, IPS / TFT, backlighting type, etc.) and that's enough for most
purposes.

But, if you've got both a colorimeter _and_ a spectrometer, along with
intelligent profiling software like Argyll, you can get the best of both
worlds. Look into Argyll's "ccmxmake" function, and you can see how you can
make a pair of measurements with both instruments that lets Graeme's code do
all the math it needs so that, when you later do your measurements with just
the colorimeter (on that particular display), Argyll also (effectively) knows
the spectral shapes associated with those colorimetric readings. The net effect
is that you get the speed and low-light performance of the colorimeter but with
the spectral precision of the spectrometer.

And...the results are amazing. Especially if you do an XYZ LUT profile with an
insane tens-of-thousands-of-patches chart. Yes, insane
tens-of-thousands-of-patches will take a few hours to read...but you can let
that run after you're otherwise done for the day. And I'm not sure if it's even
a practical option to attempt that sort of thing with any other system. Shadow
detail, evenness _and_ discernibility of small-increment step wedges, color
fidelity, lack of artifacts in a Grainger Rainbow...I've never seen anything
like it before. I can't sing Argyll's praises high enough.

Cheers,

b&

Attachment: signature.asc
Description: Message signed with OpenPGP using GPGMail

Other related posts: