[argyllcms] Re: White Point

  • From: Iliah Borg <iliah.i.borg@xxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Sat, 7 Nov 2015 22:33:13 -0500

Dear Hening,
Dear Ben,

With Sony lossy compressed files, you want to setup "Full-well limited" in
overexposure detection options in RawDigger.

The limit is actually 15860 (after applying normal workflow, that is the tone
curve, decompressing, and subtracting black level.

Looking at raw, there are very few hot pixels on the target, 5 in G1 channel
and 6 in G2 channel. The rest is the hotspot on the teflon tape, on the pin (I
always use dulling sprays on any metal parts, or paint them myself), and on the
plastic frame. That is - there is practically no overexposure, and RPP will
have no difficulty here.

Light setup is a little unfortunate, not even enough - also in white balance.
There is a reason why we have flat field option in RawDigger. It also allows to
use only one light source (with a Fresnel lens) to minimize flare; light being
positioned at a good angle to the lens to avoid reflections. If the goal is a
daylight profile, it is better to use some Rosco or Lee filters on the light to
raise the colour temperature to approximately 5500K.

Teflon tape is better applied to a matt white backing.

Some screenshots http://s3.amazonaws.com/IliahBorg/HeningBettermann_01.zip of
the histograms to help viewing how I determine the overexposure: everything
starting from 15860 is overexposure; in my work I stay 1/6 EV down from that
number, or the data becomes not very reliable.

To get CGATS out of RawDigger, it is better to check "Selection/Sample stats:
discard abnormal pixel values" in Preferences - Data processing.

On Nov 7, 2015, at 9:39 PM, Ben Goren wrote:

On Nov 7, 2015, at 5:07 PM, Iliah Borg <iliah.i.borg@xxxxxxxxx> wrote:

FTP refuses the connection, I will try later.

I was able to get in with the credentials in the original post. But I don't
feel qualified to comment. It's a Sony camera that I'm not at all familiar
with. The image opens in Raw Photo Processor v4.8.0 superficially looking
perfect -- white balance, exposure, the works. That's a subjective
evaluation; I didn't run any numbers. In Raw Digger v1.2.1...something's
clearly amiss, because many of the patches are clipped at zero in one, two,
or three channels.

So...I'm not quite sure what to suggest, though I do think it's well worth
considering using RPP. Chances are superlative that the built-in profile for
your camera that ships with RPP is going to be better than anything you
(Henning) are going to be able to be able to make with a ColorChecker.
Iliah's profiles are damned hard to beat...I should know....

I've had lots of stuff come up the past few weeks, including an email server
die on me...I've got all the pieces in place for low-cost spectral camera
profiling, but it's going to be at least another couple weeks before I can
get back to photography stuff. In the mean time...if this profiling exercise
of yours is for general-purpose photography, you're not going to do better
than RPP unless you're using either a monochromator or a spectroscope as an
integral part of your profiling workflow. If this is an exercise in fine art
reproduction or similar copy work, you _can_ do better, but only with a
several-dozens-to-hundreds-of-patchs chart with lots of spectrally diverse
samples, with said samples drawn from a similar palette as the art itself.
And, even then, what you're _really_ profiling isn't the camera so much as
the illuminant....


Best regards,
Iliah Borg
LibRaw, LLC

Other related posts: