[argyllcms] Re: help with camera profile

  • From: "alberto.ferrante@xxxxxxxxxx" <alberto.ferrante@xxxxxxxxxx>
  • To: "argyllcms@xxxxxxxxxxxxx" <argyllcms@xxxxxxxxxxxxx>
  • Date: Thu, 19 Jul 2012 09:14:07 +0100 (GMT+01:00)

Dear Iliah and Edmund,
thanks for your replies. Following my replies 
to 
your questions.

> when you mean "profile" do you mean Raw or Jpeg? 
If 
Raw, which converter?

Raw; the converter is ufraw, an open source 

converter based on dcraw.

>First, how even was the light at shooting 

the target, what was the light 
>source, how much flare you are 

getting, how linear is the greyscale step wedge 
>on the shot, for 
each 
of 4 channels?

The light source was sunlight (in a day with 
clear sky 
with sun at the zenith). I used some black paper to shield 
the target 
from possible reflexes on the bottom and on the right and 
left. The 
illumination was even (I checked it on the images and, 
before taking 
the shot, with a Sekonic light meter). I do not see any 
flare in the 
image and I positioned the target to avoid glare as much 
as possible (I 
don't see any on the final image). I can post the 
picture of the target 
tonight when I get home, if it could be useful.


>Second, what colour target are 
you using for creating the profile - 
how does it 
>warrant -qu for a 
LUT-based profile?

I am using the 
target made by Wolf Faust (coloraid.
de).
As I mentioned already, I 
used it to profile other cameras in the 
past.

> Third, how do you 
obtain the source rgb image for profiling? 
Is it through 
> ufraw in 
profiling mode?

Yes, through Ufraw, by  setting the output to absolute 
colorimetric, 16 bit, tiff, with no input 
color profile and linear 
base and output curves.

> Forth, why -al 
instead of -ax; if not a -G 
which is a much more stable solution?

I tried them both, but with no 
good results.
In the past I tried all of the available profiling 
options and at the end I preferred the -al, but 
I am ready to change 
my mind if any other option works better!


> 
Fifth, does ufraw allow 
for a linearization tone curves over all 4 
channels? If 
> not - ask 
them to implement this option. If they are 
unwilling, linearize 
> 
yourself. Linearity must be calibrated before 
profiling.

Can you 
please give me some references about this? What is 
a linearization 
tone curve?

Thanks again!

Regards,
     Alberto

P.S. I apologize for 
my previous post, but I hit the "send" button before the message was 
ready...





Invita i tuoi amici e Tiscali ti premia! Il consiglio di un amico vale più di 
uno spot in TV. Per ogni nuovo abbonato 30 € di premio per te e per lui! Un 
amico al mese e parli e navighi sempre gratis: http://freelosophy.tiscali.it/

Other related posts: