Apache-Talk @lexa.ru 

Inet-Admins @info.east.ru 

Filmscanners @halftone.co.uk 

Security-alerts @yandex-team.ru 

nginx-ru @sysoev.ru 




      :: Filmscanners
Filmscanners mailing list archive (filmscanners@halftone.co.uk)

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: filmscanners: Re: Nikonscan v VueScan

Try it.  The colour channels, of course, don't change.  But a change of
intensity combined with no change in saturation ultimately leads to a
perceived change of colour if the change is large enough.  This can be
really ugly.

Or, quite lovely:


I created this version simply by pulling a curve in the L channel really
high up.  It has created an effect much like a kind of hand-colouring.

If you take the image above into PS (it should have the sRGB profile
attached to it) and apply the enclosed curve in LAB mode (I use version 6,
dunno if earlier versions will support this file), you get back to the image
as taken (almost!).  It was bad photography on my part as the image was
about two stops+ under-exposed...  But this end result, which I discovered
while pissing about with the curve in L (out of pure frustration!) is
something that I really really like.  So, a completely lucky accident.

There's a b+w version of the image in the same folder, which was my original
idea for this picture.


I have re-cropped it to take into account Vuk's comment on the image, since
the crop matches my original idea better.

Lastly, she was playing with her hair, just before I took the picture.
Immediately after taking it, she got up and, bang, I lost the chance to take
the picture I really wanted...  Sigh.

(Note since it is quite easy to get a 2 stops under-exposed image out of
Vuescan - when aiming for an "all data" scan, using None as the Color
balance - an L channel curve similar to the inverse of the one included
would, theoretically, be required (not necessarily as severe).  In doing so
the colour loss would be so dramatic that one would realise, pretty quickly,
that the right thing to do is to increase the gamma setting in Vuescan...
and re-scan.

Or, apply the excessive L channel curve, convert back to RGB and compensate
by reducing gamma in the "RGB-combined" channel, e.g. by 0.85 (which will
also add back some saturation).

In fact, using just this technique, I've got results with a Vuescan image
which have bettered all my previous attempts.  Funny eh?!!!!  Damned ironic.
Naturally, I worked in 16-bits, in case you're wondering.)


> -----Original Message-----
> From: owner-filmscanners@halftone.co.uk
> [mailto:owner-filmscanners@halftone.co.uk]On Behalf Of Maris V. Lidaka,
> Sr.
> Sent: 23 November 2001 01:54
> To: filmscanners@halftone.co.uk
> Subject: Re: filmscanners: Re: Nikonscan v VueScan
> The L channel should not change colors - it is Luminosity only.
> The A & B channels contain the colors, so I'm mystified by what
> your are seeing!
> Maris

Attachment: 2357 Undo.acv
Description: Binary data


Copyright © Lexa Software, 1996-2009.