Re: filmscanners: Vuescan trials

["Dana Trout" <dana@troutcom.com>]

The reason that either set of mask values gave acceptable images is
that
 1. The image you were scanning was reasonably balanced
    (had good whites and blacks), and

 2. You had selected (probably) "White balance" in the color
    tab, as well as "Auto white point" and "Auto black point".

What happened is that even when the image was scanned with non-optimal
mask values the color tab selections rebalanced the image based on its
highlights and shadows (white and black points), which were assumed to
be neutral.

You would have encountered problems if the scenes that you were
scanning did *not* have neutral whites or blacks.
  --Dana
----------
From: shAf <rarewolf@roadrunner.nf.net>
To: film scanner list <filmscanners@halftone.co.uk>
Subject: filmscanners: Vuescan trials
Date: Thursday, September 13, 2001 9:00 AM

  I appreciate using Vuescan for many reason, but I have to admit ...
sometimes its behavior is inconsistent and confusing.

  Using version 7.1.9, I wanted to setup Vuescan for a general
exposure and
for scanning many frames of Supra 400.  I wanted to use auto-exposure
of a
representative frame to determine the general manual exposure settings
..
and then wanted to include the non-exposed part of the film with
"auto-mask"
for determining manual mask settings.  I am a bit confused with the
results.

  Loading default settings and setting appropriate parameters for
negative
film (Kodak/Supra/400), and scanning a full represenative frame (with
reds,
greens, blues, plenty of "white", and deep shadows) resulted in the
following "auto mask" values:

R =.948
G =.705
B =.725

  ... values which imply "orange".  However, if I now choose "manual
expose", and scan a frame which includes non-exposed film, I get these
"auto
mask" values:

R =.995
G =.920
B =.969

  ... not very orange at all.  However, both settings for the mask
resulted
in acceptable images given the same "color" settings ... I have no
idea
which settings were optimal, and I don't know how to interpret the
second
set being so close to a neutral shade, but want to believe they were
the
better determined set.

  Lastly ... the IR exposure was automatically determined to be 2.24. 
At
the time I thought this was a bit high (previous experience with
earlier
versions of Vs usually implied IR exposures near 1), but decided to
check
later and after a final scan.  Indeed, when I asked Vs to deliver a
16bit IR
Tif and scanned from memory, it delivered a nearly white image with a
few
gray defects.  I thought this should be closer to generally neutral
gray so
that defects could show as "dark" or "bright".  But it seems I cannot
experiment with manual IR settings ... even when using similar
exposure
values I cannot duplicate the "white" over-exposed IR ... but instead
get
very dark grayscale images in spite of my exposure values ranging
between
1.0 - 3.0 ... weird!!!

  (Ed ... if you want me to create a log file and submit it to you,
please
give me a brief desciption about how ... and whatever else you might
need)

shAf  :o)