Filmscanners mailing list archive (firstname.lastname@example.org)
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[filmscanners] Re: Some findings/thoughts on the Sprintscan120(comments verywelcome please)
A CCD scanner will not read the deepest shadows of transparency film, and
one can really only asses how well it's doing in this regard by comparing
the scan to the image on a light table. The "brick wall" shadow end
histogram is inevitable scanning chromes, but shouldn't be scanning negs
since that's the least dense part of the film. I've noticed some scan
programs clip neg scan shadows anyway however, and I would avoid using those
personally. But again, if the information in the film makes it to the scan
the histogram may be misleading or irrelevant.
If a program compresses the scale to less than 255 steps, but leaves all the
image information available, end point expansion in Photoshop is a viable
way to work. One may prefer hi-bit files for that approach in order to end
up with 255 discrete levels (no gaps in the histogram). Some programs can
accurately set end points to near 0 and 255 and leave all the image
information available, and that approach may not require hi-bit editing
because the critical tonal expansion to end points is done by the scan
driver, so you still end up with the full 255 steps in an 8 bit file.
However, one may find that with certain images it's preferable not to have
end points all the way to 0 and 255, and the "conservative" approach of
"short" end point settings allows the scan operator to set preferred end
points later in PS. Certain programs seem much better than others at
setting end points at or near 0 and 255 without loosing image information,
Nikonscan, and apparently Flextight and Silverfast also, but I haven't used
----- Original Message -----
From: "Simon Lamb" <email@example.com>
Sent: Sunday, April 14, 2002 1:07 PM
Subject: [filmscanners] Re: Some findings/thoughts on the
Sprintscan120(comments verywelcome please)
I forgot to mention one BIG thing!
On many of the images that I scanned the histogram in SF and Insight showed
that there were no pixels below 18-23, i.e. there was no true black in some
images. Considering that one image was of a jet black belt, I found this
odd. As I said before, at about 18-25 there is a big straight line, as if
the scanner just didn't want to record anything below that threshold or
didn't find anything below it. I did notice when I scanned a Delta 100
image at the dealer that the Flextight histogram went from 0-255 on the
black belt image, with pixels at every step, whereas the SS 120 histogram
stopped at 23.
That said, the images looked fine after some curves and levels adjustment,
but I am concerned that I may be losing some information in the scans.
Is it just that the SS 120 is not finding any pixels below that level, i.e.
the belt does not have a real true black as far as it is concerned? If so,
how come the Flextight and my LS30 think there are, and the SS 120 thinks so
when scanning the same neg using Vuescan? Has anyone else seen this
histogram behaviour with the SS 120?
On 14/4/02 11:38 am, "Arthur Entlich" <firstname.lastname@example.org> wrote:
> Hi Simon,
> I am pleased you have brought these issues into the public list because
> there are people here who have more experience with Insight than I,
> especially in terms of how it functions with the SS120.
> Further, I had another thought. It is possible that there are aspects
> of the software that work better in the PC environment than the Mac.
> Just to make it clear, to avoid any flame war, I am not implying that
> the PC is a better platform for scanning, or color management, or
> whatever. What I am suggesting is that perhaps some aspect of the Mac
> OS causes differing results with Insight, and that either there are
> adjustments that can be made within the Mac operating system or Insight,
> which I would not be aware of, which can remedy some of the problems you
> are experiencing.
> Therefore, the reported experiences which seem so divergent, might be as
> a result of different platforms or OS. Someone who is running an SS120
> with Insight on a Mac might be able to provide you with some guidance
> which I am unable to if that is the case, since all my experience with
> Insight has been under the PC Windows OS platform.
> Simon Lamb wrote:
>> I have had some private communication with Art Entlich regarding the SS
>> and he has been, and is continuing to be, very helpful and supportive. I
>> though I would share some of the comments that I made with the group to
>> if anyone has any comments to make. All would be welcome.
>> I have scanned quite a few slides so far and have had varying results.
>> film term in Insight for Kodak Porta 160 is way, way off. The resulting
>> scan bears absolutely no resemblance to the original neg. or even the
>> prints, and required so much work in Photoshop to recover it that I gave
>> it just seemed unrecoverable. I think the film terms in general within
>> Insight need to be reviewed. Scanning using the generic slide terms when
>> using Provia also produced bad results. The only good scan I got was
>> the generic slide term when scanning Kodak E100VS. As for black and
>> after over ten attempts with Scala, Delta 100 and Delta 400 I gave up.
>> scans where very dark, the black point stopped dead on all scans at about
>> 30, as if all the pixels at that end had been pushed up against a wall.
>> So, I moved on to using Vuescan. The Provia scanned well although the
>> colour accuracy was not too good. My Nikon Coolscan LS30 produced a
>> scan from a colour perspective, although obviously not as detailed. The
>> Delta 100 scan was one of the best black and white scans I had ever seen.
>> Absolutely perfect tonal balance and immense detail. The Scala was good
>> lost some detail in the highlights. The Porta 160VC was detailed but
>> was a significant amount of white speckling all over the scan. I assume
>> from this, and the fact that the Insight term produced the worst scan
>> imaginable, that the SS 120 just has difficulty with the Portra emulsion.
>> shame, since my LS30 scans it very well.
>> So where am I now. Well, the SS 120 can obviously produce detailed
>> but I will have to rely on Ed Hamrick's Vuescan to get them, especially
>> black and white which is a big proportion of my work. I am not really
>> about relying on third party software because should Ed decide to pack it
>> then I will have a scanner from which it will be difficult to get the
>> results I need. The carriers are fiddly. In many cases the 6X6 film
>> not lie perfectly flat in the carrier, and it is impossible to line up a
>> strip of 35mm unless you leave the carrier slightly undone. Most of the
>> scans I did using Insight required a lot of work in Photoshop to get them
>> close to what I wanted, and some were just too far out to be workable.
>> I think I will return the SS 120 and try the Flextight Photo. I did find
>> when comparing it side by side with the SS 120 in the store that the
>> just about always reproduced the image as near as possible to the
>> colour, contrast etc. The built in film profiles seemed to be accurate.
>> may take twice as long to scan, but I may save that additional time not
>> having to do so much in Photoshop.
>> Your comments on my findings would be welcomed. I know a lot of people
>> the SS 120, so either I am doing something wrong or they just put the
>> in to correct images post scanning.
>> I was using Insight 5.5.1.
Unsubscribe by mail to email@example.com, with 'unsubscribe
or 'unsubscribe filmscanners_digest' (as appropriate) in the message title
Unsubscribe by mail to firstname.lastname@example.org, with 'unsubscribe
or 'unsubscribe filmscanners_digest' (as appropriate) in the message title or