VueScan film types

[Oliver Vecernik]

Hi all,

I use VS to scan my archive. Most of them are negatives, some of them
are slides. I use a Nikon LS-5000 on a Linux box. I initially decided
to save them as JPEGs with 2000 DPI. This gives me approx. 2000x3000
pixels, enough for printing and for the web. One nice feature I like in
VS are film types.

Until now I was quite happy, but I encountered some film types *not* in
the list. One of them is a KODAK VR 100-3 for example. Although I
bought the professional version I haven't done any profiling (and I have
no IT8 targets). It is very difficult to decide if the colors are ok,
the best choice I found is FUJI SUPER HG 100 Gen 2. But this is my
personal taste.

Then I read some articles on the net and in this NG it would be wise to
scan "raw" and do the editing later. VS can do that (I scan with light
infrared clean, medium grain reduction and 4 times multi-passing) but
images get huge (30 MB with 2000 DPI, 120 MB with 4000 DPI) and I still
don't know what film type to use. :-\

My first question are: has someone found a better film type or can
confirm that this is a good choice? Do I need color calibration anyway?
What is the "correct" way to cope with this?

If I use Gimp (2.2) for editing this has some drawbacks. It lacks color
profiles (until 2.4) and it can't handle 16-bit TIFFs. It should be no
big deal for printing 4x6 or just editing for the web. There is no PS
for Linux, but Gimp can handle PS plugins. Are there any PS plugins for
"developing" these "raw" scans?

 

[Roger S.]

If you find a film profile that's useful to you, that's great, but
generally they don't do a whole lot of good (I never use them). They
are only intended to save time. You can do just as good a job in Gimp
without a plug-in.
I find it more useful to use "white balance", try to set color by right
clicking on a neutral area, or scanning with color balance set to
"none" and doing everything in post-processing (basically a raw
workflow). IT8 targets aren't really made for color negative films.

I'd also suggest not using grain reduction and doing that to taste in
your image editor (I use noise ninja myself- not sure if it's available
for Linux/gimp). Try it with and without VS grain reduction and see
which looks better.

Good luck,
Roger

 

[Raphael Bustin]

Hi all,

I use VS to scan my archive. Most of them are negatives, some of them
are slides. I use a Nikon LS-5000 on a Linux box. I initially decided
to save them as JPEGs with 2000 DPI. This gives me approx. 2000x3000
pixels, enough for printing and for the web. One nice feature I like in
VS are film types.

I never found them useful at all. Regardless of
what scanner or scanner driver I'm using, I scan
negatives as positives, 16-bit (well, 14 anyway)
and then "invert" in Photoshop. It's no big deal
at all.

If you get a good raw scan, you've captured
all there is to capture and anything else can be
done in Photoshop. A good raw scan is the thing.

All this fuss over film profiles always struck me
as strange.


rafe b
www.terrapinphoto.com

 

[Bart van der Wolf]

SNIP

I never found them useful at all. Regardless of
what scanner or scanner driver I'm using, I scan
negatives as positives, 16-bit (well, 14 anyway)
and then "invert" in Photoshop. It's no big deal
at all.

There is, however, a bit more to scanning Raw and inverting.

If you get a good raw scan, you've captured
all there is to capture and anything else can be
done in Photoshop. A good raw scan is the thing.

Agreed, but that requires 'optimal exposure' (maximum per channel
exposure, without clipping and while 'removing' the film's base color,
especially important for color negatives).

All this fuss over film profiles always struck me
as strange.

Actually, after 'correcting' the per channel exposure levels, there
still are residual incompatibilities between the R, G, and B curve
shapes *and* intensities (saturation). Profiling potentially takes
care of those anomalities, but it is not all that simple (even
shooting a target object such as an 'IT8' requires skill). Having
created a number of profiles myself, for various scanner
illumination/sensor-sensitivity versus film dye set families myself,
it makes arriving at a plausible baseline much easier. Then there
still is the tone-mapping challenge to fit the film's (partially
compressed) representation of the scene's luminances and saturations
to the potential output media.

IMHO, there's enough to 'fuss' about ;-) .

Bart

 

[Don]

I use VS to scan my archive. Most of them are negatives, some of them
are slides. I use a Nikon LS-5000 on a Linux box. I initially decided
to save them as JPEGs with 2000 DPI. This gives me approx. 2000x3000
pixels, enough for printing and for the web. One nice feature I like in
VS are film types.

Until now I was quite happy, but I encountered some film types *not* in
the list. ....
Then I read some articles on the net and in this NG it would be wise to
scan "raw" and do the editing later. VS can do that (I scan with light
infrared clean, medium grain reduction and 4 times multi-passing) but
images get huge (30 MB with 2000 DPI, 120 MB with 4000 DPI) and I still
don't know what film type to use. :-\

As others have pointed out "film profiles" are of very limited use to
start with. The same goes for scanner profiles.

In a nutshell, film profiles are only an approximation. Each film will
differ for a number of reasons such as different batch, different lab,
different chemicals, etc. Scanner profiles try to compensate for their
bias, but sometimes this bias goes in the right direction. So a
scanner profile may actually cause corruption which you will then have
to remove later.

Profiles are only useful for editing i.e. monitor profiles, so you see
what's really in the file, and for printing i.e. to compensate for
various differences between the screen and the printer.

As to scanning raw, there are a number of issues. First of all Gimp
(as far as I know) is only 8-bit so you will not be able to even edit
raw 16-bit images. But more importantly, if you're happy with JPG and
2000 DPI that's way below the capabilities of the scanner. So scanning
raw may be an overkill because it will produce way too much
information you are not really interested in. Also, that means that
many of the Vuescan bugs will be "below the radar" although you may
still occasionally suffer from "0-byte" file and other more basic
bugs.

Don.

 

[Raphael Bustin]

Actually, after 'correcting' the per channel exposure levels, there
still are residual incompatibilities between the R, G, and B curve
shapes *and* intensities (saturation). Profiling potentially takes
care of those anomalities, but it is not all that simple (even
shooting a target object such as an 'IT8' requires skill). Having
created a number of profiles myself, for various scanner
illumination/sensor-sensitivity versus film dye set families myself,
it makes arriving at a plausible baseline much easier. Then there
still is the tone-mapping challenge to fit the film's (partially
compressed) representation of the scene's luminances and saturations
to the potential output media.

IMHO, there's enough to 'fuss' about ;-) .

If I needed "objectively accurate" color or anything like that
then I'd shoot chromes, in a studio, with controlled lighting,
and the scanner would be profiled. (Or I'd shoot with digital
and profile the camera.)

But I don't. Need objectively accurate color, that is.

Results need to be pleasant and believable, is the
main thing. Or consistent with the mood and content
and intent of the image.

If analog gains are set right, and white/black points
set for each channel after the scan (assuming no
clipping in any channel) then usually what's left is at
most a few minutes tweaking the curves - either
the overall gamma, or per-channel.

With C41, there is no such thing as objectively
accurate color. It's impossible. Doesn't exist.


rafe b
www.terrapinphoto.com