From: "Mark Hahn" <markhahn2000@...>
To: <DigitalBlackandWhiteThePrint@yahoogroups.com>
Sent: Friday, January 09, 2004 7:48 AM
Subject: [Digital BW] Epson 2450 - negs on glass [was:Re:
Scanning Tri-X]
here's and article showing why your probably *don't* want to lay
your
negatives on the glass:
http://www.normankoren.com/Tutorials/Epson_flatbeds.html#Focus_spacing
(see Focus and Spacing)
This made a big difference in the quality I get out of the
scanner.
mark
Mark,
For the best film scanning with the Epson 3200 I have made a
variety of filmholders. There are several issues with that
scanner to consider. I have described the carrier I'm using now
and why it is used that way in some messages on the Epson 3200
mailing list. A summery is included in this message.
For anyone who thinks of a scanner purchase: there's a new Epson
4870 based on the 2450/3200 arriving soon with ICE etc included
and a wider transparency window. Epson claims an even higher
resolution and better Dmax.
Related pages of earlier Epson carrier/methods, Nikon 8000
filmholder:
http://members.chello.nl/e.dinkla/EpsonCarrierSolutions.pdf
http://members.chello.nl/e.dinkla/MF_FM_carrier.htm
Summary:
The
webpage doesn't represent the last stages of "carrier" concepts
anymore. I've written a message what I'm doing now and that is
quite easy to replicate. Right now I try to make a page with some
information from a scanner test published in C't and there I will
add the other "carrier" with some pics. For the moment a quote of
the other message:
The method I'm using now is that I make a mask of 1.7 mm thick
(I'm still trying to find an even nicer focus) polycarbonate,
matte black paper underneath for every film format that I
encounter: MF lengthwise, 665 Polaroid film, 4x5" film and what
else I will get. All formats in the middle of the scanlength and
at the start the initialisation window for the scanhead. A
separate white (transparent, this is a commercial name for
almost colourless windowglass) floatglass plate of 2 mm thick is
used to wetmount
the film on, larger than the scanwindow but smaller than the
mask. Most of the time an extra sheet of transparent polyester
foil larger than the film over it. Antistatic brush etc highly
recommended for all surfaces. Mounting fluid dripped in between
with a pipet and squeezed out with a polyethylene squeegee. Thick
paper underneath the glass to even out the pressure and to absorb
the fluid around. When I squeeze out the fluid I use an extra
piece of polyester to avoid scratching the first foil or the
film. Clean tod to wipe away excessive fluid. Check the sandwich
on both sides in reflected light to see if there's an airbubble
on the wrong spot.
Quite often a second mounting works better.
Then the glassplate sandwich goes on top of the mask in the
scanner but with the film at the underside. Hanging underneath so
to speak. The emulsion will be approximately at the focus
distance. The film is plane. The glassplate keeps most of the
heat away. The fluid gives a much better light transmission,
especially with matte emulsion surfaces. Films that have had a
rough treatment in the past benefit a lot by this method as well.
With the polyester foil there's hardly any evaporation of the
fluid so long scans are possible.
By using a separate glassplate to mount on it is easier and less
messy than doing it on the scanner glassplate itself. Aligning it
to the maskwindow isn't difficult either as the glassplate is
smaller so it can be positioned at an angle.
There's one thing that may improve this method even more:
replacing the glassbed of the scanner itself with a mask of metal
sheet that leaves the widest scan opening free. Two surfaces less
in the path. I am yet not mentally prepared enough for that step.
Materials used:
Focus measuring:
Small glass rectangles from slide mounts. Superglue. Some
dissolved black acrylic paint or ink. A spiral staircase with the
inkspots at the circumference is probably the best design. Five
or six pieces will get you at 4mm height.
Mask:
Any transparent plastic sheet like PVC, Polystyrene,
Polycarbonate will do. The kind that you cut a line in with a
Stanley knife at both sides and then tear the parts out. Scrape
the edges with the knife too. Black paper and glue or two sided
tape. That's applied in one or more layers to the plastic sheet
to get the light blocking good and the focusing distance right.
In scanning the paper will be on the underside. Use the original
carriers as a drawing mask for the initialising window + the
total size possible for the scanwindow and draw the right size
you want in the middle of all that so the glassplate can be be
rotated within the rim of the scanner surface. I've started from
1.3 mm thick PC and build it up with a 0.15 mm thick combination
thick black paper + two sided tape (have that in sheets) but glue
will work as well.
Glass:
2 mm white window floatglass is excellent. Buy more sheets and
check on air bubbles, scratches, tin spots on the surface. Use
sandpaper to soften the edges. Two good sheets will make
continuous scanning easy, mount one while the other is used in a
scan. 25 x 16 cm is a nice size with enough overlap.
Mounting fluid:
Kami SXL 2001 (Aztek distributes it in the USA) or lighter fluid
that I use. Kami is heavier, less evaporative and will need more
time afterwards. "Lighter fluid" means that it can burn, explode
and isn't free from carcinogenic effects. Usually a comment like
this is followed by comments of people that feel more
responsibility for mankind than I have. This is all I write about
that issue.
Polyester foil:
Mylar as the Dupont name, PET when they make bottles of it. Art
supply stores, Aztek, virgin Copyproof film sheets are quite
usable, prepress mounting sheets etc. Exists on rolls, in sheets
and as blocks of sheets. 0.10 mm thick is good.
Polyethylene squeegee used for adhesive marking foil etc or a
true pressure roller. I prefer the first.
Ernst
(added)
Kimoto used to have a wide range of cleaning equipment, foils,
fluids
for prepress but it looks like Policrom has taken over part of
that:
http://www.policrom.com/Scanner_Prep____Cleaning/Scanner_Prep_/scanner_prep_.html
I've had good experience with their products in Europe.
Prazio is already mentioned: www.prazio.com
And if you want to try fluid mounting without excessive costs:
lighter fluid and some sheets of transparant polyester from an
art supply store. 0.05 to 0.10 mm thick. Some tape and a clean
cloth. It is "lighter fluid" ....... "lighter" ....... so be
careful.
(other message)
3200 scanner + other tested:
http://members.chello.nl/e.dinkla/Verhaal.pdf
> > Dichtheidsbereik
Dynamic range, range available in one sampling.
> Maximale dichtheid
Dmax
> Signal/ruisverhouding
Signal to noise (at 2.85 D), this is interesting as so far the
S/N was used to qualify what the Dmax was. In the sense that when
the noise became too much compared to the information the sample
was taken as just beyond the Dmax. It seems (but I do not know
yet) that the new standard is taking the S/N at 2.85D as a
separate result and by that it will be easier to quantify the
reading and make comparing easier. I have written the author but
got no response so far. Dmax now probably is the last step before
no detail is visible, that was used in the past as well when the
S/N wasn't already getting too low.
> Dieptescherpte
Depth of focus
> Kleurfouten
Colour deficiensies in delta E.
> Helderheids-schommeling
Lighting evenness. I have asked what is actually measured there.
Per scan, in the scan path or between scans? The Epson isn't
scoring well there.
> Pixelvershuiving
Pixelshift, the true dimension scanned I guess. There's one bad
result in that test and the author mentions that scanner as
having a smaller scan than the image is. Will ask if I get a
reply on the other questions.
> Werkelijke resolutie
Real optical resolution
>
>
(other message)
Don't wait for Epson information where the optimum focus should
be. Could be that they don't know (manufacturer's tolerance :-)
but
it is sure that they will not tell you where it is.
I do not agree with your assumption that it should be halfway the
filmholders centre and the glassplate. If I had had a hand in the
design I certainly would have made the focus at the filmplane = 1
mm above the glass. There you need the highest resolution and you
will scan with the highest resolution. For scanning in reflective
mode you usually do need 300 ppi at most and 720 ppi if it has to
be enlarged. Photographs scanned do not deliver more at higher
resolutions.
The scanner has three planes that are used for scanning: the
glassbed, filmcarrier and the top of the scanner (with the lid
removed). Epson could have chosen to take a focus point halfway
the two extremes which is about 1.5 mm above the glassbed. Still
close to the film position in the normal carrier and covering the
two extremes. C't gives the 2.9 mm number to the DOF which
corresponds with that idea as there's about 3 mm between the two
extremes. Combine that with some manufacturer's slack in
tolerence and it will be between 1 and 2 mm for the focusing
distance.
Just measure it, it isn't hard to do and you get the number that
belongs to your scanner.
(other message)
>Interesting results. I think it just proves once again the
amount of variation between scanners of the same model and
production run.
The focusing spots on the 2450, 3200 models vary as much as 1.2 -
2.4 mm above the glass (Norman Koren's pages and links).
My measurement says 1.7 mm and Julio gets below 1 mm.
On depth of field I have seen 2.4 and 2.9 mm. That's an arbitrary
number as the qualification has much to do with the sharpness
accepted. If Julio's focus is that low on the bed he can't
measure the DOF that has been reported by others as 1/3 will be
under the glassbed.
A further testing and reporting by other 3200 list members would
be nice to see what the spread actually is.
I've had problems in the past with multi-pass scans. Using
another method I'm sure that the main problem must have been the
heat of the lamp.
Now I've made a very nice two sample long exposure scan of 665
film with Vuescan that was as sharp as a single scan and with
less noise, more tones.
The long exposure is a 2 pass with different exposure times so
the sampling is 4 in total. To overcome the heat etc I needed the
wet mounting method I've described before.
I would never do a 16 pass though, 2 maybe 4 is enough. There's
hardly any gain with more passes than that.
End of summary