3/12/97
going to buy a flat bed scanner, probably a Microtek with a transparency
adapter. We would primarily use it for scanning in TEM and SEM (polaroid)
negatives. Should we go for the 24-bit color, 300 x 600 dpi version, or
the more expensive (ouch) 30 bit color, 600 x 1200 dpi version? Our
resident computer guru says that we should go with the cheaper version
because we are unlikely to have an output source with better than 600 dpi
for most stuff. It's true; the printers around here are mostly 300 - 600
dpi. Is there a compelling reason for the better model? Will I be really
sorry in a year if I don't?
Thanks you all for all the expertise and advice on so many different
topics!
Aloha,
Tina
http://www.pbrc.hawaii.edu/bemf/microangela
****************************************************************************
* Tina (Weatherby) Carvalho * tina@pbrc.hawaii.edu *
* Biological Electron Microscope Facility * (808) 956-6251 *
* University of Hawaii at Manoa *
24 bits will not provide. Now, you may ask 'my software can only
handle 24 bits so what good does 30 bits do when I will have to
convert it down?'. When you convert the file to 24 bits, the system
will choose from the best 24 bits to produce the converted file. I
would equate it to a survey or poll - the larger your sample, the more
accurate your results.
John D. Warren
Area Sales Manager
Digital Products
Polaroid Corporation
WARRENJ1@cliffy.polaroid.com
24 color bits gives one 256 grey shadows.
If one want only to store and reproduce images - this is more than enough
(128 is O'K for viewing).
If one want to quantify grey levels for processing, 256 levels is also
usually enough because of narrow linear range of film, grain, etc.
Interpolation will give better results.
If one want to waste money - I have no reasons not to this :)
Andrey L. Chuvilin
dusha@catalysis.nsk.su
old HP-IIcx here that easily beats a newer cheaper scanner hands down. The
software is also not as convenient as HP's deskscan package.
Regarding resolution, remember it will take a few of those printer pixels to
dither up the gray scale for each of your digitized pixels. Thus, a 600 dpi
printer may only be able to handle about a 200 dpi image. Of course if you
are scanning 35 mm slides and enlarging them, you will need the high res
scanning.
Warren Straszheim
wes@ameslab.gov
afford - not so much for the SEM stuff but for the TEM negatives. The
scanner should have enough optical (not interpolated) resolution to over
sample the film resolution by at least a factor of 2.5 to 3 (the Nyquist
sampling limit you know) otherwise you will not be able to treat the
digitized images like the original negative - i.e. information will be
lost! The company that makes tha scanner probably also makes a 1k x 2k
version which would be even better.
Best Regards - Bill Miller
microbill@MOHAWK.NET
I have found that what is 'adequate' for one opportunity may not be
adequate for another. In any event, for those who are unsure of a
decision between digital products based on varying specifications, I
suggest you look at the end result. Take an image and scan it in on a
24 bit scanner and a 30 bit scanner, print them out on the printer(s)
you would typically use and compare the results. If the 24 bit image
is sufficient for your application, then buy the thing. The street
price should be about $220. If you like the 30 bit converted to 24
better, then buy it. Its street price should be around $525.
As far as the comment regarding scanning 35 mm slides, I would not
recommend using a flatbed scanner to scan 35mm images that will be
enlarged more than 1x the original size, otherwise the image gets soft
even with a 600 dpi scanner. If you are scanning 35mm on a regular
basis, you should use a film scanner.
John D. Warren
Area Sales Manager
Digital Products
Polaroid Corporation
WARRENJ1@cliffy.polaroid.com
get you more contrast discrimination. Another reason is that, even with
300-600dpi printers, you might want to select a subregion of the scan and
expand it to full page. In that case, the extra resolution of the better
scanner will show.
Kal
Kalman Rubinson
rubinsnk@is2.nyu.edu
We recently purchased a film scanner to digitize images and diffraction
patterns from 3.25"x4" negatives. Prior to this we had an old Microtek
flatbed with a transparency adapter (300 dpi). It was ok as long as we
didn't try to enlarge the image much.
I don't know how you will be using the digitized images. If all you want
to do is digitize the negatives and then print the images without enlarging
them, you don't need a high-resolution scanner. But I would recommend that
you get a scanner with a larger bit-depth.
We often need to do some significant image analysis, so we needed a scanner
which would give us very high optical resolutions and more than the usual
256 gray scale (24-bit color). We wanted at least 12 bits per color and
2400 dpi optical resolution. Our requirements may be more than you need.
I think John Warren's recommendation is good: try before you buy.
Russell E. Cook
Scientific Associate
Electron Microscopy Center for Materials Research
Argonne National Laboratory
Building 212
9700 South Cass Avenue
Argonne, IL 60439
(630)252-7194
FAX: (630)252-4798
cook@aaem.amc.anl.gov
don't! One of the uses I have found for the high resolution end of the
scanner is to check small details (eg is that cell junction really tight?)
in an EM negative without the hassle of a large print - it's amazing just
how much it is possible to see on the computer in a fraction of the time. I
have the ScanMaker 3 and am very happy with it. BUT, it is no good for
35mm, only prints and large format negatives.
Diana van Driel
Dept Ophthalmology
Sydney University C09
AUSTRALIA 2006
dianavd@eye.usyd.edu.au
You are all a font of all kinds of information! To summarize:
1) (Obviously) get the highest res scanner with the most pixel depth you
can afford - it's hard to overdo.
2) Printers that are 300-600 dpi with dither and print grey levels at
100-200 dpi; for continuous tome they need 3x3, 4x4, or 9x9
points (depending on the type of printer) for each pixel. So the higher
the resolution of the image scanned in the better the print (also
obvious).
3) Opinions vary - some say that 24 bit color depth gives 256 grey levels
which is "more than enough", while others say that 30 or 36 bit depth
gives more contrast enhancement, even when your software only handles 24
bit color. The best bet is to try out 24 and 30 bit scanners and see
which does the job for you.
4) The big problem lies in enlarging images. A 4x5 SEM negative scanned
in at 300 dpi can really only become a 4x5 print. A TEM negative scanned
in at 600 dpi can be enlarged <3x, whereas a TEM negative can be enlarged
10x photographically. This brings up the discussion at MSA '96 - learn to
take your micrographs at the magnification at which you will want to use
and reproduce them rather than enlarging and therefore gaining only "empty
magnification".
5) Remember to check for the true resolution, not the interpolated res.
of a scanner.
6) Be careful of getting ghosts from glossy Polaroids and Newton rings
from TEM negatives on some scanners.
7) None of the flatbed scanners seem good enough for 35mm negatives - buy
a film scanner instead.
8) Remember that higher res images mean larger files. WIll you be able
to store them? You will soon be needing ZIPs, JAZs, and CD-Rs. (Our SEM
students are alloted 10 MB storage space on their department accounts.
The full-size, full-res, colorized Mexican Ant on my web site is over 8
MB by itself!)
9) See "Tips & Tricks" at http://www.biotech.ufl.edu/~emcl/ for a
previous discussion on this topic.
As you can see from my "cheap frame grabber" and "cheap scanner" posts, we
are trying to operate within an embarrassingly small budget this year.
ALthough I am still fairly happy to prepare everything photographically
for myself, when I talk about image presentation and look at the bright
shining faces of my current SEM students, I know that I am not going to be
teaching them darkroom techniques. They start with Polaroid 55 pos/neg
film. One needs to make slides right away; another wants to know how to
make camera ready illustrations for a journal by next week. Most of these
students ultimately have access to some decent hardware and software, and
I need to point them in the right direction. I need about $10K to bring
my lab up to speed. Too bad I can't invite you all to a huli-huli chicken
sale!
What brought this up? Microtek 300 dpi scanners are down to $195, which I
could buy out of my own pocket. It was tempting.
Aloha,
Tina
http://www.pbrc.hawaii.edu/bemf/microangela
****************************************************************************
* Tina (Weatherby) Carvalho * tina@pbrc.hawaii.edu *
* Biological Electron Microscope Facility * (808) 956-6251 *
* University of Hawaii at Manoa *