> My father has something like 2000 slides taken when I was a young'un.
> Some of them have great sentimental value and I want to preserve them
> digitally (probably not all of them you understand! I will need to look
> at most of them though!)

I've been scanning slides for about 10 years now, the last few years
under Linux. No idiots guide from me, but I can somewhat reliably give
you the points within which you're operating.

First off, no matter how you look at it and how you do it, it's going to
cost you money. More money if you don't want to do it manually yourself
spending two weeks full time nonstop.

I assume you want a quality result, so you don't have to touch the film
ever again.

Commercial prices run from about $2.50 to $10 (manually adjusted) per
image. In other words, you're buying your own equipment, or otherwise
borrow/hire it, or form a group of people and share the cost, or some
combination thereof. (Count me in. I have a film scanner already which
I'd be prepared to make available.)

Be aware that the technical knowledge and understanding of the so-called
professionals in town is non-existant. They're nice people, but
completely clueless as soon as it comes to dividing two numbers
(essential for assessing quality). I'm not kidding.

Forget about flatbed anything, unless your film is larger than 35mm, and
the flatbed is an expensive model. I wouldn't consider it.

A good slide projector and a good digital camera, set up carefully, may
do the job. Expect to get parallex and geometric distortions in your
result. Colour control is an issue - decent scanner software can correct
for colour shifts of well-known film brands due to aging. With your
projector, you'd be shooting raw format and running it through a
self-contrived correction mechanism. Possible, takes some time (and
software) to set up.

Make sure you're sitting down (and your other half isn't looking over
your shoulder) when you look up the prices for film scanners. Actually
they've got cheap - $2000 (they used to be $4000), plus $1000 for an
automatic feeder. Buying a used one would help. There are some good
ones, and many I wouldn't look at twice, let alone pay for. They do
become obsolete too, though not nearly as fast as digital cameras.

The speed of the scanner is a top consideration unless you buy a
feeder!! Even then, at your numbers you're either getting the feeder
anyway or you have too much time.

Do not even think about unmounting the slides. You'll do more damage
than it's worth the extra millimetre of film area. All scanners can take
mounts as well as film strips. I save myself pointing out that film is
film, whether it's positive or negative is not something the scanner
ever cares about.

All film scanners are plug and go under Linux. Most flatbeds are too,
with few exceptions. Fork out for hamrick.com, or don't expect any
sympathy.

Think about storage. Remember CD-R is spelled D V D. Think very
carefully whether jpeg will do what you want, taking into account that
you don't know what you want in 10 years. Most professionals don't even
consider it for what they want today. The problem is not only the loss
factor, but also the distortions introduced by the compression algorithm
- square patterns of sorts (the compression happens in blocks of
pixels). Consider png - not as good compression rates, CPU-intensive,
but lossless. You could be looking at 50MB per photo. Fills up a
spindle or two...

I could tell a few stories about the uni film scanner... You'd need a
uni internal computer account to access it btw., i.e. be either student
or staff. The scanner is very old, no longer that good, and so worn it's
not funny. In short, forget it unless you're deperate. I used to make a
scan in under 2 minutes, but it took full-time organisation and
concentration.

The shop on Riccarton Rd is Copyland. Make sure they aren't giving you
the proprietory Kodak Photo CD(TM) stuff!!! That image format is
undocumented and there are no guaranteed functional open source programs
to unpack the top resolutions (meaning much above 512x768).

Volker

-- 
Volker Kuhlmann                 is possibly list0570 with the domain in header
http://volker.dnsalias.net/             Please do not CC list postings to me.

Reply via email to