Home of Professional Cinematography since 1996

Still Scans

Published : 15th September 2003


>I like having a slide scanner around.

Without wanting to ask a dumb question like "what's a good camera", I'd like to ask a similar one : what's a good film scanner?

Either - the best (the Canon seems to be popular) or the best value for general work or the fastest - and what do you rate most highly? - dpi? colour depth or consistency? cost? speed?

It probably doesn't need to be 6k for my needs at present so I've breached the topic area for this list already,

Geoff…Sorry…(Wanna fight?) but it seems appropriate to get the question in quick.

Dominic Case
Atlab Australia



Dominic Case wrote:

>what's a good film scanner?…and what do you rate most highly?

For a desktop scanner I would look for :

Large dynamic range - 3.7 for retaining highlight details in neg
Bit depth - 12 to 16 bits per channel
Pixel multi-sampling - to reduce CCD noise
Colour management - with a choice of ICC colour gamut profiles high DPI
Dust & scratch removal - 'Digital ICE' is by far the best.

Mark Harmon
Digital Film Technician
Animal Logic Film
Sydney, Australia
http://www.animallogic.com



Dominic Case wrote :

>Either - the best (the Canon seems to be popular)

If I had to scan 90' of film per running minute (to fulfil the list requirement), I'd want to choose the scanner with :

1) The best applicable resolution
2)The scanner with the appropriate software.

The Canon software is okay, Arclight and Vuescan are better, but the Nikon software really is a step beyond them all. I have both the Canon & Nikon 2700dpi scanners currently and am looking forward to moving into the 3000-4000 dpi range. The Canon is faster, it moves the film past the fixed scanning head, the Nikon moves the head past stationary film and qualitatively produces better looking scans from my chromes & negs.

IMHO,

Michael Vitti



Michael Vitti :

>The Canon & Nikon 2700dpi scanners currently and am looking forward >to moving into the 3000-4000 dpi range.

If you want to spend almost no money then get a HP photosmart that will do 2400 for almost nothing.

The results are pretty good.

I use it for scanning the half frame shots that I take with my PL mounted Pen F, the very poor mans time-lapse, reassembled into 3 second sequences in Commotion.

Cheers

Geoff Boyle FBKS

Director of Photography
EU based
www.cinematography.net



I would suggest that of the items on Mark's list, dynamic range and bit depth are the most critical.

I have a Minolta Dimage Multi scanner that's a few years old now...2800 dpi from 35mm negs. I recently took some transparencies over to a local lab that just got a new Nikon scanner (8000 dpi?)…Don't remember exactly.

The increased resolution wasn't all that spectacular. (Kodachrome 200 and Velvia originals...with an Agfa Optima B&W transparency thrown in for fun). But wow, did I notice the improvement in dynamic range. Aesthetically I think that makes a bigger difference than resolution, when it comes to reproducing the image you see on the transparency.

Also, the Digital ICE did make for some pretty clean scans - that would save a lot of retouching time. (They used it in an automatic mode, and there were very few spots left that I would have to retouch. I have yet to try it out with a thin negative, which would be a good dirt-removal test.)

George Hupka
Director/DP, Downstream Pictures
Saskatoon, Canada



I have 775 ish pictures and will need to keep 730 ish. It’s about $30 per roll @ $1.00 per scan at 12MB per picture at ISGO's.

I don't know what type of scanner they use but I am footing the bill until some curator of a gallery is interested in having a showing of the pictures somewhere…but I doubt it.

Please educate me here…

Vistavision scanner Genesis or any other Motion picture scanner vs a Still image scanner as mentioned by others that was intended for doing slides or neg scans and can the genesis save the scans to a Hard drive and write to DVD as well as write to a Datatape format DTF, DLT, D5 or HDCAM whatever.

Please forgive the elementary nature of these questions.

Geoff, thanks for setting up this new list and I hope these kind of questions will not be thwarting your original intent.

Once I get a handle on this issue I will understand much more about how the process works and will be able to follow more closely the post and CGI folks as these issues get discussed.

B. Sean Fairburn
LA DP
Love to learn



George Hupka wrote :

>The increased resolution wasn't all that spectacular (Kodachrome 200 >and Velvia originals...with an Agfa Optima B&W transparency thrown in >for fun).

Its interesting that you noticed the difference with various slide films.

It just occurred to me that the need for a large dynamic range would only be necessary for scanning transparencies. If you are only scanning negative, a dynamic range of at least 2.5 should be more than enough.

Even though negative contains a larger range of scene brightness, the densities on the neg film are of much lower contrast compared to print transparency's.

Have I missed anything here ?

Mark Harmon
Animal Logic Film



B. Sean Fairburn wrote:

>and can the genesis save the scans to a Hard drive and write to DVD as .well as write to a Datatape format DTF , DLT, D5 or HDCAM whatever.

With the Genesis we scan everything as a 10bit log cineon file but we can convert that to 16bit log tiff file that can be transferred to CD, DLT, DTF,AIT.

16bit log images look very flat so they need to be graded but at least you know you have all the shadow & highlight information, and from there you can convert to linear images with a soft clip in the highlights etc.

Similar to what happens with the Thomson Viper.

Mark Harmon
Animal Logic Film



>If you are only scanning negative, a dynamic range of at least 2.5 >should be

Mark,

The figure 2.5 represents what ? Is it 2.5 stops ?

Many Thanks

Tom Gleeson D.O.P.
Sydney Australia
www.cinematography.net



Tom Gleeson wrote:

> The figure 2.5 represents what ? Is it 2.5 stops ?

Your really testing me now…

Measuring dynamic range of a scanner is related to how well it can reproduce the fixed density on film.

Basically it is the highest density minus the lowest density it can reproduce.

You could call it a reproducible density range.

1 stop = 0.3 density (only between the straight line of the neg’s characteristic curve).

A dynamic range of 2.5 = (2.5 / 0.3) stops 2.5 dynamic range = 8.3 stops (that can be reproduced). Its actually more than 8.3 stops because of the compressed range in the highlights & shadows.

Mark Harmon
Animal Logic Film



>Its interesting that you noticed the difference with various slide films.

It's the result of having a friend with a Cibachrome processing machine, which can be really quite addictive.

Your analysis makes sense to me. Although I've often wondered why printers still seem to prefer transparencies. I'd think that a neg would give them a bit more range to work with.

George Hupka
Director/DP, Downstream Pictures
Saskatoon, Canada



George Hupka wrote :

>Although I've often wondered why printers still seem to prefer >transparencies I'd think that a neg would give them a bit more range to >work with.

Well, a transparency looks like it's supposed to look, so all a printer needs to do is capture what's on the slide. A negative is so open to interpretation that a printer probably has to go through more test prints to get something perfect that satisfies the customer.

Jeff Kreines



>Well, a transparency looks like it's supposed to look, so all a printer >needs to do is capture what's on the slide…

Hehe...

So, we're shooting a modest local TV commercial....Client (a small local print agency doing its first TV spot) wants some stills for a newspaper campaign at the same time, can I look after it? I call in a friend of mine to do the stills. (Hasselblad, transparencies requested by the agency.)

I get a call from the agency a couple of weeks later...the printer says the stills are overexposed and unusable. I've seen the transparencies, and I know they're just fine.

So I go to the agency where we look at the transparencies on a light table. They're just fine. The client says, "they must be overexposed because we use this printer all the time and he wouldn't say they were overexposed unless they're overexposed."

We pull up the scans the printer delivered and they are the most awful scans I've ever seen, not even close to the transparency.

Client : "You must not know how to expose properly for images to be scanned."

I take the transparencies home, scan them in presets on my scanner and send the agency a CD-ROM.

Client : "How long did it take to fix the scans? We're not paying you extra for that, the transparencies should have been properly exposed in the first place."

They've never called me back. I've never lost any sleep over it.

George Hupka
Director/DP, Downstream Pictures
Saskatoon, Canada



>Although I've often wondered why printers still seem to prefer >transparencies

Grain seems to be the issue. Transparencies have vastly tighter grain than negatives.

Some of the newer negative still stocks are being touted as "designed for scanning”…finer grain, protective coatings, more latitude. It's only a matter of time. A number of magazines already accept digital images, they don't particularly care what they originated on as long as they'll print up to 8.5x11.

Art Adams, DP
Mountain View, California - "Silicon Valley"



Mark Harmon of Animal Logic Film asked

> Have I missed anything here ?

You got it right. A projection-contrast original is likely to have a density range of over 3.0 (1000:1), whereas a color negative film has a much lower gamma, so the original scene information is contained within a much lower density range, typically only about 1.20 density from shadows to highlights for a typical scene :

http://www.kodak.com/US/en/motion/support/h61/

http://www.kodak.com/US/en/motion/support/h1/structure.shtml

John Pytlak
EI Customer Technical Support
Research Labs, Building 69
Eastman Kodak Company
Rochester, New York 14650-1922 USA

Web site : http://www.kodak.com/go/motion