I am planning to buy a film scanner. A few questions popped up as I read the specs. 1. What is dynamic range? 2. How much difference is there in color fidelity between 12 bits per color and 10 bits per color? 3. Is Nikon's LED lighting unique in the under $1000 range or are they the only ones to make a big deal about it?

-- chuck fan (chaohui@msn.com), March 04, 2000

Answers

Here's my understanding:

1. The dynamic range defines the amount of contrast the scanner can resolve. Using a low density scanner will give you black shadows and white highlights without any details. So far no consumer scanner can give you the full density range of a film but some come close (those with D3.6).

2. I think there're two issues in regard to the 12 vs 10 bit color depth difference: a) for pre-processing e.g. in PhotoShop. Modifying the image characteristics will ultimately result in a reduction of colors. The more colors you have on the input side the better obviously. In the end you'll usually store/print the image with 8 bit / color. b) the color graduation should be more smooth (less noisy)

3. This is more a marketing thing.

-- Klaus Schroiff (klaus@photozone.de), March 05, 2000.

Dynamic range is similar to exposure latitude on film: it is a measurement of the difference between the lightest value and darkest value a scanner can capture. Take these measurements with a large dose of skepticism as there is no industry standard for measuring dynamic range.

Not much. Color fidelity is as much a matter of software as of hardware, if not more.

Nikon's LED lighting is unique.

One thing no manaufacturer talks about is signal to noise ratio. This is an important measurement, espeically in cheaper equipment. Unfortunately, with no published specs there is no way to compare SN of different scanners without trying them yourself.

-- Darron Spohn (dspohn@photobitstream.com), March 05, 2000.

It's not true to say that scanners can't capture the full density range of film. Take a look at at the density curve for a typical slide film:

Ektachrome 100 professional curve

You'll see that the useful density range of the slide is only just over 3, which is a ratio of 1000:1. A 12 bit/channel scanner has a theoretical density range of 3.6 (4096:1), and most manufacturers quote figures between 3.2 (1500:1) and 3.4 (2500:1). So 12 bit scanners have more than enough dynamic range for the job. A 10 bit scanner would be struggling, but thankfully 10 bit scanners are becoming the exception, rather than the rule. In short, go for a 12 bit scanner.

Beside the Imaging Resource site there are excellent scanner reviews at:

Ed Hamrick's site

and:

Tony Sleep's Scanner comparison page

PS. Strictly we shouldn't refer to the range of a scanner in terms of density, but brightness. However, since in this instance brightnesss is just the reciprocal of density, we can.

-- Pete Andrews (p.l.andrews@bham.ac.uk), March 06, 2000.

It may be that dmax 3.6 is enough to capture the full density range ... theoretically but in a recent (serious) magazine test none of the scanners came even close to that value regardless of the specified dmax 3.6 so specs are the one thing reality is something else.

-- Klaus Schroiff (klaus@photozone.de), March 06, 2000.

I agree with Klaus.... You just aren't going to get all the shadow details from a slide by scanning with any equipment you are likely to want to buy.

OTOH, I've had very good results with most slides of typical density and negatives are another issue altogether.

-- Mike Milton (mike.milton@sympatico.ca), March 06, 2000.

Mike, I didn't claim that you'd get all the shadow detail onto a scan, just the useful shadow detail. Sure, the total density range of slide film exceeds even 3.6 (and incidentally, I never made any claim that a real scanner could reach that) but the film's colour balance goes way off in the shadows.

Klaus, it would be stupid of any scanner maker to claim a dynamic range of 3.6, since the electronic analogue to digital converters that scanners rely on can't be made accurate to the least significant bit, which realistically brings the dynamic range down to a maximum of ~3.3.

The makers of my scanner make the more realistic claim of 3.2, which I've no reason to doubt. Scanning some Fujichrome Velvia slides, which have a useful density range in excess of 3.3, I was able to drag all but the very deepest shadow detail out of them. However, in order to see that detail, I had to bend the gamma curve so drastically that the overall scene looked washed out. Remember, a change from 1 RGB level to 2 represents a doubling of brightness, but most people would be very hard-pressed to see that change on screen.

And once you've got all that precious shadow detail scanned in, what are you going to do with it? You can never get it into printed form without drastically distorting the tonal range, and unless you sit and view your monitor in total darkness with the brightness full on, you'll never see it on screen either.

Consumer scanners aren't perfect, but then neither are printers or monitors.

-- Pete Andrews (p.l.andrews@bham.ac.uk), March 07, 2000.

oops - sorry Pete, I was not trying to be argumentative. Heck - I have a relatively inexpensive (HPs20) scanner and am very satisfied with the VAST majority of scans.

I even have seen reasonable results with quite dense night shots on slide film.

-- Mike Milton (mike.milton@sympatico.ca), March 10, 2000.