For information, I have made a short comparison between two inexpensive flatbed scanners for 4x5", the 'old' Epson Perfection 1200 Photo and the new Canoscan D2400UF, and one film scanner, the Polaroid Sprintscan 45.

The results are presented on the following web page: http://elbereth.obspm.fr/~kervella/Scanners/index.htm

As I have been helped a lot by the contributions of the members of this forum for my first steps in large format, I believe it is natural that I bring also my two cents :-). As a general summary, I find this scanner a very good piece of equipment, especially considering its price.

Of course, I would apreciate any comment you may have on these results.

-- Pierre Kervella (pkervell@yahoo.com), September 25, 2001

Answers

I'm intrigued by these results. There actually appears to be no difference in sharpness between any of them, only a change in digital resolution. 4000dpi should be enough to resolve the grain of any film, yet I see no sign of any grain structure in the 4000dpi scan.
The scan from the Epson 1200 is also at odds with what I was able to obtain from one. There was a distinct flare, or blooming, on mine which made the poor optical resolution look even worse. There's also been noticeable flare on other samples I've seen from the D2400U. Perhaps colour scans look different? Any chance of anything other than B&W samples, Pierre?

-- Pete Andrews (p.l.andrews@bham.ac.uk), September 25, 2001.

From a 1200 x 2400 flatbed: http://mmebers.aol.com/forgeniuses/I/FarmWEB.jpg

-- David Stein (DFStein@aol.com), September 25, 2001.

there's not point to make a scan (from 4x5) bigger than 300Mo because 300Mo, 8 bits, TIFF, RVB it's about all the details available on your film, assuming that the resolution of the scan is a true optical resolution !!! Second point, i've heard that there is a lot of trouble when using the canon on a Mac G4, any comments ?

-- dg (sacripant@online.fr), September 25, 2001.

I have added a number of new tests including color and processed images. The adress is the same as before:

http://elbereth.obspm.fr/~kervella/Scanners/index.htm

-- Pierre Kervella (pkervell@yahoo.com), September 25, 2001.

My college purchased a polaroid 4x5 scanner which we than returned when we discovered that the film holder cropped a full 3/8 inch off the 5 inch length of the film. Didn't matter how highly resolved the scans were.

-- David Goldes (david_goldes@mcad.edu), September 25, 2001.

David, are you saying that the Polaroid scanner cropped the useable image area by 3/8", or that it was 3/8" short of 5"? The exposable length of a 5x4 is only 120mm, about 4.72", and this is only 1/10th of an inch away from 4 and 7/8". 5x4 film isn't fully 5x4 inches in size.

-- Pete Andrews (p.l.andrews@bham.ac.uk), September 26, 2001.

I've just looked at the specs for Polaroid's scanners, and that explains a lot.
The scanner that Pierre was using must have been the discontinued 45i, which had a 4000 x 2000dpi resolution. So we're not comparing a 2400dpi scanner with a 4000dpi scanner, but with a 2000dpi scanner interpolated up to 4000. No wonder there's very little difference!
The new Polaroid 45 ultra only claims a flat 2500 dpi.

-- Pete Andrews (p.l.andrews@bham.ac.uk), September 26, 2001.

I was told when using the Polaroid scanner that it was truly a 4000 dpi model which, considering its price, was possible. I did not check by myself on the web, but I should have.

Nethertheless, the comparison is still valid, in the sense that the Canoscan D2400UF delivers scans that are similar to the Polaroid 45, which is 2000 dpi optical resolution. I find this already impressive considering the price difference ! :-)

As a general point, the performances of the scanners are currently following an "asymptotic" curve with respect to price. This means that above say $1000, the improvement in scanning quality is very small between models of vastly different prices.

It is quite clear that the low-level consumer models are bad quality/low price and the high-end scanners (Imacon, Tango,...) are top quality/unbelievable price. The compromise in-between seems to me the best way to go currently, and the scan quality, while not identical to top-notch scanners, is in the same order of magnitude.

A point I have not adressed on my web page, and I plan to do it soon, is the Dmax of the three scanners or more precisely their ability to "see in the dark". As the image I have is from RVP, I will have plenty of density range to test the responses of the three devices. The quality of the shadows produced by the D2400UF already impressed me visually, but I want numerical values before reporting on it. As an indication already, the D2400UF "sees" easily through unexposed (and processed...) RVP, which is reputed for its "infamous Dmax" (from another post in this forum).

Also, I plan to do a more thorough analysis of the true resolution of the scanners using Fourier transformation (which is not possible in Photoshop unfortunately). I hope this will give an objective measurement of the modulation transfer function of the three scanners.

-- Pierre Kervella (pkervell@yahoo.com), September 26, 2001.

Having used the Polaroid 45 at work, I wouldn't suggest that one AT ALL! It does crop your image, it does ruin film once in awhile and it did crap out ALOT! Spend your money wisely and don't go this route!

-- Scott Walton (scotlynn@shore.net), September 26, 2001.

I have added a section on the Dynamical range, Dmax and Dmin of my two current scanners, the D2400UF and the FS2710, both from Canon. You are kindly invited to visit this web page :-).

The adress is the same as before (see above). There is a BIG surprise from the D2400UF ! :-)

-- Pierre Kervella (pkervell@yahoo.com), September 26, 2001.

The direct adress to the new page is:

http://elb ereth.obspm.fr/~kervella/Scanners/dmax.htm

-- Pierre Kervella (pkervell@yahoo.com), September 26, 2001.

I have completed my series of tests with the measurement of the modulation transfer function (MTF) of the Canoscan D240UF. Its resolution is really 2400 dpi for a modulation of 50%, and it is probably even higher optically but limited by the detector sampling.

There is also a paragraph on a suggested "optimal unsharp masking" procedure for this scanner, considering its MTF.

All comments will be appreciated :-).

-- Pierre Kervella (pkervell@yahoo.com), September 28, 2001.

Thanks for all of your work Pierre.

I too have an Epson 1200 Photo and have struggled for sharpness. Noticing a distinct improvement the larger my neg size.

I now struggle with Newtons rings with 4x5 B+W negs (and some colour). Am experimenting with making some new neg holders and working out how to tension the neg to prevent the sag. Blu-tack is am obvious, but inelegant candidate.

Are the neg holders any better with the Canon. I'm very impressed with the performance.

-- Baxter Bradford (baxter.bradford@ntlworld.com), October 01, 2001.

The film holders of the D2400UF are in my opinion very well designed. The film is not touching at all the scanner window, and I have never seen Newton rings even on 4x5". I suppose this is one of the most important parameters to have good quality scanning.

With my previous Epson Perfection 1200 Photo, the Newton rings were really a nightmare, even with medium format film.

-- Pierre Kervella (pkervell@yahoo.com), October 03, 2001.