Scanners for 4x5" comparison

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.

<p>

The results are presented on the following web page:

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

<p>

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.

<p>

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

these results.

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.<br>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?

From a 1200 x 2400 flatbed:

http://mmebers.aol.com/forgeniuses/I/FarmWEB.jpg

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 ?

I have added a number of new tests including color and processed

images. The adress is the same as before:

<p>

<a href="http://elbereth.obspm.fr/~kervella/Scanners/index.htm">

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

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, 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.

I've just looked at the specs for Polaroid's scanners, and that

explains a lot.<br>The scanner that Pierre was using must have been

the discontinued 45i, which had a 4000 x <i>2000</i>dpi 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!<br>The new Polaroid 45 ultra only claims a flat

2500 dpi.

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.

<p>

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 ! :-)

<p>

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.

<p>

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.

<p>

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).

<p>

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.

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!

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 :-).

<p>

The adress is the same as before (see above). There is a BIG surprise

from the D2400UF ! :-)

The direct adress to the new page is:

<p>

<a

href="http://elbereth.obspm.fr/~kervella/Scanners/dmax.htm">http://elb

ereth.obspm.fr/~kervella/Scanners/dmax.htm</a>

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.

<p>

There is also a paragraph on a suggested "optimal unsharp masking"

procedure for this scanner, considering its MTF.

<p>

All comments will be appreciated :-).

Thanks for all of your work Pierre.

<p>

I too have an Epson 1200 Photo and have struggled for

sharpness. Noticing a distinct improvement the larger my neg

size.

<p>

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.

<p>

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

impressed with the performance.

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.

<p>

With my previous Epson Perfection 1200 Photo, the Newton rings were

really a nightmare, even with medium format film.