----- Original Message -----
From: "Tony Sleep" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Thursday, September 21, 2000 2:28 AM
Subject: Re: Real resolution of a 4000 dpi scanner?
> Your own tests show a clear differential in MTF, although only a small
> increase in resolution. In terms of image quality MTF is *vastly* more
> important than ultimate resolution. Look it up. In fact look up the whole
> matter of resolution and MTF and their relative perceptual value.
>
The MTF defines the resolution. They are mathematically related. Photo types
with no formal training in the theory argue endlessly about which is more
important but it is a facile debate.
The USAF Resolution chart is widely used because it shows the upper limits of
the MTF curve and as long as that is in the visual pass band (and it is on all
of these scanners) it has a large perceptual impact.
I agree that further evaulating the MTF at different points is also worthwhile.
It would be nice to add such a test to the group scan tests but takes more
expertise, equipment and time than the average contributor has. If you new
site will be plotting the full MTF for each scanner then I think that's great.
Very useful.
Anyway all this is theoretical debate. Great for obfuscation but little else.
The fact is these images don't show a significant difference between the
Polaroid and the rest. The jury is still out as to why.
> Your resampling down to 3000ppi shows nothing except the in/effectiveness
> of resampling algorithms. But you are looking at the wrong parameter
> anyway.
Simply incorrect. It shows there is no difference in detail above 3000 dpi.
This means the combination of contrast and resolution of the data above 3000
dpi is too small to be signficant. Exactly the point and negates the main
purpose of buying a 4000 dpi scanner if this is indeed typical.
For this test, one just has to make sure they use a reasonable resampling
algorithm (bicubic is fine) and the original film data must indeed have the
higher detail to start with. Those are the two most common mistakes people
make with the test but it doesn't invalid the test approach. I've done this a
gadzillion times on countless scanners and it works very well.
The bottomline is that if you can't see a detail difference between a 3000 dpi
image and a 4000 dpi image then the difference isn't significant...
Cheers,
Byron
====================================================================
The filmscanners mailing list is hosted by http://www.halftone.co.uk
To resign, <mailto:[EMAIL PROTECTED]> with UNSUBSCRIBE FILMSCANNERS in the
title, or UNSUBSCRIBE FILMSCANNERS_DIGEST if you are reading the Digest.
