Olympus-OM
[Top] [All Lists]

[OM] Re: Film vs. Digital @ equivalent 9 x 6 print then cropped and en

Subject: [OM] Re: Film vs. Digital @ equivalent 9 x 6 print then cropped and enlarged
From: W Shumaker <om4t@xxxxxxxx>
Date: Sun, 15 Aug 2004 13:27:49 -0400
Yes you don't get more image detail in high res scans, but because of
the grain and the sampling nature of film scanning, you need to scan to
the point where all you are getting is more grain structure in order to
avoid aliasing issues that would affect resolution. It does not mean
the resolution of a 22mp film scan is 3-4 times that of a 5-8mp digital
camera. Rather high scanning ppi is necessary for film to fully
preserve its character. The high spatial frequency caused by the grain
requires over sampling. Otherwise the film needs to have an anti-alias
filter applied first (that is blur the image) before scanning at a
lower ppi resolution. The alternative is to scan at high ppi, then low
pass filter and down-sample digitally. Both will loose some
information. Most DC's already have such a filter before the sensor.

The best way to compare film vs digital is to sample film at the
highest scanner resolution, then up-sample the digital camera image to
the film scan resolution and compare the two. Otherwise you are putting
film at a slight disadvantage. Grain does not have a uniform spacing
like pixels do, while up sampling the digital image does not loose
information. In addition, I would suggest using something like
http://scanhancer.iddo.nl/ to help de-collimate the scanner light to
further reduce grain scanning artifacts. This would get the result
closer to that of an ideal projector.

\A/yne - I apologize if I missed a post here and there lately...

At 08:12 AM 8/15/2004, you wrote:

>Not being a professional or a scientist or anything, I should probably 
>shut up, but my own impression is that you don't get a lot more useful 
>information out of film by scanning it at super-high resolutions. Yes, 
>there's more there to be seen at higher magnifications, but not really 
>image data. It seems that you're just getting a clearer look at the 
>grain structure past a point.
>
>On Aug 15, 2004, at 1:00 AM, Simon Worby wrote:
>
>> Again, I don't think scanning a negative at 4000 dpi does film any 
>> justice at all (please see previous e-mails). It's simply not a level 
>> playing field.


==============================================
List usage info:     http://www.zuikoholic.com
List nannies:        olympusadmin@xxxxxxxxxx
==============================================

<Prev in Thread] Current Thread [Next in Thread>
Sponsored by Tako
Impressum | Datenschutz