Olympus-OM
[Top] [All Lists]

Re: [OM] New E-1 (noise reduction)

Subject: Re: [OM] New E-1 (noise reduction)
From: Joe Gwinn <joegwinn@xxxxxxxxxxx>
Date: Sun, 21 Sep 2003 12:45:01 -0400
At 1:34 AM +0000 9/21/03, olympus-digest wrote:
>Date: Fri, 19 Sep 2003 19:42:58 -0700
>From: Jim Brokaw <jbrokaw@xxxxxxxxxxx>
>Subject: Re: [OM] New E-1
>
>on 9/19/03 2:39 PM, W Shumaker at om4t@xxxxxxxx wrote:
>
> > Note that noise is most often measured on a flat gray area. Many noise
> > reduction schemes can be done to flat areas of tone. Consider what Neat
> > Image can do. What would really be telling is if all the cameras could
> > be measured for noise before any processing. This would tell you what
> > the real sensor and electronics are capable of, rather than what
> > tradeoffs were made between sharpening, noise reduction, etc. Noise in
> > the presence of detail is more interesting. How to measure, I'm not sure.
> > 
> > Wayne
>
>Would the basis noise level be evident in a 'raw' file...? Most of the newer
>prosumer-grade digicams seem to be offering ability to save in 'raw' format,
>would this be available for inspection without processing out the noise
>information, perhaps in Photoshop? Would a good test be to image a gray
>card, then examine the 'raw' file that results for any non-gray pixels in
>the gray area? Gross count or 'noise bits' percentage could be figured.

Any noise-reduction scheme will blur the image, if we are talking about random 
noise.  Sharpening schemes increase the noise, so one often blurs a little to 
reduce noise, then sharpens a bit.  Depending on the details, the two 
operations don't completely cancel, but it's a bit of a black art.

Pattern noise (variations in pixel sensitivity, for instance) is the same from 
image to image in a given camera, and so can be calibrated out, but always at 
the expense of increasing the random noise.  One can make the increase minimal, 
but there is always an increase.  If the random noise is low enough to start 
with, the increase will not be a problem in practice.

A good measurement is to photograph a uniformly-lit surface (which need not be 
gray) and compute the average and standard deviation of the values of the 
pixels in a patch containing at least 50 pixels.  Do this computation three 
times, once per color.  The larger the ratio of the standard deviation to the 
average the greater the noisiness of the image.  This is a standard way to 
measure such things.

Joe Gwinn


< This message was delivered via the Olympus Mailing List >
< For questions, mailto:owner-olympus@xxxxxxxxxxxxxxx >
< Web Page: http://Zuiko.sls.bc.ca/swright/olympuslist.html >




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