Yours is better but I think half of the improvement is getting rid of
the ugly reddish color cast around the eye.
Chuck Norcutt
Moose wrote:
> usher99@xxxxxxx wrote:
>> Oh,
>> I found the site that piqued my interest in the deconvolution
>> routine.(R-L iteration).
>>
>> http://www.clarkvision.com/imagedetail/image-restoration1/index.html
>>
>
> The R-L processing is interesting and, in the example, clearly a bit
> better than USM. But lets get right down to it. Anyone interested in
> this stuff, I'd appreciate your votes on the alternates here.
> <http://www.moosemystic.net/Gallery/tech/ARL_Sharp/Fox.htm>
>
> Problems I see with his presentation and examples, which also mostly
> explain the above:
>
> - He refers to an end use of a print, yet never shows scans of prints
> nor comments on any visible differences in actual prints of various
> sizes. He doesn't seem to know about, at least doesn't mention, the
> different sharpening needs for web display vs. printing. Images
> optimized for printed output will generally look over sharpened, even
> 'crunchy' when viewed on screen.
>
> - He does not compare it to more sophisticated approaches to sharpening.
> The techniques in Bruce Fraser's book "Image Sharpening", Fred Miranda's
> sharpening plug-ins and undoubtedly numerous others offer different and
> often more effective sharpening than simple uses of USM - and without
> the extreme processing overhead penalty.
>
> - He only tries one up-sampling option. There are better tools
> available. Qimage and other, more expensive and often specialized RIPs,
> may use both more sophisticated up-sampling algorithms and different
> sharpening algorithms, both optimized for printing, rather than viewing.
>
> Based on what I see in his examples, it appears that the R-L processing
> makes full pixel and upsampled images look better on screen than does
> simple USM. My rough guess is that in a print of his side by side
> example at his specified magnification, the USM version will look
> better, sharper, clearer, than the R-L version at any normal viewing
> distance.
>
> For the screen? I think I've equaled or bettered his AR-L result using
> far less exotic and processor intensive tools. I would not use any of
> them for printing.
>
> Moose
--
_________________________________________________________________
Options: http://lists.thomasclausen.net/mailman/listinfo/olympus
Archives: http://lists.thomasclausen.net/mailman/private/olympus/
Themed Olympus Photo Exhibition: http://www.tope.nl/
|