Re: [OM] Enhance Details

[Mike Gordon via olympus]

<<I admit I am not familiar with deconvolution sharpening. Someone pointed to 
the Point Spread Function (similiar to impulse response). How this PSF is 
determined for the sharpening I don't know and wonder if some optical 
assumptions are involved or if it is derived from the <<actual image? 

Moose is spot on about Canyon.  They have full wave solutions for their 
profiled lenses  and use that in the DLO (dig lens optimizer) in DPP.  Since  
Ver 3.11 they have had diffraction correction which  deconvolutes with the 
exact diffraction PSF for the lens and aperture.  Now 
one may not predict a tremendous benefit for the diffraction component as there 
is quite sharp cut-off in the frequency info lost.  

See previous post:

http://lists.tako.de/html/Olympus-OM/2017-01/msg00765.html

http://lists.tako.de/Olympus-OM/2013-05/msg02068.html

Unfortunately the link to R-L deconvolution  to improve a diffracted softened 
image within the second link is gone--not even in web archive--please see 
Roger's analysis at lensrentals though.  

So Canyon a priori has the exact PSF for the lenses and deconvolutes using 
them.   


"Blind" deconvolution techniques like R-L used by FM typically assume Gaussian 
blur as the PSF--it works rather well especially when applied in a layer and 
masking out solid colors like skies where can easily get ringing artifacts.  
Back in the day one  could call DXO and speak with an engineer.  They use some 
form of proprietary iterative deconvolution  to improve lens softening but have 
never said if it is R-L.  The method of application differs however as it is 
applied in a "contextural approach" considering both local noise and detail. 
Empirically this type of "blind" deconvolution also does improve diffraction 
softened images though one might not predict it to do so in an effective 
manner. Their "profiles" take all forms of aberrations into account and use 
different methods of amelioration. 
Some data is available from a 2004 patent but they keep their cards close to 
their chest.  Ringing artifacts in sky is almost unheard of with that software. 
 The deconvolution is applied where it is needed most and not to the whole 
image equally.  Sometimes it can be used to even out lens performance and then 
finish in PS using FM to further "sharpen" select areas. Moose judiciously  
applied FM in layers is undoubtedly as good as any of the above.

Mike
-- 
_________________________________________________________________
Options: http://lists.thomasclausen.net/mailman/listinfo/olympus
Archives: http://lists.thomasclausen.net/mailman/private/olympus/
Themed Olympus Photo Exhibition: http://www.tope.nl/