Frank van Maarseveen schreef: > On Sat, Nov 28, 2009 at 04:14:25PM +0100, Jos De Laender wrote: > >> Frank van Maarseveen schreef: >> >>> On Sat, Nov 28, 2009 at 01:41:04PM +0100, Jos De Laender wrote: >>> > [...] > > >>> The human perception comes into play because the algorithm should >>> do something which ultimately pleases the eye (or more precise: the >>> human brain). For example, the implementation contains a sqrt() which >>> I think is an approximation of a gamma compensation >>> >> My understanding is that the sqrt is introduced to denoise shadows and >> highlights +/- equally. >> > > Reading carefully, both explanations could be correct at the same time :-) > Yes , you're probably right. Including the reason why it is left out in the gimp (which is already gamma compensated). Thanks for the insight. > <snip> > > [...] > >>> I've had a look at the hat transform function: two out of three loops are >>> there for handling pixels close to the edge of the image. The function >>> looks ok to me from a logical/programmer point of view. >>> >> No offence, but how can you say without knowing the underlying algorithm ? >> > > hat_transform() considers pixels to the "left" and "right". How it > handles this close to the edge when either left or right side is not > available must at least be symmetrical. It is a corner case which might > not even be described in the original math. The middle loop is the > real algorithm. The 'st' (step) multiplier is not really part of the > algorithm but how to do code it once for >1 dimension. > Yes , of course ... You made me see the light ! It is simply the implementation of the h = (0.25,0.5,0.25) linear Lagrange interpolation (as I thought), with some hocus-pocus at the edges. So the complete algorithm is "A trous" algorithm with a linear interpolation filter. Once more thanks for your insight Frank. (and if you would know by any chance the math/algorithm behinde the noise array ...).
Jos ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ ufraw-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/ufraw-devel
