It is always wise to start with the simplest functional implementation, but if you want something significantly better that takes just about the same amount of time both to code and to compute, I'll again suggest nearest neighbor vertically, with a nice Gaussian horizontally. As a bonus, it works great with scanline-based file formats, like many of those supported by OIIO.
On Mon, Apr 2, 2012 at 2:21 PM, Chris Foster <[email protected]> wrote: > On Tue, Apr 3, 2012 at 3:08 AM, Larry Gritz <[email protected]> wrote: > > I would use nearest neighbor, or perhaps bilinear resampling first, and > then > > see if there is any point in doing anything better. Let's see what it > feels > > like the simplest way possible, then decide if more is needed. > > Good idea. I recently used the most braindead nearest-neighbour > sampling to make individual image icons for the aqsis framebuffer and > image viewer app. The results aren't exactly pretty (hello severe > aliasing!), but it's so much more useful than nothing, and extremely > easy to code. The size of the iv thumbnails will be a lot bigger, so > the aliasing won't be so bad... > > ~Chris > _______________________________________________ > Oiio-dev mailing list > [email protected] > http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org >
_______________________________________________ Oiio-dev mailing list [email protected] http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org
