> If adopted; I imagine this would be added as an additional> option/parameter > to the current ImageResize() function, like> 'isMultiPass'. A global code > change would taunt the dependency gods, no?> Some people may rely on the > classic behavior for quirks (those who want> processor speed and not quality, > for example). Well it is slower when done from CF, but if it was implemented internally the speed difference would be much smaller... My thought was if you wanted the normal behavior you would just use bicubic (since it is the default anyway and is just as fast as bilinear) - I kind of liked the idea that it wouldn't require any changes to the calling interface...
What if it were added as another quality option? Instead of just bicubic, bilinear, and nearest you added say "multistep binear" or whatever it should be called. Then use the same logic as I originally outlined (you don't want to do this if the new size is bigger than 50% of the original size - it does more harm than good and it gets a lot slower to boot). Then again... My understanding is the bad results that ImageResize() gets are not due to the interpolation method - it is a regression that should be fixed in Java itself. This whole "multistep" thing is just a workaround that shouldn't be necessary. On Nov 26, 10:56 pm, Alan Holden <[email protected]> wrote: > It seems pretty impressive alright. > > I would have liked to able to access the raw file paths, rather than > have the pix pass through this mediafire.com presentation gizmo. Some > people might worry that it's 'doing something' to taint the results of > your experiment. > > If you don't have a public server, email the images to me off-list and > I'll be happy to host the (jpg, png, etc) examples for you. > > It would be even better if we could host a running "example engine" for > people to try with their own images! > > If adopted; I imagine this would be added as an additional > option/parameter to the current ImageResize() function, like > 'isMultiPass'. A global code change would taunt the dependency gods, no? > Some people may rely on the classic behavior for quirks (those who want > processor speed and not quality, for example). > > My .02 for the topic > Al -- online documentation: http://openbd.org/manual/ google+ hints/tips: https://plus.google.com/115990347459711259462 http://groups.google.com/group/openbd?hl=en
