On Fri, Jul 6, 2012 at 6:25 AM, Stefan Stavrev <[email protected]> wrote: > In summary, the two cases are: > > 1. User specifies width and number of bins, but we modify number of > bins to fit width. > > 2. User specifies number of bins and X, and we decide about the width. > > So, if you want a fixed histogram window like in Photoshop use > option 1. If you want variable width window like in Matlab I think, > use option 2. > > I think there is no escape from snapping the number of bins in case 1, > it is just the nature of the drawing logic.
What you're talking about here is just like resampling an image from one size (the natural width based on number of histogram buckets) to another. Doing snapping is just like doing the resampling using nearest neighbour interpolation (which is normally thought of as pretty crude since it leads to aliasing/pixellation). Never mind though, pixellation is exactly what we want here for clarity and simplicity. I was going to suggest using option 2 but Larry's suggestion is much better: Simply require the output to have number of columns equal to the width of the histogram and the user can always resize it later. ~Chris _______________________________________________ Oiio-dev mailing list [email protected] http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org
