Chris,
Great article and clever substitute implementation. Do you have any sense of how much of your "multi-step" reduction is strictly necessary to smooth things out to the point where they look good? I would like to perform subsampling when I read in huge-gantic images from ImageIO, and am wondering what degree of blunt pre-scaling I could get away with. Was there a particular tile size for the old averaging filter?
Also, is there no chance that someone could just sneak a one-line warning into the legacy javadocs for Image.getScaledInstance? This article won't help somebody without Internet access, on a multi-generational starship to Alpha Centauri, etc.
Dave |
java.net RSS Feeds
