On Fri, Jun 4, 2010 at 04:45, Juan Vuletich
<juan@...2357...> wrote:
Bryan Hoyt | Brush Technology wrote:
On Thu, Jun 3, 2010 at 12:06, Juan Vuletich <juan@...2357... <mailto:juan@...2357...>> wrote:
rendering, but the Inkscape versions shown side-by-side
don't, it doesn't seem like it's a fair comparison.
But then, fair comparison would be to turn on subpixel rendering
on Inkscape. Can that be done?
I don't believe it can be done.
Could you tweak your program to turn off subpixel rendering,
and give us the same samples with subpixel turned off in both?
How can I claim mi technique is better if I disable its
advantages? If I did as you say, someone could say, "hey Inkscape
does just only 1 pixel wide anti aliasing. could you tweak your
filter to do that, to do a fair comparison?" Then, my results
would be exactly those of Inkscape, and I'd kill all the progress
I did!
Sorry, it sounds like I misunderstood the intent of your algorithm -- I understood it as having advantages for non-subpixel rendering, and the subpixel stuff was just an extra bonus. But it sounds like the subpixel rendering is the main advantage of your algorithm. Is that right?
But I did something else. I opened the web page on several displays. Tried it in 3 LCD screens, with both DVI and VGA interfaces, and at native resolution and at lower resolution. In all cases the subpixel rendered version shows less pixellation and less color artifacts. Then I tried on 2 15" CRT screens, one Trinitron and the other regular, at 1024x768 and at 800x600. The whole pixel version looks consistently better than the Inkscape rendering, although both show some colored shadows that are due to the misalignment between logical pixels, the display's color mask and the actual colored phosphor on the screen. The subpixel version looks less pixellated and also with a slight colored shadow, but in different places. Overall, I find the subpixel version to also look better in these 2 CRT displays! Finally, I tried on an OLPC XO, both in color and b/w modes. There, all alternatives look equally good and it is hard to declare a winner.
So I say that given that the majority of displays sold and in use are RGB LCDs, given that it makes sense to obtain the highest possible quality from the best and most used displays, and given that the subpixel version looks better even in CRTs and doesn't hurt in the XO; it is advisable to use the subpixel version on any display, at least as a default.
When hardware can tell us about the physical geometry of the pixels, then software might make a better choice automatically.
You did mention "The geometry of the pixels (i.e. the sample positions) is a parameter to the rasterized. It is pretty easy to adapt it to any target or turn it
off completely". In light of how I now understand it, turning off subpixel geometry would mean turning off your algorithm altogether.
Well, no. My algorithms allows for easily using different filters and different pixel geometries. So, even if using a 1 pixel step filter and no subpixel position, and giving similar results to conventional renderers, it would still be using different algorithms. BTW, in my work, all coordinates are float, and it supports subpixel positioning with float precision, regardless of the sampling grid and the filter. I'm not sure if Inkscape does this. AGG claims to do it.
However, it would be nice to see subpixel rendering in Inkscape, so I applaud your project.
However, I believe many people use Inkscape for creating, not primarily for viewing, and fairly often for exporting (at least, that's how I use it). When you're creating an image file for other people to view, you generally want it be compatible with all types of screens. I use Inkscape mostly for web work, which means I can't use subpixel rendering at all. So I guess that means I can't really benefit from your rendering technique?
Well, as I said before, I believe subpixel rendering is better on almost all displays in use, and should be generally adopted. Even if you decide to turn it off, my filter gives better results (as shown by tigerWholePixel.png), and you also have the seam removal from polygon edges (as shown by the last Clinton image).
Good work, and don't take my questions as criticisms!
- Bryan
Thanks. I welcome criticism too!
Cheers,
Juan Vuletich