![]() ![]() > resolution, does the compositing, and then shrinks back to what the user > OK, why can’t we do that? Have a mode that internally runs 2X or 4X > image in this way will change the perceived line width. > Reason 3: Line width is measured in pixels. To render such large images with images/movies, it renders image You are likely to have a rendering context of the original > Reason 2: There can be technical difficulties in rendering images that It can be a noticeable difference when rendering interactively. Increasing the resolution 4X increases the compositing time by up > Reason 1: Increasing the resolution 2X increases the compositing overhead by > Thus, when it shrinks by 2, you get normal, poor man’s AA lines. ![]() > 3) Have the line widths automatically written 2X requested width. ![]() > 2) We do it (when rendering off screen) all the time when making movies > will be slower (when lines are visible). > 1) Make it 2X – i.e., 4 times the pixels, if the user asks for better > To: Kenneth Moreland >, David E DeMarle > important enough now to have a software push for it? The question is, do you really think it is > We turned off antialiasing in ParaView years ago because it caused way more > features (OpenGL2, OpenSWR, Ospray, IndeX). > happening in the middle of a push for other (more important) rendering > destabilize several features relying on the rendering. > development cycles, add complexity to the rendering system, and potentially > non-trivial change to the rendering system that will take significant We know how to achive this effect, but it requires a > We should keep it on the table for future work. > or real) probably isn’t worth a non-trivial software change at this point. I would say AA (either the hack mentioned below On Tue, at 10:08 PM, Scott, W Alan > wrote: This way, 2D charts, for example, will always have AA turned on. What do folks think about using AA byĭefault for builtin or client-side render and not when compositing? While the technical arguments are indeed sound, we tend to get askedĪbout this quite periodically. On Tue, at 4:27 PM, Utkarsh Ayachit > wrote: Subject: Re: Re: Remind me - AAĪs long as it doesn't affect selection, sounds pretty snazzy. I think transparency (depth peeling) was one.įrom: David E DeMarle Ĭc: Scott, W Alan > Moreland, Kenneth > paraview-developers at I think there were others, but I don’t remember them all. I definitely remember picking as an issue that cropped up. Using antialiasing only on client side used to be our solution, but it turns out antialiasing screws up other stuff, too, and we eventually threw up our hands and got rid of it.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |