That is very impressive indead! That performance, without HW acceleration, and the helpfulness of the author makes it an easy choice.
Any thoughts on an interface API? Regards, McKinley --- Jarrad Hope <[EMAIL PROTECTED]> wrote: > Check out this example here: > http://antigrain.com/stuff/particle_demo2.exe > > Check bitmap caching, move particles to 1000, check > start universal. > > I get 80-100FPS :| and thats maximised to full > screen. > > That is incredible! Im totally convinced this is the > vector library for us > > On Sat, 17 Jun 2006 05:17:46 +0800, Maxim Shemanarev > > <[EMAIL PROTECTED]> wrote: > > > Hi Jarrad, > > > > > Once AGG is ported to nekoVM I am going to be > starting an animation > > > software. > > > Which will possibly need to handle complex > vector graphics > > > (importing/manipulating/displaying swf's) > > > > In this case, yes, the most recent addition can > help you handle the > > fancy Flash data model (rasterizer_compound_aa, > see > > examples/flash_rasterizer.cpp). As far as I know > there are no other open > > libraries that support it directly. In Amanith the > tessellator works > > only with traditional single-style polygons. In > Flash it's very > > different, that is, instead of cyclic contours > there's a set of acyclic > > edges with left and right fill attributes that > eventually form a closed > > planar graph. So that, you can rasterize or > tessellate the whole > > multi-color scene in one pass. But it isn't > trivial to reconstruct the > > polygons in order to use the traditional polygon > fill machinery. > > > > > I am abit concerned when you said > > > "you can't use it for heavy duty realtime > graphics" > > > well the graphics in the app would be a mix of > realtime and moving > > > things around. > > > > I mean really fancy real time graphics, like in > most modern games (not > > obligartory 3D). At least AGG isn't slower than > the original Flash > > player (maybe even faster). Still, very complex > scenes, with high > > resolution, like 1600x1200 may work too slow in > Flash. The criterion > > here can be as follows. If you are satisfied with > the performance of the > > original Flash, AGG is most probably suitable for > your needs too. If you > > need something dramatically faster, it's a long > way to go to HW > > accelerated solutions, which are typically very > complex, tedious to > > implement and expensive. And most of all, the > resulting quality isn't > > guaranteed. > > > > But if you have a well defined interface and > practically working > > software backend renderer you can replace it in > future with some other > > solution. > > > > > Because I was looking at the lens example on > the lion and the alpha > > > example > > > - both ran in realtime on my machine. The > MultiClip at fullscreen > > goes > slow, but holding space down on the flash > rasterizers is fine. > > > > That's right, the multi-clip isn't the best > solution; typically these > > things are implemented with alpha masking. > > > > > Perhaps theres a way of caching to make it run > faster??? > > > > Some things can definitely be bitmap-cached. For > example, the latest > > Flash 8 uses caching technique too. It's not that > trivial to implement, > > but doable: > http://antigrain.com/stuff/particle_demo2.exe > > > > BTW, you are welcome to the mail list: > > http://antigrain.com/maillist/index.html > > Perhaps, your ideas may be interesting to other > people too and you can > > get more help there. > > > > McSeem > > -- > Neko : One VM to run them all > (http://nekovm.org) > __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com -- Neko : One VM to run them all (http://nekovm.org)
