A place to discuss everything related to Newton Dynamics.
Moderators: Sascha Willems, walaber
by Julio Jerez » Mon Nov 28, 2011 3:48 pm
ha ok I will debug it tonight with that sample mesh.
thanks for the sample.
-
Julio Jerez
- Moderator

-
- Posts: 12426
- Joined: Sun Sep 14, 2003 2:18 pm
- Location: Los Angeles
-
by Bird » Mon Nov 28, 2011 4:02 pm
Julio Jerez wrote:all right I re enable my implemnetaion again.
In the cow mesh that has abput 1500 vetex, Mamou take abput 40 secund to make it, Netwon's take less that halh a secund, about 70 time faster.
tonight I will work to fix the convexity metric bug and add the fix for the concaves faces left overs. that should complete this task for good
Wow! That's a lot better than I was expecting I tried the latest HACD on a cow with 2770 vertices and it parsed in about 15 seconds using the HACD default settings. Here's a little video of how it works in my project now.
http://www.hurleyworks.com/media/flash/CompositeCow/CompositeCow.htmlLooking for to testing the new version out
Thanks!
-Bird
-
Bird
-
- Posts: 636
- Joined: Tue Nov 22, 2011 1:27 am
by Julio Jerez » Mon Nov 28, 2011 4:21 pm
I download the lates version on saturday, so I thionk we have the same,
teh cow took a very long time, so I deciomated in 3dxMax to 1400 vertices.
my 40 secund convex constrution is in debug mode. in release is about 10 to 15 like you say.
the newton one is less than a secund, but it is not for prime time yet.
tonight I will go over again over the concavity heuristic, and tomorroy I will add the extension to handle the concave island.
then maybe you can try, and see if it work.
then I will send the sugestion to him again to see if he wants to added to the original code base.
-
Julio Jerez
- Moderator

-
- Posts: 12426
- Joined: Sun Sep 14, 2003 2:18 pm
- Location: Los Angeles
-
by Julio Jerez » Tue Nov 29, 2011 11:07 am
Ok I see the problem, my heusritic to preserve the conncet pirce is no good, because I simple add the confiction face as new edge at teh end of the list.
that thgsi in fact make only one segmen and all right, but each shared vertes generate one new segment.
I have a better histic now, and I am fixxing it. I will take me more than one night to get it, but it will be very, very good.
thank you for showing me that case, very very good because improves reliability of the MeshEeffct tool.
stan by I will upload a newer version as sonn as I get it going.
-
Julio Jerez
- Moderator

-
- Posts: 12426
- Joined: Sun Sep 14, 2003 2:18 pm
- Location: Los Angeles
-
by Bird » Tue Nov 29, 2011 11:40 am
Julio Jerez wrote:Ok I see the problem, my heusritic to preserve the conncet pirce is no good, because I simple add the confiction face as new edge at teh end of the list.
that thgsi in fact make only one segmen and all right, but each shared vertes generate one new segment.
I have a better histic now, and I am fixxing it. I will take me more than one night to get it, but it will be very, very good.
thank you for showing me that case, very very good because improves reliability of the MeshEeffct tool.
stan by I will upload a newer version as sonn as I get it going.
I'm glad to help! Looking for to testing the new solution.
-Bird
-
Bird
-
- Posts: 636
- Joined: Tue Nov 22, 2011 1:27 am
Return to General Discussion
Who is online
Users browsing this forum: No registered users and 2 guests