HeightField-Collision

A place to discuss everything related to Newton Dynamics.

Moderators: Sascha Willems, walaber

Re: HeightField-Collision

Postby Julio Jerez » Tue Jan 28, 2014 11:32 am

well yes, the cell are the spaces in between the points,
a heighfield if n x m point will have (n -1) x (m - 1) cells

if you for example have a map that is 256 x 256, and you subdivided into 4 sub maps of 128 x 128
the firs tile is gores form 0 to 128 points that's a 129 x 129 tile the second for 128 to 256 ( notice that point 256 does not exit what people do is that the either add a guard row and the end, or use mod so it wrap to point 0)
that is (256 - 128) + 1 = 129 x 129 tile
Julio Jerez
Moderator
Moderator
 
Posts: 12426
Joined: Sun Sep 14, 2003 2:18 pm
Location: Los Angeles

Re: HeightField-Collision

Postby Julio Jerez » Tue Jan 28, 2014 1:13 pm

a guard row should only be needed in the last row and column on the map, intermediated tile share the dividing line with their neighborgs.
Julio Jerez
Moderator
Moderator
 
Posts: 12426
Joined: Sun Sep 14, 2003 2:18 pm
Location: Los Angeles

Re: HeightField-Collision

Postby Julio Jerez » Tue Jan 28, 2014 6:26 pm

do you have a test that can be tryed?
Julio Jerez
Moderator
Moderator
 
Posts: 12426
Joined: Sun Sep 14, 2003 2:18 pm
Location: Los Angeles

Re: HeightField-Collision

Postby Julio Jerez » Wed Jan 29, 2014 10:35 am

you can serialize you scene an import it the sandbox demo, and see if you can recreate the error there.
Julio Jerez
Moderator
Moderator
 
Posts: 12426
Joined: Sun Sep 14, 2003 2:18 pm
Location: Los Angeles

Re: HeightField-Collision

Postby Julio Jerez » Wed Jan 29, 2014 2:25 pm

yes.

I will check that serilized file later.
Julio Jerez
Moderator
Moderator
 
Posts: 12426
Joined: Sun Sep 14, 2003 2:18 pm
Location: Los Angeles

Re: HeightField-Collision

Postby Julio Jerez » Thu Jan 30, 2014 11:19 am

Ok thank for the test, yes there was a bug there.
It is fixed now.
Please sync and try again
Julio Jerez
Moderator
Moderator
 
Posts: 12426
Joined: Sun Sep 14, 2003 2:18 pm
Location: Los Angeles

Re: HeightField-Collision

Postby Julio Jerez » Fri Jan 31, 2014 11:38 am

Ok I have it,
you are right there is a bug in the AABB calculation here
CalculateMinExtend3d (data->m_p0, data->m_p1, boxP0, boxP1);

basically the function does not takes into account the shape aabb can be straddling a terrain boundary, an dit does nop clip it.
It is eassy to fix, I am doing it now.
Julio Jerez
Moderator
Moderator
 
Posts: 12426
Joined: Sun Sep 14, 2003 2:18 pm
Location: Los Angeles

Re: HeightField-Collision

Postby Julio Jerez » Fri Jan 31, 2014 11:58 am

Al right fixed, you can sync and try again

Thank you the Bug report and the test
Julio Jerez
Moderator
Moderator
 
Posts: 12426
Joined: Sun Sep 14, 2003 2:18 pm
Location: Los Angeles


Return to General Discussion

Who is online

Users browsing this forum: No registered users and 7 guests

cron