Jump to content

  • Log In with Google      Sign In   
  • Create Account


#ActualAshaman73

Posted 05 November 2012 - 06:44 AM

The problem at mirrored seams is, that the tangentspace changes the handness at the mirrored faces (=> dot(cross(T,B),N)<0.0f is a way to determine this). When using shared vertices you only save one tangent space, so one face will use the wrong one often resulting in inverted lighting. The solution is quite simple, just duplicate the vertices when you detect a change in the handness of the tanget space much like you would duplicate a vertex when the uv coords change.

#4Ashaman73

Posted 05 November 2012 - 06:43 AM

The problem at mirrored seams is, that the tangentspace changes the handness at the mirrored faces. When using shared vertices you only save one tangent space, so one face will use the wrong one often resulting in inverted lighting. The solution is quite simple, just duplicate the vertices when you detect a change in the handness of the tanget space much like you would duplicate a vertex when the uv coords change.

#3Ashaman73

Posted 05 November 2012 - 06:42 AM

The problem at mirrored seams is, that the tangentspace changes the handness at the mirrored faces. When using shared vertices you only save one tangent space, so one face will use the wrong one often resulting in inverted lighting. The solution is quite simple, just duplicate the vertices when you detect a change in the tanget space much like you would duplicate a vertex when the uv coords change.

#2Ashaman73

Posted 05 November 2012 - 06:41 AM

The problem at mirrored seams is, that the tangentspace changes the handness at the mirrored faces. When using shared vertices you only save one tangent space, so one face will use the wrong one often resulting in inverted lighting. The solution is quite simple, just duplicate the vertices when you detect a change in the tanget space much like you would duplicate a vertex when the uv coords changed.

#1Ashaman73

Posted 05 November 2012 - 06:41 AM

The problem at mirrored seams is, that the tangentspace changes the handness of the mirrored faces. When using shared vertices you only save one tangent space, so one face will use the wrong one often resulting in inverted lighting. The solution is quite simple, just duplicate the vertices when you detect a change in the tanget space much like you would duplicate a vertex when the uv coords changed.

PARTNERS