POV-Ray : Newsgroups : povray.programming : Anyone up to testing? : Re: Anyone up to testing? Server Time
29 Jul 2024 00:23:39 EDT (-0400)
  Re: Anyone up to testing?  
From: Mike
Date: 25 Jan 1999 03:48:22
Message: <36AC2F16.33EAC67C@aol.com>
You're right.  I had done a few more tests and what I had changed hadn't
fixed the problem. Changing Radiosity_Trace_Depth to 2 causes the
problem all the time.  If it is being set to two by an interrupted
render, it must only happen when the render is interrupted at a specific
portion of the radiosity code, which would explain why the problem is
intermittent and unpredictable.

In radiosit.c in Deinitialize_Radiosity_Code(), add this at the end:

if (Radiosity_Trace_Level !=1)
  {
	  Radiosity_Trace_Level = 1;
  }

This should ensure that if the render is interrupted that this variable
gets reset to 1.

Let me know if it works.

-Mike

Ben Birdsey wrote:
> 
>         It looks like you've gone a long way toward curing the problem with
radiosity
> for *continued* renders.  However, my problem is that if the scene looks like
> junk I want to be able stop right in the middle and tweak it and then re-render
> it. But interrupting the trace almost always crashes Radiosity.
> 
>         As far as I can tell, the real problem with interrupted renders is that the
> global variable radiosity_trace_depth is not re-initialized every time an image
> is rendered.  From my investigations, it looks like the global
> radiosity_trace_depth is only initialized when POV loads.  So, if you interrupt
> a render when radiosity_trace_depth = 2 and you render a *new* image, the
> renderer will *begin* with radiosity_trace_depth = 2.
> 
>         A simple fix might be to move the radiosity_trace_depth initialization to
the
> code which initializes radiosity ( in Radiosity.c ) and to set
> radiosity_trace_depth = 0 if not continue_trace.
> 
>         In Him,
>         Ben


Post a reply to this message

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.