In XSI 5.11, we have a plugin here that automatically sets the FGMap to
$TEMP/fgmap if it's set to Overwrite. This way, each machine generates
it's own fgmap locally with no risk of conflict between them. If it's
set to Append or Read Only, it sets the FGMap to a network path
(Render_Pictures) so that all machines can access the same FGMap.
Normally, you pre-generate it on one machine by setting it to Append,
and then you launch the render an many machines by setting it to Read
Only. Depending on the shot, that pre-generation process can be made by
rendering the whole sequence in step 10 with a very crappy anti-aliasing.
In XSI 6.0 and mental ray 3.5, everything has changed. I assume you can
launch a render on many machines when setting it to Multi Frame, but
don't take my word for it. I still need to read about it to make sure.
Wayne Williams wrote:
Hey guys,
I've a question about process and potential problems of trying to
create a fgmap via a renderfarm.
Firstly, is it even possible to build a "proper" fgmap via appending
from nodes on a farm.
Im guessing that potentially there is a problem where rendernode 5
can't append to the fgmap for frame 5 because rendernode 4 isn't
completed appending for frame 4. Does the fgmap appending work in this
fashion....where it is a chronological process that has to be followed
in a consecutive frame by frame basis? If so, how do you actually
build up a fgmap to use for final renders using the farm? Render the
fgmap locally?
I did get a fgmap to be appended to from the farm here....but if it is
a serial kinda process....what actually happened with the map? Is it
completely borked now?
How are you guys actually using FG for animation in a production
environment? or are you :)? Thanks for any tips.
-Wayne
---
Unsubscribe? Mail Majordomo(at)Softimage.COM with the following text in body:
unsubscribe xsi
