HtoT is intended to work with Tractor 2.3, Houdini 17 and up.
These steps should be done by your system administrator (or IT team)
- Make sure houdini-17+ and tractor-2.3 are both correctly installed on every machine
- On every machine that needs to send Tractor jobs, add Tractor's Python API path to every
machine's PYTHONPATH environment variable. The path should look like :
C:/Program Files/Pixar/Tractor-2.3/lib/python2.7/Lib/site-packages
- Copy htot.hdanc to your Houdini asset library. Ideally this should be a shared directory for every machine. For more details, check out : sidefx.com/docs/houdini/assets/install
- Since version 2, you do not need to copy the python files anywhere, they are just a copy of the contents found in the HDA's Scripts tab
- Create a
htot
node in a/out
context - Specify the Output Driver (e.g.
/out/mantra1
or/out/RIS1
). This is the only mandatory field before you can spool a job to Tractor. - Set your job parameters :
- Frame range : start and end frame (default :
$FSTART
,$FEND
) - Title : title of your tractor job (default :
[$HIPNAME][<renderer>] Render frames <start> - <end>
) - Start Paused : if enabled, the job will be spool and immediately paused (default : disabled)
- Projects : name of the projects attached to this job, separated by commas
- Priority : priority of the job (default :
1
) - Max Active : cap the number of simultaneously active tasks from that job (default : disabled)
- Debug Mode : debug mode will only print your job to the command output instead of spooling it (default :
False
)
- Frame range : start and end frame (default :
- Hit 'Spool To Tractor' !
- For more information, you can click the Help button on HtoT's parameters interface
- Spool Mantra and Renderman jobs
- Generate ifd/rib archives locally or as part of Tractor job
- Renderman jobs will only render rib files locally. This should be fixed in a later release
- Arnold is only partially implemented although it won't go further as I have no way to test
- No support for other renderers
- At the moment archives are generated then frames are rendered from these archives. This is to avoid using one Houdini license AND one render engine license per blade for the entire duration of the render. In the future it would be interesting to be able to choose (for cases where licenses are not a problem and/or archives will take too much space).
You're welcome to contribute to this project by creating a branch and issuing a merge request like a normal human being. I've made it so implementing a new render engine should be easy.