Forums Houdini Lounge Rendering issues with mantra through a objnet node

Rendering issues with mantra through a objnet node

   1062   5   1
User Avatar
jgoulet
Member
13 posts
Joined: Feb. 2016
Offline
 
Hi. I am working a tool for artists and to make it cleaner I am hopping to have a objnet responsible for the render aspect of the tool inside the hda itself. We were using the opengl render node, but I would need to move away from that so that I can have a solution working on blades without gpus. I am wondering how I could get the same render working with mantra if it doesn't have the "scenepath" parameter. The scene path parameter is the only reason we managed what we were doing with the opengl node as I understand it.

If any knows of a way to render with mantra flipbook like sequence that would also interest me. My main issue stays the fact that it is hard to render through sub levels of objnet. I know the same issue can rise when rendering with arnold, but usually doing the rop net and geometry network inside the same objnet made it work just like the opengl render node.
The scene is attached. Any help is welcomed.

Attachments:
renderExample.hipnc (362.1 KB)

 Report
User Avatar
tamte
Member
8518 posts
Joined: July 2007
Online
 
from my tests it seems like Mantra uses the scene subnet based on the camera path
so if you put camera in the same Objnet and point Mantra to it then your pig should render
Tomas Slancik
FX Supervisor
Method Studios, NY
 Report
User Avatar
jgoulet
Member
13 posts
Joined: Feb. 2016
Offline
 
Nice. I can work with that. Do you know if there is an easy way to make an object reference like object merge in sop but for obj for the camera? I will work on putting expressions on a dummy camera to match all attributes of the target camera but ideally I think having a reference could be better.

Thank you.
 Report
User Avatar
tamte
Member
8518 posts
Joined: July 2007
Online
 
jgoulet
to match all attributes of the target camera but ideally I think having a reference could be better.
I'd use chops for that, you can get cameras world transform using Object Chop and all parameter values you need using Fetch Chop, then just merge them and export to your embedded camera
Edited by tamte - Oct. 20, 2021 11:39:10
Tomas Slancik
FX Supervisor
Method Studios, NY
 Report
User Avatar
jgoulet
Member
13 posts
Joined: Feb. 2016
Offline
 
Can you throw away a quick exmample scene of that. I know little of CHOPS yet.
 Report
User Avatar
jgoulet
Member
13 posts
Joined: Feb. 2016
Offline
 
I used this little python script to help attach the parms of the dummy camera to the real one, if someone needs it later.

camera = hou.pwd().node("../dummy_camera")
# cam parm to target is the  sop path parm that users use to target a camera
targetCamParm = hou.pwd().node("../../camParmToTarget")
hScript_template = 'ch(chsop("../../camParmToTarget") + "/{}")'


skip_types = [hou.parmTemplateType.Folder, hou.parmTemplateType.FolderSet,
              hou.parmTemplateType.Separator, hou.parmTemplateType.Label]
# with this I am probably loosing the support for a lookat setup, but they cause problems when I hook them up with 
# an expression even if they aren't in use.
skip_parms = ['constraint_path', 'lookatpath', 'lookupobjpath', 'pathobjpath', 'vm_lensshader']
for parm in camera.parms():
    if parm.parmTemplate().type() in skip_types or parm.name() in skip_parms:
        continue
    parm.setExpression(hScript_template.format(parm.name()),
                       hou.exprLanguage.Hscript,
                       replace_expression=True)
Edited by jgoulet - Oct. 21, 2021 16:45:36
 Report
  • Quick Links