Hi,
So I can get a list of sublayers in a usd with GetCompositionAssetDependencies() [graphics.pixar.com] but is there any way to preserve the ordering? This returns an alphabetical list of sublayers.
Cheers,
J
Found 69 posts.
Search results Show results as topic list.
Solaris and Karma » Querying sublayers with GetCompositionAssetDependencies()
-
- protean
- 75 posts
- Offline
Solaris and Karma » Mute Layer and Payloads
-
- protean
- 75 posts
- Offline
If I drop a configure layer LOP (with no changes) before referencing the cube then it works as expected
Solaris and Karma » Mute Layer and Payloads
-
- protean
- 75 posts
- Offline
Hi,
I have some odd issue with layer muting and payloads.
In this very basic example I have a shot USD with an animation layer which consists of a payload to to the cube and a reference to the animation of it.
When I load the shot.uds and mute the animation layer everything looks normal (the cube no longer 'exists'
). However, when I load the reference to the static cube again as a payload the animation is still occurring. If I load the cube as a regular reference the cube is static, as I would expect.
The idea here being I want to mute the layer and rebuild it rather than retain what was there before.
I have some odd issue with layer muting and payloads.
In this very basic example I have a shot USD with an animation layer which consists of a payload to to the cube and a reference to the animation of it.
When I load the shot.uds and mute the animation layer everything looks normal (the cube no longer 'exists'

The idea here being I want to mute the layer and rebuild it rather than retain what was there before.
Edited by protean - Feb. 27, 2023 05:14:02
Solaris and Karma » Alembic and primvars
-
- protean
- 75 posts
- Offline
Ah... It does seem to. Thanks!
I actually notice velocity seems to come through but I've not inspected it's accuracy yet. So perhaps it's just the format of the vector being myvec.0, myvec.1, myvec.3 instead of r,g,b or x,y,z?
J
I actually notice velocity seems to come through but I've not inspected it's accuracy yet. So perhaps it's just the format of the vector being myvec.0, myvec.1, myvec.3 instead of r,g,b or x,y,z?
J
Solaris and Karma » Alembic and primvars
-
- protean
- 75 posts
- Offline
Hi,
Referencing an ABC(exported from Houdini) directly doesn't seem to to allow me to use the custom attribs in a 'geometry property value' shader for material x. Using a SOP create to load the ABC does work and I notice the formatting of the primvars is a little different between the the two import methods. I assume something to do with this:
https://www.sidefx.com/forum/topic/87928/ [www.sidefx.com]
So for example; if I have an FX fluid sim as an ABC with some custom attribs, the only way to access those in a shader would be to write the geometry out again as usdc via something like sopcreate. Is that so? Perhaps I need to find a way to write to usdc in the first place to avoid duplicating these sims?
J
Referencing an ABC(exported from Houdini) directly doesn't seem to to allow me to use the custom attribs in a 'geometry property value' shader for material x. Using a SOP create to load the ABC does work and I notice the formatting of the primvars is a little different between the the two import methods. I assume something to do with this:
https://www.sidefx.com/forum/topic/87928/ [www.sidefx.com]
So for example; if I have an FX fluid sim as an ABC with some custom attribs, the only way to access those in a shader would be to write the geometry out again as usdc via something like sopcreate. Is that so? Perhaps I need to find a way to write to usdc in the first place to avoid duplicating these sims?
J
Edited by protean - Jan. 24, 2023 13:07:15
Technical Discussion » Switch-like input connectors
-
- protean
- 75 posts
- Offline
Hi,
I was wanting to make my own custom switch in ROPs in this case, but I believe the same applies elsewhere. Is it possible to show which input is 'active' (solid vs. dotted line)? In this example there happens to be a switch in the HDA but it seems the live input display is not propagated through subnets.
I did read the bit about reference inputs but am currently unsure how to set that other than manually in the operator type properties panel.
Cheers,
J
I was wanting to make my own custom switch in ROPs in this case, but I believe the same applies elsewhere. Is it possible to show which input is 'active' (solid vs. dotted line)? In this example there happens to be a switch in the HDA but it seems the live input display is not propagated through subnets.
I did read the bit about reference inputs but am currently unsure how to set that other than manually in the operator type properties panel.
Cheers,
J
Edited by protean - Sept. 12, 2021 12:21:45
Technical Discussion » MuteLayer Python LOP
-
- protean
- 75 posts
- Offline
Hi,
I was trying to recreate the functionality of Configure Stage LOP using a Python LOP, primarily the muting function. My very basic test below returns 'True' which implies the layer is muted but it does not mute in the viewer/layer panel and Houdini also bombs when I change frame. I'm assuming I'm missing some kinda Solaris caching thing.
And I'm not using asset resolving for now. I was hoping to get wildcard functionality in the Configure Stage LOP 'mute paths' parm
Thanks,
J
I was trying to recreate the functionality of Configure Stage LOP using a Python LOP, primarily the muting function. My very basic test below returns 'True' which implies the layer is muted but it does not mute in the viewer/layer panel and Houdini also bombs when I change frame. I'm assuming I'm missing some kinda Solaris caching thing.
node = hou.pwd() stage = node.editableStage() stack = stage.GetLayerStack(includeSessionLayers = False) pth = 'p:/mypath/mylayer.usda' for lyr in stack: if 'mylayer' in lyr.identifier: print lyr.identifier stage.MuteLayer(pth) print stage.IsLayerMuted(pth)
And I'm not using asset resolving for now. I was hoping to get wildcard functionality in the Configure Stage LOP 'mute paths' parm
Thanks,
J
Technical Discussion » NPR/PBR shading
-
- protean
- 75 posts
- Offline
Hi,
I've been doing a lot of NPR shading using illuminance loops and it's all fine and well until people want it to interact with PBR. A little research indicates it's not a trivial process to do anything beyond exporting cell shading as Ce for PBR. I think even the toon shading VOP does exactly that.
Other issues occur as well, including very bad sampling of area and geo lights. I mean I have something that works but with many caveats to consider. Illuminance lops and gather are my friends but I'd like to try something new
In the interest of making things more robust and having recently discovered https://www.sidefx.com/docs/houdini/vex/functions/cvex_bsdf.html [www.sidefx.com] I wondered if it was possible to build my own BSDF for cell shading? It seems feasible right?! (although I wonder why I can't find a cell shading BSDF example anywhere)
I don't imagine it is trivial although the documentation has clear examples of traditional models.
The other issue that's in the back of my mind is if Karma will be more limited and any custom BSDF work I do now might not make the migration.
I hope I can post a couple of examples later but it might not be allowed :/
I've been doing a lot of NPR shading using illuminance loops and it's all fine and well until people want it to interact with PBR. A little research indicates it's not a trivial process to do anything beyond exporting cell shading as Ce for PBR. I think even the toon shading VOP does exactly that.
Other issues occur as well, including very bad sampling of area and geo lights. I mean I have something that works but with many caveats to consider. Illuminance lops and gather are my friends but I'd like to try something new
In the interest of making things more robust and having recently discovered https://www.sidefx.com/docs/houdini/vex/functions/cvex_bsdf.html [www.sidefx.com] I wondered if it was possible to build my own BSDF for cell shading? It seems feasible right?! (although I wonder why I can't find a cell shading BSDF example anywhere)

I don't imagine it is trivial although the documentation has clear examples of traditional models.
The other issue that's in the back of my mind is if Karma will be more limited and any custom BSDF work I do now might not make the migration.
I hope I can post a couple of examples later but it might not be allowed :/
Solaris and Karma » Python command to refresh Context Options Editor?
-
- protean
- 75 posts
- Offline
Hi!
Was there any update to this functionality? I'm running into the same issue where context options can be made, deleted and imported but still needs user input to refresh the UI. I'm gonna go and break stuff by editing XMLs now.
Cheers,
j
Was there any update to this functionality? I'm running into the same issue where context options can be made, deleted and imported but still needs user input to refresh the UI. I'm gonna go and break stuff by editing XMLs now.
Cheers,
j
Solaris and Karma » Sublayer Vs Reference
-
- protean
- 75 posts
- Offline
Ah cool thanks!
I think I get but I'll see if I can break it when I have a more complex setup :p I'm thinking in general that references make more sense when building/assembling stuff (such as individual asset or the assembly for an environment) and that sublayers work better for storing changes to the original assemblies. Although I know they could be used interchangeably they do also seem to manifest themselves differently in the Scene Graph Layers Panel.
J
I think I get but I'll see if I can break it when I have a more complex setup :p I'm thinking in general that references make more sense when building/assembling stuff (such as individual asset or the assembly for an environment) and that sublayers work better for storing changes to the original assemblies. Although I know they could be used interchangeably they do also seem to manifest themselves differently in the Scene Graph Layers Panel.
J
Solaris and Karma » Sublayer Vs Reference
-
- protean
- 75 posts
- Offline
Hi,
I hope the title doesn't provoke too much eye-rolling! I was reading Matt Estela's notes about weaning himself off sublayers but I'm not able to use reference to perform the same function.
Attached is a network that describes a hypothetical lookdev scenario where the decisions (material assign and geometry settings) for LD are stored as a seperate layer. Importing them back as a sublayer works (in that it renders as expected).
I'm clearly fumbling with some basic concepts but to me it seems that reference is working mainly for importing ‘things’ into the scene graph in the current layer and sublayer is working for importing decisions (e.g. prim attribute modifications).
The second network is a hypothetical sublayer workflow describing two lookdev stages (basic + advanced) where the later, advanced decisions are in the later, stronger layer. Again, not something I'm getting with the reference node workflow.
Cheers,
J
I hope the title doesn't provoke too much eye-rolling! I was reading Matt Estela's notes about weaning himself off sublayers but I'm not able to use reference to perform the same function.
Attached is a network that describes a hypothetical lookdev scenario where the decisions (material assign and geometry settings) for LD are stored as a seperate layer. Importing them back as a sublayer works (in that it renders as expected).
I'm clearly fumbling with some basic concepts but to me it seems that reference is working mainly for importing ‘things’ into the scene graph in the current layer and sublayer is working for importing decisions (e.g. prim attribute modifications).
The second network is a hypothetical sublayer workflow describing two lookdev stages (basic + advanced) where the later, advanced decisions are in the later, stronger layer. Again, not something I'm getting with the reference node workflow.
Cheers,
J
Technical Discussion » openGL UDIM resolution
-
- protean
- 75 posts
- Offline
Hi,
I'm noticing that UDIM textures in openGL don't display full resolution. Also not rendered in full res in the openGL renderer. Is this there a workaround or is it an necessary limitation to of openGL?
This is for 17.5 sadly.
Cheers
J
I'm noticing that UDIM textures in openGL don't display full resolution. Also not rendered in full res in the openGL renderer. Is this there a workaround or is it an necessary limitation to of openGL?
This is for 17.5 sadly.
Cheers
J
Technical Discussion » LPE per light
-
- protean
- 75 posts
- Offline
Technical Discussion » LPE per light
-
- protean
- 75 posts
- Offline
Hi,
I was hoping to get combined direct and indirect diffuse as an AOV per light but not sure if LPEs work with light exports in Mantra. At least I haven't managed it in 17.5
Cheers,
J
I was hoping to get combined direct and indirect diffuse as an AOV per light but not sure if LPEs work with light exports in Mantra. At least I haven't managed it in 17.5
Cheers,
J
Technical Discussion » Opacity independent shading
-
- protean
- 75 posts
- Offline
Hi,
I'm struggling a bit with the challenge of getting some shading, in this case specular, over a 100% transparent surface in the context of a non-photoreal shader. I know Mantra doesn't conform to a ‘priority fill’ type renderer but I would still like to try to get it to work.
I would like to avoid rendering twice for the shading I need over the transparent objects.
Cheers
J
I'm struggling a bit with the challenge of getting some shading, in this case specular, over a 100% transparent surface in the context of a non-photoreal shader. I know Mantra doesn't conform to a ‘priority fill’ type renderer but I would still like to try to get it to work.
I would like to avoid rendering twice for the shading I need over the transparent objects.
Cheers
J
Technical Discussion » Hair and Clump length
-
- protean
- 75 posts
- Offline
Hi,
I have this issue when I want to use clumps of different lengths (lets assume that I'm using guides as clumps here for simplicity). When there are guides that differ relatively significantly in length then all the hairs lengths around them are naturally interpolated. However this does produce the undesired effect of some hairs being associated with clumps that are shorter leading to hair tips that bend backwards.
I know I can use the clump profile to force a certain minimum amount of clumping along the whole length of the hair but that has an undesired effect. I thought maybe I could force the lengths of hair to be identical to the clump or guide they are associated with (thereby removing the interpolation of length) but I can't see if that is possible?
I could also generate a huge amount of guides in sops of course but then I may as well just render those guides directly and bypass fur.
Cheers
John
I have this issue when I want to use clumps of different lengths (lets assume that I'm using guides as clumps here for simplicity). When there are guides that differ relatively significantly in length then all the hairs lengths around them are naturally interpolated. However this does produce the undesired effect of some hairs being associated with clumps that are shorter leading to hair tips that bend backwards.
I know I can use the clump profile to force a certain minimum amount of clumping along the whole length of the hair but that has an undesired effect. I thought maybe I could force the lengths of hair to be identical to the clump or guide they are associated with (thereby removing the interpolation of length) but I can't see if that is possible?
I could also generate a huge amount of guides in sops of course but then I may as well just render those guides directly and bypass fur.
Cheers
John
Technical Discussion » REC709 Look-up Tables
-
- protean
- 75 posts
- Offline
protean
Hi,
Sorry to dig up this old thread but does anyone know how to produce the LUT table (via script etc) with the log sampling used? I'd quite like to produce some LUTs with log sampling that I cannot because of the ‘use only blue nodes’ limitation.
J
Nevermind. As usual the act of writing down my question promoted the solution. I found the log sampling method in the HDK docs:
http://www.sidefx.com/docs/hdk12.1/hdk_io_lut.html [sidefx.com]

Technical Discussion » REC709 Look-up Tables
-
- protean
- 75 posts
- Offline
Hi,
Sorry to dig up this old thread but does anyone know how to produce the LUT table (via script etc) with the log sampling used? I'd quite like to produce some LUTs with log sampling that I cannot because of the ‘use only blue nodes’ limitation.
J
Sorry to dig up this old thread but does anyone know how to produce the LUT table (via script etc) with the log sampling used? I'd quite like to produce some LUTs with log sampling that I cannot because of the ‘use only blue nodes’ limitation.
J
Technical Discussion » coverage pass
-
- protean
- 75 posts
- Offline
Hello again,
Does mantra have a pixel filter or default AOV to generate a coverage pass or does that have to be done in a shader?
Cheers
John
Does mantra have a pixel filter or default AOV to generate a coverage pass or does that have to be done in a shader?
Cheers
John
Technical Discussion » Writing camera matrix information into an exr's metadata
-
- protean
- 75 posts
- Offline
-
- Quick Links