Unreal Engine Paragon asset settings .json

NVlabs / nvdiffmodeling

Differentiable rasterization applied to 3D model simplification tasks

Other

455 stars 30 forks source link

Unreal Engine Paragon asset settings .json #2

Closed TLJParadox closed 3 years ago

TLJParadox commented 3 years ago

could you tell me something about Automatic cleanup. in figure 7 this paper. What's your param settings(json) in Unreal Engine Paragon asset.

jmunkberg commented 3 years ago

Hello, In that experiment, we first reduced the model from 81k tris to 15k tris using the mesh reduction tool in Maya 2019, and used the reduced model as initial guess. The relevant parameters in the .json files are:

    "texture_res" : [1024, 8192],
    "train_res" : 2048,
    "batch": 8,
    "learning_rate" : 0.003,
    "iter" : 5000,

TLJParadox commented 3 years ago

Thank you. And are parameters "random_textures": true,"min_roughness" : 0.25 in this .json file ?
And how to calculate the diffuse map , such as in dancer_displacement "mtl_override": "data/dancer/dancer_diffuse.mtl" ?
dancer_1k.obj and dancer_8k.obj are also generated by Maya 2019? I use the Unreal4 to LOD in Paragon asset, but the experiment result isn't better than base_mesh what we got.

jmunkberg commented 3 years ago

"random_textures" : false, and no mtl_override so the optimization will use the textures from the reduced initial guess as a starting point.

min_roughness is not set in the json file, so will get the default value of 0.08

IIRC, the dancer models were generated in Simplygon as that produced better approximations than Maya. We haven't tested Unreal to generate LODs so cannot comment on the quality there.