Hi there!
Thanks for your awesome work!
But I met some problems in conversion.
In my case, I loadSMPL_m_unityDoubleBlends_lbs_10_scale5_207_v1.0.0.fbx for conversion while in this line and this line, the lRootNode.FindChild returns null values. And I found all child names exactly start with 'm_avg_' (or 'f_avg_' for the SMPL_f_...fbx).
Then after adding the prefix, the conversion can be processed. I wonder if it is the right operation.
Besides, I found the final .fbx file is large where motion information (rotation, global velocity, etc) only occupies a little space (i.e. 300MB for .fbx and 330kB motion for .npy). Also, for the CMU-MOCAP .fbx file, a sequence is also small. I wonder if is there any extra information restored.
2020-3-2 fbx python SDK is used.
Hi there!
Thanks for your awesome work!
But I met some problems in conversion.
In my case, I load
SMPL_m_unityDoubleBlends_lbs_10_scale5_207_v1.0.0.fbxfor conversion while in this line and this line, thelRootNode.FindChildreturns null values. And I found all child names exactly start with'm_avg_'(or 'f_avg_' for the SMPL_f_...fbx).Then after adding the prefix, the conversion can be processed. I wonder if it is the right operation.
Besides, I found the final
.fbxfile is large where motion information (rotation, global velocity, etc) only occupies a little space (i.e. 300MB for .fbx and 330kB motion for .npy). Also, for the CMU-MOCAP.fbxfile, a sequence is also small. I wonder if is there any extra information restored.2020-3-2 fbx python SDK is used.