I've been working on addons for blender (modeling, sculpting). I use millions of vertex positions/normals through Numpy. Which Float Type is better for Blender? float16, float32, float64? What type of floats are in Blender for vertices/UVs?

Right now I'm using float64 like this:

import bpy
import numpy as np

act_obj = bpy.context.active_object
np_test = np.zeros(len(act_obj.data.vertices) * 3, dtype=np.float64)
act_obj.data.vertices.data.foreach_get('co', np_test )


You must log in to answer this question.

Browse other questions tagged .