Back
Similar todos
render out two videos from neural networks on tiny datasets to test new method of capture & pipeline 
#mused
w/larger datasets, test lowering number images to sample from, but results too blurry to be usable :/ so go back to lowering fps of input data #mused
test & fail 3d environment reconstruction from only 50 images instead of video #mused
test pipeline choosing least blurry image out of every 120 frames for smaller datasets #mused
render from acropolis small model & start processing data for full #mused
try generating 360 video scenes (using vid<>vid from existing 360 video rendered from Unity), not great results 
#3dmystore
training on full dataset is taking too long so export just the Kings Chamber 360s for testing #mused
fix memory bug 
#interiorai by scaling input image smaller
finish processing and aligning data for 360 images for big sur nerf, total about 28 hours (last step 24) #mused
Experiment with rendering performance 
#morflaxstudio
import new 360 video dataset & realize 360 image textures are rendering inverse, so fix sphere/texture issue #mused
Experimenting with faster process to create videos 
#theportal
new method for tighter control of transitions between 360 images for walking around #mused
test new method eq 2 perspectives for processing 360 video frames #mused
