3 ms·Real-Time NeRF Talking Portrait Synthesis via Audio-Spatial Decomposition1 points by mcoliver 4y ago