airsplay/R2R-EnvDrop

Difference between aug_paths and Speaker-Follower (Literal) Paths?

siddk opened this issue · 2 comments

siddk commented

The "augmented" paths in the download file (http://www.cs.unc.edu/~airsplay/aug_paths.json) do not seem to be the "ground truth" paths from the original Speaker-Follower paper (http://people.eecs.berkeley.edu/~ronghang/projects/speaker_follower/data_augmentation/R2R_literal_speaker_data_augmentation_paths.json).

Is there a difference? If so, what is it?

The augmented paths of Speaker-follower are not released when this project is developed. We thus collected these augmented paths on our own following the setup in their paper.

The paths are always the shortest paths from the beginning to the end thus the intermediate viewpoints are not logged.

siddk commented

Understood - thanks so much!