Beautiful capture! Is the black background just the SuperSplat viewer background, or did you remove the museum background with masks before training? If masked, was it done in RealityScan before exporting to COLMAP/LichtLeld, or cleaned afterward? Also, was this captured from video or multi-view still images？