README: note the six transparent cases

README.md

@@ -38,6 +38,8 @@ The track queries are the benchmark's grid of at most 2048 points on frame 0: 51
 
 23 real videos cannot be redistributed and are rebuilt by Dataset-Real-World's `prepare_videos.py`, whose output can slightly differ from the benchmark's copy by H.264 encoder noise. The estimates here are of the benchmark's copy.
 
+Six cases show their main object through a transparent surface: `internet_01_purple_granular_sand`, `internet_19_fabric_laundry_tumbles`, `internet_22_viscous_honey_dripping`, `phys101_01_solid_object_dropped`, `wisa_80k_06_pouring_red_wine` and `wisa_80k_46_wooden_pestle_crushing`. The video shows what lies behind the surface while a reconstruction rasterises the surface itself, so the benchmark does not score Flow, Track2D, Depth error or Uni3D MoGe on them. Their files are released for completeness only, and the models can fail on them; RAFT's flow of `wisa_80k_06_pouring_red_wine`, for one, puts hundreds of pixels of motion on its static background.
+
 ## Synthetic cases
 
 A synthetic case has the world it was rendered from, so its depth, flow and tracks are computed exactly rather than estimated. `scripts/extract_synthetic.py` writes them in the same layout as the real cases, from the meshes, material trajectories and camera of each world, with the same code the scorer uses to read a submission. No model checkpoint is needed.