Dataset Viewer
Auto-converted to Parquet Duplicate
face_size
int64
levels
int64
format
string
faces
string
entries
list
16,384
15
BC6H_UF16 (GL_COMPRESSED_RGB_BPTC_UNSIGNED_FLOAT)
+X -X +Y -Y +Z -Z (GL order)
[ { "level": 0, "face": 0, "offset": 0, "size": 268435456, "w": 16384 }, { "level": 0, "face": 1, "offset": 268435456, "size": 268435456, "w": 16384 }, { "level": 0, "face": 2, "offset": 536870912, "size": 268435456, "w": 16384 }, { "level": ...

audio-reactive-shaders data

Precomputed data for the Kerr-Black-Hole-Visualizer shader in audio-reactive-shaders. Fetch it with that repo's download_data.py (resumable, SHA-256 checked); the paths here match the repo's textures/ layout.

Proof of concept, not maintained.

file what it is
textures/kerr_v7/kn/paths32.npy Null geodesics around a Kerr-Newman black hole (a* = 0.997, Q* = 0.07), 32 path samples per pixel for a 2410x2410 view (kn_bake_modal.py)
textures/kerr_v7/kn/sky.npy Where each pixel's light ends up on the sky, up to 4 images per pixel
textures/kerr_v7/kn_disk_table.npy Accretion disk emission table (kn_disk.py)
textures/starfield/sky64k/* NASA SVS Deep Star Maps 2020 at 64k, as a BC6H cube map with mips (skycube_unpack.py, skycube_bake.py)
textures/pulsar_orrery/*.npy Star, planet and ring surface maps (pulsar_orrery_prep.py)

paths32.npy has one bake defect repaired: 57 rows where sample #9 came out zero.

Sources and licences

  • Geodesic, sky and disk tables: generated by the scripts above, GPL-3.0-or-later like the repo.
  • Star map: NASA/Goddard Space Flight Center Scientific Visualization Studio, Deep Star Maps 2020 (Gaia DR2 (ESA/Gaia/DPAC), Hipparcos-2, Tycho-2). NASA media, public domain.
  • Sun and Saturn textures and the ring profile: Solar System Scope, CC BY 4.0.
  • Chromosphere: NASA STEREO/EUVI + SDO/AIA 304 Å Carrington map. Public domain.
  • Moon surface: NASA LRO / LROC. Public domain.
Downloads last month
32