Sora Video ArchiveI picked 3 random modules
{
"id": "material_and_microstructure_l... by @reflectgen
I picked 3 random modules
{
"id": "material_and_microstructure_library",
"description": "High-fidelity material BSDFs and microtexture maps to match lens-resolved detail.",
"required_keys": [
"material_id",
"albedo_spectrum_wavelength_samples",
"micronormal_heightmaps (sub-mm scale)",
"roughness_and_anisotropy_maps",
"specular_F0_values",
"subsurface_scattering_profiles"
],
"critical_datasets": ["measured_BSDFs","microdetail_texture_bank.exr"],
"usage": "Supports macro/micro shots where lens resolves tiny surface detail; essential for believable close-ups."
},
{
"id": "occlusion_parallax_and_depth_synthesis",
"description": "Algorithms and data to synthesize plausible occlusion and parallax for tele/wide conversions and remapping.",
"required_keys": [
"layered_depth_maps",
"multi-plane_image_specs",
"parallax_preservation_constraints",
"occluder_edge_smoothing_rules"
],
"critical_datasets": ["depth_layers_samples","parallax_kernels"],
"usage": "Crucial for pushing lens-specific parallax behaviors in CG and converting between focal lengths."
},
{
"id": "artifact_and_degradation_bank",
"description": "Comprehensive set of camera and lens imperfections to add authenticity.",
"required_keys": [
"dust_spots_masks (per sensor format)",
"lens_scratch_maps",
"sensor_hot_dead_pixel_lists",
"db_of_flares_by_angle_wavelength",
"compression_artifact_profiles (codec-specific)"
],
"critical_datasets": ["flare_bank.exr","dust_spots_catalog","codec_profiles.json"],
"usage": "Layered application of real imperfections to prevent 'CG-clean' look."
}
SoraStats is an independent archive and analytics site, not affiliated with OpenAI. Video availability reflects public archive data and may differ from source platform availability.
SoraStats interpretation
Archive analysis
This stable archive record combines playback availability with the latest stored public engagement snapshot, so the video can be compared with other Sora-era work even when source availability changes.
- Likes per view
- 1.6%
- Remixes per view
- 0.3%
Ratios are descriptive, not quality scores. Counts may lag the source platform, and deleted or private activity is outside archive coverage.