loading...
running...
How to use
Drag to orbit, wheel to zoom, camera buttons for presets. Things to try: (1) Flat space preset, then Wide view: see how gravity wraps the far side of the disc over the hole. (2) Edge-on: the upper and lower arcs are the back of the disc seen over the top and under the bottom. (3) Photon ring with the overlay on: the thin bright ring hugs the 1.5 Rs circle, and the white curve is the analytic shadow edge. (4) Raise the spin to 0.95 and compare the inner edge and the bright side. (5) Set the object to a stellar-mass hole and watch the tidal readout while you drag the probe orbit to 3 Rs. Space pauses, R resets time.
Textures (star field, disc plasma) and the probe with its baked PBR materials were made in Blender; Three.js r160 and a WebGL fragment shader do the live ray tracing.
The science
Geometry. Outside a non-rotating mass M the Schwarzschild metric has an event horizon at the Schwarzschild radius Rs = 2GM/c² (about 2.95 km per solar mass) Light can orbit on an unstable circle at 1.5 Rs (the photon sphere), and the innermost stable circular orbit of a massive particle is at 3 Rs = 6GM/c².
Bending of light. With u = 1/r and azimuth φ, a light ray obeys d²u/dφ² = -u + (3/2) Rs u² far away the deflection is 4GM/(c²b) = 2Rs/b (Einstein 1915); rays with impact parameter below bc = (3√3/2) Rs = 2.598 Rs are captured. That circle of capture is the black hole's shadow; for an observer at radius r its apparent half-angle is sin α = bc/r · √(1 - Rs/r).
The renderer. Every pixel of the picture starts a light ray at the camera and follows it backwards through the curved space. In Cartesian form the same equation reads a = -(3/2) Rs h² x / r⁵, h = |x × v| and is integrated with a midpoint step that shrinks near the hole. A ray that falls inside r = Rs is black; a ray that crosses the equatorial plane between the disc's inner and outer radius picks up the disc's emission (several crossings give the secondary and tertiary images: the thin bright rings and the halo above and below an edge-on disc); a ray that escapes reads the star map in its final direction. The camera's local view angle is converted to the coordinate direction by scaling the radial part by √(1 - Rs/r), which is what makes the shadow size match the formula above.
The disc. A thin Shakura-Sunyaev disc radiates as a black body whose temperature falls as T(r) = [3GMṀ/(8πσr³) (1 - √(rin/r))]1/4 with a maximum at 49/36 rin. Its inner edge is the ISCO: 6M for a = 0 and, for spin a* (Bardeen-Press-Teukolsky), risco = 3 + Z2 - √((3 - Z1)(3 + Z1 + 2 Z2)) down to 1.24 M at a* = 0.998; the efficiency 1 - Eisco grows from 5.7 % to about 32 %. Each patch is Doppler-boosted and redshifted by g = √(1 - Rs/r) / [γ(1 + β φ̂·n̂)], brightness ∝ g3 to g4, colour from a blackbody at g T. The approaching side of the disc is therefore brighter and bluer.
The probe moves on an exact circular geodesic: Ω = (M/r³)1/2 in Schwarzschild time (Kepler's third law keeps its Newtonian form), and its own clock runs at dτ/dt = √(1 - 3M/r), so 71 % at the ISCO. Static clocks tick at √(1 - 2M/r).
Honest simplifications. (1) The ray tracer is exact Schwarzschild; spin only changes the disc's inner edge, efficiency and brightness, not the shadow shape or frame dragging (a Kerr tracer needs the Carter constant and is not done here). (2) The Doppler angle uses coordinate directions rather than a full local-frame transformation. (3) The disc is a thin opaque sheet with a painted turbulence texture, not a magnetohydrodynamic simulation; the display colour temperature slider decouples the visible palette from the physical peak temperature printed in the readout (for stellar-mass holes that is in the X-ray). (4) The probe is drawn without lensing and hidden while it is behind the shadow. (5) Quantities such as mass, Rs and time scales are real for the chosen object; the ship, 0.4 Rs wide, is exaggerated.