How to Save Memory While Filming a Black Hole ✛ ✎

Original: "CoportSL: A Contribution-constrained Hybrid Slow-light Framework for Time-dependent Polarized GRMHD Imaging"

· Fan Zhou, Jiewei Huang, Yuehang Li, Minyong Guo, Bin Chen

arXiv:2609.06607v1 · 2026-09-06 · arXiv non-exclusive · independent analysis · 2 min

An editor decides which plasma frames can be frozen so a black hole movie fits on an ordinary computer.

A cinematographer filming a black hole faces a problem: light from different parts of the disk arrives at the camera at different times. The near edge flares up today, the far edge a week ago, and the jet stretches months behind. If you stitch all frames together without regard to age, you get not a movie but a collage of moments from different times.

To account for this, you need to hold hundreds of 3D plasma snapshots in memory simultaneously. For one M87* model, that means 903 layers occupying 65 gigabytes, too much for an ordinary computer. The new method CoportSL acts like an experienced editor who first scans all footage on fast forward. For each region of space, it evaluates the contribution to emission, absorption, and Faraday effects, then keeps the time dependence only where the contribution exceeds a set threshold. Elsewhere, time is frozen, but light is still traced along the entire ray.

Faraday rotation is when a magnetic field twists the direction of light's oscillation, like a spiral staircase twists a person climbing it. Even weak light, after passing through such a field, carries information about it in its polarization.

In tests with an M87*-like disk, the method reduced the number of snapshots from 903 to 223 for the near region and from 897 to 496 for the jet. The error in the image did not exceed 4 thousandths compared with the full calculation. Memory usage dropped from 65.7 to 20.2 gigabytes for the near image and from 64.1 to 37.3 gigabytes for the jet, fitting within a workstation. The editor returns to the film, now knowing which scenes must not be touched: near the horizon, fast and slow calculations almost coincide, but for the jet, differences in polarization remain significant. This means that ignoring the finite speed of light can lead to a wrong understanding of the magnetic structure near the horizon. The method can already be used to prepare libraries of images for comparison with observations from the Event Horizon Telescope and the future ngEHT.

For M87*, one gravitational radius is about 9.6 billion kilometers, and the time for light to cross it is roughly 8.9 hours. In the paper, delays are measured in these units, and even a 'short' window of 110 gravitational radii means a spread in emission times of almost 41 days.

In the movie 'Interstellar', time delay near the black hole Gargantua is shown as gravitational time dilation, but here it is different: light from different parts of the disk reaches the camera at different times, and without accounting for this, the image would not just be shifted but assembled from pieces of different ages, like a collage.

P_i = (1 - \eta_i) P(x_i, t_a) + \eta_i P(x_i, t_b), \quad \eta_i = \frac{t_{\mathrm{fluid},i} - t_a}{t_b - t_a}

The plasma state at a point along the ray at the required time is obtained as a linear combination of states from the two closest snapshots in time.

\epsilon_S(t) = \sum_{\boldsymbol{x}} \frac{|S(\boldsymbol{x},t) - S_{\mathrm{ref}}(\boldsymbol{x},t)|}{\sum_{\boldsymbol{y}} |I_{\mathrm{ref}}(\boldsymbol{y},t)|}

The sum of absolute differences over all pixels, divided by the total intensity of the reference image, gives a dimensionless error measure.

Scientists

Stephen Hawking Jacob Bekenstein Fritz Zwicky Bernhard Riemann Joseph Weber Karl Schwarzschild

Concepts

black hole Hawking radiation gravitational lensing Bekenstein-Hawking entropy Einstein field equations no-hair theorem Lense–Thirring effect

I am the author

How to Save Memory While Filming a Black Hole ✛ ✎