If a black hole traps light, what are we actually seeing?

Start with a real observation

The EHT image uses 2017 radio measurements and was released in 2019. Its orange palette displays radio brightness. It is reconstructed from observations, not visible orange light from a glowing solid surface.

Give a direction a job

Choose a position in the source map. The model traces the corresponding viewing direction inward from a static observer. A returning backtrace connects to a source; physical light follows that connection toward the observer in reverse.

Keep the boundaries distinct

The horizon is a causal boundary. The photon sphere is a radius of unstable circular null orbits in this geometry. Neither is the conserved impact parameter used to label an incoming ray.

Find where it turns

A returning path reaches an outer turning point and then the invented background sphere. Its closest areal radius differs from its impact parameter. A b=6 ray turns near r=4.453, not r=6.

One source, more than one connection

Some paths accumulate an extra winding. The same aligned patch can then form a broad main image and a much narrower additional feature. The source has not been duplicated.

Inspect the narrow feature honestly

The fine inset changes a labeled radial scale. The additional image is far below one pixel in the full impact map. A thick painted ring would hide the actual numerical scale.

Change illumination

With a single bright patch, many returning rays are dark because they sample an unlit direction. The capture classification remains the same when the invented source is recolored.

Ask what the model leaves out

This spherical nonrotating vacuum model has no emitting accretion flow, spin, plasma, telescope reconstruction or physical travel time. It isolates geometry; it is not fitted to M87*.

A one-way causal boundary

In the classical description, the event horizon separates events that can send future-directed light signals to distant observers from those that cannot. It is not a solid surface glowing in the observed image.

One explicit geometry

The Schwarzschild exterior is nonrotating, uncharged and spherical. Set r_g=GM/c². The chosen static observer is at 50 r_g and the invented source sphere at 1000 r_g. These are authored positions, not Earth’s location relative to M87*.

Areal radius is not every distance

A sphere of areal radius r has area 4πr². Plotting x=r cosφ and y=r sinφ gives a useful coordinate projection, but does not reproduce every proper spatial length or local angle in curved geometry.

Exterior null paths

With u=1/r and r_g=1, u″=3u²−u. The first integral is u′²+u²−2u³=1/b². The incoming derivative at the observer is determined by this relation; the code does not use a Newtonian force on a light particle.

The critical distinction

The horizon has r=2, the unstable circular null orbit has r=3 and the critical impact parameter is b_c=3√3. For the chosen inward directions, b<b_c is captured, b>b_c returns outward, and b=b_c approaches the unstable orbit asymptotically.

Static angle and image radius

The static observer measures sinψ=b√(1−2/50)/50. A local camera uses radius proportional to tanψ. An impact map instead uses radius proportional to b. Switching projection keeps the selected physical ray unchanged.

Unwrapped angles matter

The accumulated angle can exceed 360°. First interpolate the full angle, then use its sine and cosine to find a source direction. Interpolating wrapped directions across 360° would produce a false intermediate source.

Closest approach

For a returning path, the outer turning radius satisfies b²=r₀³/(r₀−2), with r₀>3. It is not correct to place the turn at r=b. Finite-endpoint accumulated angle is also different from deflection defined at infinity.

A bounded lookup

The map uses 512 logarithmically spaced knots over 1.001≤b/b_c≤3.1. Independently checked midpoint angular error is below 4.781×10⁻⁶ radians. Returning mappings closer to criticality are deliberately unresolved; a colored or hatched indication is computational scope, not matter.

Two images of a finite patch

For the fixed aligned 2° source patch, centers occur near b=15.337679493438 and b=5.202001640146 r_g. The latter corresponds to an additional winding. Its radial width is about 0.000409681 r_g. Geometry alone does not supply observed radiometric brightness.

Shadow and illumination

Under the stated background illumination, captured backward-traced directions lack a returning source connection. An unlit returning direction can also be dark. The independent status label prevents color from being mistaken for proof of capture.

Observation is reconstructed evidence

The EHT combined interferometric radio measurements. Independent image reconstructions, tests with synthetic data and comparisons across observing days supported the principal structure. Reconstruction does not make the evidence an arbitrary invented drawing.

A measured size is not a direct horizon ruler

The 2019 analysis reports an emission-region diameter of 42±3 microarcseconds. Inferring mass or a horizon scale also uses distance and emission modeling. The published broad structure did not separately resolve every predicted photon subring.

Compactness is not unlimited suction

A hypothetical object with the Sun’s same gravitational mass concentrated into a black hole would not automatically pull Earth out of its distant orbit. This is a same-mass thought experiment, not the Sun’s predicted future.

Reading scientific images

Ask what the instrument measured, what reconstruction was required and which colors were assigned for display. Those questions apply to radio astronomy, medical imaging and microscopy.

Lensing as a mapping problem

Connect an image direction to possible sources before interpreting multiple features as multiple objects. The same source may have several exterior light-path connections.

Testing a scientific computation

Compare independently formulated calculations, inspect boundary cases and state where resolution ends. Numerical agreement is evidence about an implementation, not automatic confirmation of every physical assumption.

What supplies light in the EHT image?

• A luminous solid horizon
• Emitting matter outside the horizon
• Light escaping from inside

Which quantity is an areal-coordinate radius?

• b_c=5.196 r_g
• Photon-sphere r=3 r_g
• 42 microarcseconds

In this model, where does the returning b=6 ray turn?

• r=6
• r≈4.453
• Inside r=2

Two image features from the same patch mean…

• The matter was duplicated
• There are different exterior connections
• The light escaped through the horizon

What does the near-critical hatching mean?

• A material shell
• An individually measured photon ring
• Background mapping is unresolved

The EHT image released in 2019 was…

• An orange-light snapshot taken in 2019
• A reconstruction of 2017 radio observations
• A purely invented drawing

Our circular capture boundary establishes…

• M87* has zero spin
• The symmetry chosen in our model
• The brightness of the patch

In single-patch mode, a dark direction…

• Always proves capture
• Can also connect to an unlit source
• Must be the photon sphere

Exterior null geometry and strong-deflection limit

Bozza 2002, sections II and III A, equations 38–50. Its length unit is 2 r_g; conversion is explicit. No figures copied.

Static observer angle and vacuum shadow

Perlick, Tsupko and Bisnovatyi-Kogan 2015, equations 38 and 47. Only the vacuum Schwarzschild limit is implemented; plasma extensions are omitted.

Historical escape-cone result

Synge 1966, MNRAS 131, 463–466. Historical primary treatment recovered by the cited modern vacuum formula; no article figures copied.

Observed image, causal boundary and reconstruction

EHT 2019 Paper I, introduction and sections 5–7, Figure 3. The separate public JPEG has its own item provenance and CC BY 4.0 image license.

Image-size measurement and model-dependent inference

Paper VI reports the 42±3 microarcsecond emission diameter and relates image size, distance, models and inferred mass.

Broad emission structure versus much narrower subrings

Johnson et al. 2020, introduction and section 2. Do not claim every predicted subring was separately resolved in the 2019 image.

Accretion context and outside emission

NASA anatomy page explains emitting matter and lensing. The implemented horizon uses the more precise causal-boundary definition; no third-party NASA-hosted artwork is copied.

Same-mass thought experiment

NASA Goddard distinguishes black holes from indiscriminate suction and discusses hypothetical replacement by the same solar mass, not the Sun’s predicted fate.

Exact image identity, date, wavelength and credit

ESO eso1907a identifies the 2017 1.3 mm observation campaign, 10 April 2019 release and EHT Collaboration credit. Original served publication JPEG and SHA-256 are retained locally.

Visual reuse conditions

ESO public image policy: CC BY 4.0 unless excepted, full visible credit and no endorsement. The item has no separate exception. This does not relicense scientific papers or code.