Plate Nº 77 · recorded November 15, 2026
Space & AstronomyReported finding
Independent Check Confirms First Black Hole Photo Got It Right
A Nature study re-analyzed the original EHT data with a new method and confirmed M87's black hole ring, about 55 million light-years away, is real and the right size.
By Nathan Brooks3 min read630 words
In brief
- The first black hole image was released on April 10, 2019, showing the supermassive object in galaxy M87, about 55 million light-years away.
- The new independent analysis, published in Nature, confirmed the ring's diameter matches a black hole of 6.5 billion solar masses.
- The team's 'closure invariants' method combines triplets of telescope measurements instead of pairs, canceling atmospheric and instrumental errors.
- The validated machine learning imager, GenDIReCT, was trained on cats, dogs and cars — not astronomical images.
- The method was applied to the M87 data on December 4, 2025, after training and testing were frozen to prevent bias.
The first-ever image of a black hole, released on April 10, 2019, has passed an independent audit. A new study published in Nature re-analyzed the original Event Horizon Telescope (EHT) data using a fresh technique, and the iconic ring came out the same.
The object in question sits at the center of galaxy M87, about 55 million light-years from Earth. Its shadow, outlined by light from swirling hot gas, matched Einstein's general relativity predictions and became one of the most recognized scientific images of the modern era.
Not everyone was fully convinced, though. Scientists have since disagreed over the ring's fine details, and one study went as far as challenging whether the source is ring-like at all.
Why did the image need checking?
Photographing M87's black hole is, by the researchers' own comparison, like photographing a doughnut on the moon from Earth. To get that sharpness, the EHT combined signals from radio telescopes across the globe into one planet-sized virtual instrument.
That approach has built-in fragility. Four independent imaging teams worked on the 2019 publication, and all produced a ring — but with differences in the fine-scale features, caused by known human and algorithmic biases in the imaging process. Combining sparse telescope data, instrument errors, and distortions from Earth's atmosphere all leave residuals that can skew the final picture.
What is the new method?
The team developed an approach called "closure invariants." Standard analysis combines pairs of telescope measurements; closure invariants combine triplets — three measurements at once. That triple combination eliminates the atmospheric and instrumental errors and, the authors say, reveals cleaner information about the object's structure.
To turn that cleaner data into an image, they built a machine learning imager called GenDIReCT. To avoid teaching it any preconceptions, they trained it on everyday terrestrial pictures — cats, dogs and cars — so it had no idea what a black hole looked like.
They then stress-tested it:
- A large suite of validation tests, including imaging a wide range of non-ring structures
- Validation on real data sets of other galaxies, such as Centaurus A
- A frozen training and testing procedure, locked before the real run, so human biases couldn't creep into the results
On December 4, 2025, the fully validated method was finally allowed to "look" at the EHT observations of M87.
What did it find?
Across all data sets, the ring structure was consistent. Its appearance and diameter matched a black hole weighing 6.5 billion suns, observed nearly face-on — the researchers liken it to looking straight down into the eye of a hurricane.
Compared directly with the EHT Collaboration's images, the new result agrees on diameter and overall orientation. There are differences in the fine details of ring width and brightness along the circumference, but the overall structure matches — which the authors say resolves the earlier discrepancies.
To show the answer didn't hinge on machine learning alone, the team also built a second, complementary method based on traditional statistical techniques. That entirely different approach validated the same structure using the new closure invariant data.
Why does it matter?
The study delivers a strong independent confirmation that the center of M87 really does host the supermassive black hole the 2019 image claimed — reinforcing one of the most significant scientific results of recent decades.
The technique may travel beyond astronomy, too. The authors suggest it could help wherever data from multiple sensors are combined to map unknown structures, citing medical imaging and earthquake mapping as candidate fields.
As with any imaging study, the fine-detail differences between methods remain, and the analysis covers the M87 data sets only. But on the headline question — is the ring real and the right size? — two independent pipelines now agree.
via phys.org/rss-feed/ (Source)
More from Nathan Brooks
Show full bio
Market editor covering companies, deals and strategy in general science news at SciBeat.
375 articles
Nearby plates
- Independent Method Confirms the First Black Hole Image of M87*
- Euclid's First Data Release Images 900 Million Galaxies
- Euclid Telescope Logs Nearly a Billion Cosmic Objects in First Big Data Drop
- Webb Telescope Spots Unexpected Shifts in Chariklo's Rings Over Just a Few Years
- Hubble Images Gravitational Lens to Prepare for Roman Telescope