The Repository Absolute Magnitude Space

Betelgeuse Lost Two Thirds of Its Light. One Number Was Measured.

September 19, 2026 · Runtime 13:18 · Watch on YouTube

In the winter of 2019-2020 Betelgeuse faded to visual magnitude 1.614, about 36% of its usual light, and the headlines said it was about to explode. The photometer returned one brightness in one band. Everything else was converted afterwards.

Transcript

Two thirds of the light, gone 0:00

In the winter of two thousand nineteen, one of the ten brightest stars in the sky lost almost two thirds of its visible light. By late December it had slipped from the sixth or seventh brightest star to about the twenty first. The headlines said it was about to explode. This episode is about what was actually measured.

One brightness, one band 0:21

What the instrument returned that winter is a brightness. One band of colour, night after night, from a small telescope in Pennsylvania. This episode ends on that number, and on the one set of wavelengths where the star did not fade at all.

Last episode used Betelgeuse as a ruler: two billion times brighter than a star beside the galactic centre. That ruler is not steady. It swings, on a clock of about four hundred days. And in February two thousand twenty it swung further than anyone had recorded.

The photometer: 25 years on one star 0:57

The instrument is a photometer. A telescope, a filter, and a detector that counts light. Edward Guinan and Richard Wasatonic, at Villanova University, had pointed one at Betelgeuse for twenty five years, through a V filter, the band closest to the human eye.

Magnitudes run backwards 1:16

It reports magnitudes, and magnitudes run backwards. A bigger number is a fainter star. Each step of one magnitude is a factor of about two and a half in light, and five magnitudes is a factor of one hundred. Betelgeuse usually sits between zero point one and one point zero.

The light curve, night by night 1:37

Here is what the photometer returned. In September two thousand nineteen the star read about zero point six. On the seventh of December it read one point one two, the faintest in fifty years of photoelectric records. By the twenty second of December, one point two nine. By the seventeenth of January, one point four nine.

The mean over the seventh to the thirteenth of February was one point six one four, plus or minus zero point zero zero eight. That is the floor. By the twenty second it was rising, by the end of March it was climbing two hundredths of a magnitude a day, and by April the star was back where it started.

The honest conversion: 36% 2:23

The first conversion is the honest one. One point one magnitudes below its ordinary half magnitude is a factor of two point eight in light. Betelgeuse was sending us thirty six per cent of its usual visible light. That figure is arithmetic on the measurement, and nothing else.

The light curve as sound 2:43

Before the other conversions, listen to the measurement. This is the light curve as sound. Each night's brightness becomes a pitch, with the frequency in proportion to the light received: four hundred and forty hertz for the ordinary star, one hundred and fifty seven at the floor. One day of calendar becomes a twentieth of a second.

Two hundred days, from September to April, in ten seconds. Nothing else is encoded. What you hear falling is the number, and only the number.

The other conversions: a countdown 3:27

Then came the other conversions, and they were not arithmetic. A star this big, fading this fast, was read as a star about to collapse. Betelgeuse will explode one day. It is a red supergiant, and it will end as a supernova. The dimming was reported as a countdown.

A brightness in one band cannot say that. It cannot tell a star cooling from a star hiding behind something. And it cannot tell a change on the star from a change in front of it. To separate those you need other instruments, and four of them were pointed at Betelgeuse that winter.

Infrared: unchanged in fifty years 4:08

The first was a photometer again, but in the infrared. Robert Gehrz's group at Minnesota measured Betelgeuse in eight bands, from one point two to eight point eight microns. The star was unchanged, to within a tenth or two of a magnitude. In their words, the magnitudes were the same as they were fifty years ago.

Most of a red supergiant's energy comes out in the infrared. A star that had cooled all over, by enough to lose two thirds of its visible light, would have dimmed in the near infrared as well. Both models the Nature team ran predict a drop of twenty to thirty per cent in the J band. The measured drop was two per cent.

That is the K band from last episode, where the ruler was drawn. In the visible, the star had lost two thirds of its light. At two point two microns it had barely moved. Whatever happened, happened to the visible.

Spectrograph: fifty kelvin cooler 5:09

The second instrument was a spectrograph. On the fifteenth of February, Emily Levesque and Philip Massey took the spectrum of Betelgeuse at Lowell Observatory, with the same setup they had used in two thousand four. Titanium oxide bands in a red star's spectrum are a thermometer. The cooler the star, the deeper the bands.

The bands gave three thousand six hundred kelvin, plus or minus twenty five. In two thousand four, the same method gave three thousand six hundred and fifty. Fifty kelvin cooler. And fifty kelvin is not enough.

Cooling explains a third, at most 5:49

Run it through the models. Fifty kelvin of cooling dims the V band by about zero point one seven magnitudes. Even a hundred kelvin, the most their error bars allow, gives zero point three eight. The star had lost one point one. Cooling of the whole star explains, at most, a third of it.

Interferometer: the star did not shrink 6:12

The third instrument was an interferometer: GRAVITY, the same one that drew last episode's orbit. It measured the star's angular size. Forty two point six one milliarcseconds in January two thousand nineteen, and forty two point one one in February two thousand twenty. To dim by that much by shrinking, the star would have had to lose thirty per cent of its diameter. It lost one.

A radius that changed twice, unmeasured 6:40

Notice what happened to the radius before that measurement existed. On the twentieth of January the Villanova telegram converted the fading into a luminosity down twenty five per cent and a radius up nine per cent. Twelve days later, with a cooler temperature adopted, the same photometry gave a radius down eight per cent. Neither was a measurement of the radius. Both were a brightness, divided by an assumed temperature.

Camera: one hemisphere ten times darker 7:07

The fourth instrument was a camera. SPHERE, on the Very Large Telescope, with adaptive optics sharp enough to resolve the disc of the star. Miguel Montargès had imaged Betelgeuse in January two thousand nineteen. He imaged it again in December, in January, and in March.

In December and January, in every filter, the southern hemisphere of the star had a peak brightness more than ten times lower than the northern. The star had not dimmed. Half of it had. By March the dark region sat in the south western quadrant, and the whole star was on its way back.

Hubble saw it leave, months before 7:48

There was a fifth instrument, and it had looked before anyone knew to look. Andrea Dupree's group had been taking ultraviolet spectra of Betelgeuse with Hubble, resolving the disc into quarters. From September to November two thousand nineteen, before the dimming, the southern hemisphere brightened in ultraviolet, its outer layers hotter and denser, with material moving outward.

Something left the south side of the star months before the light dropped. The photosphere was expanding, and the outflow rode the outward phase of the four hundred day pulse. None of that is in the visible light curve. It is in a spectrum taken for other reasons.

Put the four together. Infrared unchanged, so the star as a whole did not cool. Diameter unchanged, so it did not shrink. Spectrum fifty kelvin cooler, so it cooled a little. And a camera showing that the darkness was on one side, where Hubble had seen gas leave.

A cool patch and a dust cloud 8:52

The Nature paper's answer is two things at once. A cool patch on the southern surface, and in front of it a cloud of dust that had condensed from gas the star had thrown off. Cool the gas enough and the dust forms. The dust blocks the visible light and lets the infrared through.

Where the model begins: a factor of four 9:12

Here is where measurement ends and modelling begins, and the paper says so. Fitting the images: a cool patch at three thousand four hundred kelvin, against a photosphere of three thousand seven hundred, covering sixty two per cent of the disc in December and seventy nine per cent in January. And a dust clump four and a half to six astronomical units in radius, about the size of the star itself.

Read the error bars. The dust in the clump is three to thirteen ten billionths of a solar mass: a factor of four. With its gas, the clump is between thirty five per cent and a hundred and twenty eight per cent of what Betelgeuse loses in a typical year. None of that is the photometer. It is what the photometer becomes once a model of dust is added.

A weather satellite settles the split 10:03

One more instrument saw the whole thing by accident. Himawari eight is a Japanese weather satellite, parked over the equator, photographing the Earth's disc every ten minutes in sixteen bands. Betelgeuse sits at the edge of its frames. Daisuke Taniguchi's group pulled four and a half years of the star out of the weather data, from half a micron to thirteen.

With the mid infrared, where warm dust glows, they could weigh the two causes against each other. Their answer: the cooling of the surface and the extra dust contributed by almost the same amount. A satellite built to watch typhoons gave the split that the ground had argued over for two years.

The 400-day clock was on schedule 10:47

And the timing was ordinary. The dimming came four hundred and twenty four days after the previous minimum, plus or minus four, right on the star's four hundred day pulse. The clock was on schedule. Only the depth was new. The pulse and the cloud pulled the same way at the same time.

Afterwards the clock did not run the same. Dupree's group reports that the four hundred day pulsation vanished, from the optical and from the star's velocity, for more than two years after the dimming. The ejection had left the star ringing differently.

The supernova: nothing measured 11:25

So, the explosion. Nothing measured that winter bears on it. The models place Betelgeuse in core helium burning, and the red supergiant phase lasts on the order of a hundred thousand years. The dimming was a cloud and a cool patch on one side of a star that pulses every four hundred days, and it was over by April.

Line the numbers up. Thirty six per cent of the light: measured, in V. Infrared unchanged: measured. Fifty kelvin cooler: measured, through a model of the spectrum. Diameter unchanged: measured. One hemisphere ten times darker: imaged. A dust cloud of a few ten billionths of a solar mass: modelled, to a factor of four. A supernova: nothing.

2026 footnote: Betelgeuse B 12:17

One footnote from this year. In July, the same SPHERE camera reported a companion, imaged in December two thousand twenty four: Betelgeuse B, two to three times the mass of the Sun. It does not explain the dimming. The paper is clear that the dust did that. But the star that was used as a ruler has turned out to be two.

So the practical version. When a star is reported dimming, ask in which band, and whether the infrared moved. If the visible fell and the infrared did not, you are looking at something in front of the star, or on one side of it. Not at its engine.

1.614 magnitudes. Next: the distance 12:59

One point six one four magnitudes, plus or minus zero point zero zero eight. The visual brightness of Betelgeuse in the second week of February two thousand twenty, read by a photometer that had watched it for twenty five years. Everything else that winter, the cooling, the cloud, the countdown, was built on top of it. Next time: how far away Betelgeuse actually is. A star ten times wider than its own parallax, and a distance nobody has pinned down. Absolute Magnitude Space. Every episode ends in a measured number.

Description and sources

Villanova's 25-year V-band photometry of Betelgeuse reached V = 1.614 +/- 0.008 over 7-13 February 2020: 36% of the ordinary star's light, which is arithmetic. 'Cooling', 'shrinking' and 'a countdown' are not. Infrared photometry from 1.2 to 8.8 microns was unchanged, the same as fifty years ago; a spectrum gave 3600 K against 3650 in 2004, enough for 0.17 of the 1.1 magnitudes lost; GRAVITY measured the angular diameter at 42.61 and then 42.11 milliarcseconds, a 1% change where 30% would have been needed. SPHERE's images show the southern hemisphere more than ten times darker than the northern, and Hubble had seen that hemisphere brighten in ultraviolet and eject material months before. The fit is a cool patch under a fresh dust clump known to a factor of four. Nothing measured that winter bears on a supernova. The episode ends on the number the photometer returned: 1.614 +/- 0.008.

PRINT-READY, FROM THIS CHANNEL

The Conversion Chain, Vol. 01 — the method, and four numbers taken apart with it

https://therepository.gumroad.com/l/the-conversion-chain

PRIMARY SOURCES

Montarges, M. et al. (2021). A dusty veil shading Betelgeuse during its Great Dimming. Nature 594, 365, doi 10.1038/s41586-021-03546-8 -- V = 1.614 +/- 0.008 (7-13 Feb 2020); southern hemisphere >10x darker (SPHERE); theta_UD 42.61 -> 42.11 +/- 0.05 mas; J band predicted 1.2-1.3x, observed 1.02; cool patch 3,400 K on 3,700 K, 62/79/43% of the disc; dust (3-13) x 10^-10 Msun, clump 35-128% of a year's mass loss; 424 +/- 4 d after the previous minimum.
Guinan, Wasatonic, Calderwood, Carona -- ATel 13341 (V = 1.12, 7 Dec 2019), 13365 (1.27-1.29, 19-22 Dec), 13410 (1.49-1.51, 17-18 Jan; L -25%, R +9%), 13439 (1.614 +/- 0.012, 30 Jan; R'/R ~0.92), 13512 (mean 1.614 +/- 0.008, 7-13 Feb; 424 +/- 4 d after the Dec 2018 minimum). Sigismondi, ATel 13601 (V = 0.93 by 31 Mar, 0.02 mag/day).
Gehrz, R. D. et al. (2020). Betelgeuse remains steadfast in the infrared. ATel 13518 -- 1.2-8.81 microns, 'the same as they were 50 years ago'.
Levesque, E. M. & Massey, P. (2020). Betelgeuse Just Is Not That Cool. ApJL 891, L37 -- T_eff 3600 +/- 25 K (15 Feb 2020) vs 3650 +/- 25 K (2004); 50 K -> ~0.17 mag, 100 K -> ~0.38 mag.
Dupree, A. K. et al. (2020). ApJ 899, 68 -- HST/STIS: southern-hemisphere UV and Mg II brightening, outflow, Sep-Nov 2019. Dupree et al. (2022). ApJ 936, 18 -- the ~400-day pulsation absent for more than two years after.
Taniguchi, D. et al. (2022). Nature Astronomy 6, 930 -- Himawari-8, 16 bands 0.45-13.5 microns, 4.5 years: cooling and dust 'by almost the same amount'.
Joyce, M. et al. (2020). ApJ 902, 63 -- core helium burning; R = 764 (+116/-62) Rsun; d = 168 (+27/-15) pc.
ESO eso2003 (14 Feb 2020) and eso2109 (16 Jun 2021) -- '36% of its normal brightness'; back to normal by April 2020. Montarges et al. (2026), A&A 711, L12; ESO eso2611 (28 Jul 2026) -- Betelgeuse B, 2-3 solar masses; the dimming was dust, not the companion.
Harper, G. M. et al. (2017). AJ 154, 11 -- parallax 4.51 +/- 0.80 mas, 222 (+48/-34) pc; angular diameter 44 mas (next episode).

ARCHIVE PLATES: all 20 are credited on screen in the video and listed in this project's manifest.

Absolute Magnitude takes one number the internet repeats and asks what an instrument actually returned. Every episode ends on a measured figure with its source on screen, and says plainly which part of it was converted afterwards.

No AI-generated imagery is used on this channel. Every frame of archive footage is real, publicly released material, credited on screen with its licence; every diagram is drawn in code from the numbers in the cited papers. The sound of the light curve is synthesised from the published magnitudes with the mapping shown on screen.

Corrections are pinned, not quietly edited.

#Betelgeuse #Astronomy #Orion #Space