The Repository Absolute Magnitude Space

Betelgeuse Is 10 Times Wider Than Its Own Parallax

September 27, 2026 · Runtime 12:21 · Watch on YouTube

Betelgeuse has been given six published distances in thirty years, from 131 to 222 parsecs, and every one of them is a measurement. What the instrument returns is an angle of about 4.5 milliarcseconds, and the star's own disc is 44 milliarcseconds across.

Chapters

Transcript

Nobody knows how far away it is 0:00

This is one of the most observed stars in the sky. You can find it tonight, with your own eyes, in the shoulder of Orion. Astronomers have pointed almost every kind of instrument ever built at this one object. And not one of them can tell you how far away it is. That is not a figure of speech.

Six published distances 0:21

Here is the distance to Betelgeuse, as actually published. One hundred and thirty one parsecs. One hundred and fifty two. One hundred and sixty eight. One hundred and ninety. One hundred and ninety seven. Two hundred and twenty two. Six numbers, thirty years, and every one of them is a measurement.

None of them is wrong, and that is the interesting part. What an astrometric instrument returns is not a distance at all. It returns an angle. The distance is what somebody computed afterwards. This episode ends on that angle, and on a second number that had to be invented before the first one would fit.

The chain: angle, parallax, distance 1:08

The chain from the instrument to the headline is short, and that is exactly what makes it fragile. A detector records where the star appears. Across a year that position traces a small ellipse. The size of the ellipse is the parallax. One divided by the parallax is the distance. Everything else is downstream of that single division.

The baseline is Earth's orbit 1:33

The baseline is the Earth's own orbit. In January we look from one side of it, in July from the other, three hundred million kilometres across. A nearby star shifts against the far background over those six months, and the size of that shift is the whole measurement. Nothing else is being observed.

The unit comes straight out of the geometry. A star whose shift is one second of arc sits at one parsec, about three and a quarter light years. The nearest star to the Sun shifts by three quarters of a second. Betelgeuse shifts by four and a half thousandths of one second of arc.

The star is a disc, not a point 2:14

That is the signal. Now here is the object producing it. Betelgeuse is close enough, and large enough, that interferometers resolve it as a disc rather than a point. That disc measures about forty four thousandths of a second of arc across.

44 milliarcseconds against 4.5 2:32

Forty four, against four and a half. The star is nearly ten times wider on the sky than the yearly wobble anyone is trying to detect inside it. It is like measuring how far a tower sways by watching one window, when the tower is ten times wider than the sway itself.

Centre of light, not centre of mass 2:52

That matters because of what astrometry actually locks onto. It does not find the centre of mass. It finds the centre of light. For a small, round, steady star those two points are the same, and nobody has to think about the difference. Betelgeuse is none of those three things.

Its surface is not uniform and it does not hold still. It boils in convective cells that are enormous, comparable in size to the orbit of the Earth, and they brighten and fade over months. Whenever one hemisphere runs brighter, the centre of light slides away from the centre of mass.

Hot spots moved it during the mission 3:32

This is not a hypothesis added later to explain a bad result. Interferometry in nineteen ninety one and nineteen ninety two caught hot spots dragging the photocentre to position angles of thirty nine degrees, and then to minus twenty nine. Those years are when Hipparcos was observing.

Cosmic noise: the term that had to be added 3:52

So the fit does not close. To get any solution at all, the Hipparcos analysis had to add an extra term of scatter on top of every measurement. It is called cosmic noise. In the nineteen ninety seven catalogue it was three point four thousandths of a second of arc. In the two thousand seven reanalysis, two point four.

Hold that against the signal. The parallax being measured is about four and a half. The correction term added to make it fit is two point four. More than half the size of the quantity itself. And the paper is candid about what that term is: its origin, it says, remains unknown.

1,561 stars needed it, then 588 4:37

It is not that the catalogue was careless. The nineteen ninety seven release needed this term for one thousand five hundred and sixty one single stars. The two thousand seven reanalysis was good enough to cut that list to five hundred and eighty eight. Betelgeuse stayed on it.

The published series, with error bars 4:56

Here are the six published values, with their error bars, in the order they appeared. Look at what they do rather than at any one of them. The bars do not cover each other. The radio solution and the revised Hipparcos solution sit two point six standard deviations apart.

Too bright for Gaia to observe 5:15

At this point the obvious answer is Gaia, the satellite that measured more than a billion parallaxes at a precision Hipparcos could not approach. Gaia cannot do this one. Its detectors saturate above about third magnitude, and Betelgeuse is far brighter than that.

The best parallax machine ever flown is blind to the brightest stars in the sky, and the paper says so in a single line. Betelgeuse is very bright, too bright for Gaia to observe. The star everyone can see without a telescope is missing from the catalogue.

28 months against 34 years 5:52

What is left is time. Hipparcos surveyed the whole sky for three point two years, and within that it watched Betelgeuse for less than twenty eight months. Radio telescopes hold positions on this star from nineteen eighty two to twenty sixteen. Thirty four years, roughly fifteen times the baseline.

There is a second reason the long record matters. Betelgeuse drifts across the sky at between ten and twenty five thousandths of a second of arc every year. That is a sizeable fraction of its own disc, annually, and it dwarfs the parallax wobble hidden inside that drift.

Why radio sees a calmer layer 6:32

Radio changes more than the length of the record. At millimetre wavelengths you are no longer looking at the boiling visible surface. You are looking at a layer well above it, where brightness responds gently to temperature instead of exponentially, and the centre of light moves less.

So the two thousand seventeen solution combines the older array positions with new ones from e-MERLIN in England, and from ALMA. That is sixty six antennas on a plateau five thousand metres up in the Atacama, linked so the array works as one instrument as wide as the plateau itself.

101 evaluations, and radio's own noise 7:09

Eighteen radio positions. Sixty five surviving scans from Hipparcos. One hundred and one separate evaluations inside the fit. And the radio positions carry a cosmic noise term of their own, added for the same reason as the optical one: the radio surface is lumpy too.

The proper motions disagree 7:30

When the fit is finished, one thing refuses to reconcile. The proper motion, meaning how fast the star drifts across the sky, comes out at eleven point three thousandths of a second of arc per year from Hipparcos, and nine point four from radio. Those are two point seven sigma apart.

The authors' conclusion is blunt, and it points at the optical data rather than the radio. Hipparcos, they write, still overestimates the parallax. Their combined answer is four point five one thousandths of a second of arc, plus or minus zero point eight zero.

What would actually settle it 8:10

The paper is explicit about what would actually settle this. You would have to measure the position and image the lumpy surface at the same time, at millimetre wavelengths, with ALMA and the array at Q band. Their own word for the scheduling effort that would take is Herculean.

Antares: the same three numbers 8:28

Before this sounds like one awkward star, look at the second object named in the same paper. Antares, the red supergiant in Scorpius. Angular diameter, about forty four thousandths of a second of arc. Distance, near two hundred parsecs. Parallax, about five.

The same three numbers, to the digit. And Antares needed three point six thousandths of a second of arc of cosmic noise, more than Betelgeuse did. This is not one difficult object. It is what happens to any star that is nearby and resolved at the same time.

The cost: absolute magnitude 9:08

Now the cost, and it is the quantity this channel is named after. Absolute magnitude is how bright an object would appear from a fixed distance of ten parsecs. It is apparent brightness with the illusion of distance removed. Removing that illusion requires knowing the distance.

Between the nearest and the farthest published value, the distance modulus moves by one point one five magnitudes. Converted to light, that is a factor of two point nine. The same star, the same measured brightness, and nearly three times the luminosity depending on which paper you open.

And everything downstream inherits it. The authors note that their own new parallax makes the star one point two times more luminous than their previous paper had it, and that was a revision of their own work, not somebody else's. Age and remaining lifetime move with it too.

620 or 1,050 solar radii 10:06

The radius does the same thing, out of the same angle. Forty four thousandths of a second of arc, at one hundred and thirty one parsecs, is six hundred and twenty solar radii. At two hundred and twenty two parsecs the identical measurement gives one thousand and fifty.

A distance obtained without an angle 10:25

There is one more value on that list, and it did not come from an angle at all. Joyce and colleagues modelled the star's pulsations, the four hundred day beat, against stellar structure and hydrodynamics, and arrived at one hundred and sixty eight parsecs.

Here is why that is worth having. Their model gives a radius of seven hundred and sixty four solar radii. Take the measured angle instead, place it at their distance, and you get seven hundred and ninety five. Two methods with no ingredient in common, agreeing inside the error bars.

4.51 +/- 0.80 milliarcseconds 11:02

So the episode ends where the instrument does. Four point five one thousandths of a second of arc, plus or minus zero point eight zero. Eighteen per cent uncertainty, on the bottom rung of the ladder that every rung above it is calibrated against.

The companion, and its 8.8 AU 11:19

One last thing, because it is the whole problem compressed into a single sentence. In December twenty twenty four, the SPHERE camera on the Very Large Telescope resolved a companion beside Betelgeuse, at six point one sigma. It now has a name approved by the union: Siwarha.

What the camera measured is a separation of fifty two point three two thousandths of a second of arc. What was published everywhere is eight point eight astronomical units. The second number is the first one multiplied by a distance that somebody had to choose. They chose one hundred and sixty eight parsecs.

Next: Siwarha 12:00

Choose two hundred and twenty two instead, and the same image gives eleven point six astronomical units. Thirty two per cent further out, with a different period and a different inferred mass. Next time: Siwarha, and what the camera recorded before the conversion that became the headline.

Description and sources

Astrometry measures the centre of light, not the centre of mass. On a supergiant whose surface boils in cells the size of Earth's orbit, those are not the same point, and they move. Both Hipparcos solutions only close after adding a term called cosmic noise -- 3.4 mas in 1997, 2.4 mas in 2007 -- whose origin, the paper says, remains unknown; that term is more than half the size of the parallax it helps to measure. Gaia cannot help: it saturates above about third magnitude. The 2017 solution sets 34 years of VLA, e-MERLIN and ALMA positions against Hipparcos's 28 months and lands on 4.51 +/- 0.80 mas. Antares, in the same paper, has the same 44 mas disc and needs even more of the same correction. The cost is the quantity this channel is named after: across the published distances the absolute magnitude moves 1.15 mag, a factor of 2.87 in luminosity, and the radius runs from 620 to 1,050 solar radii out of one angle.

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

Harper, G. M. et al. (2017). An Updated 2017 Astrometric Solution for Betelgeuse. AJ 154, 11, doi 10.3847/1538-3881/aa6ff9 -- source for every parallax here: 4.51 +/- 0.80 mas (222 +48/-34 pc) combined, 5.27 +/- 0.78 (190) from radio proper motions, 6.56 +/- 0.83 (152) from Hipparcos 2007, 131 (+35/-23) pc from the 1997 catalogue, 197 (+55/-35) from Harper et al. 2008; disc ~44 mas; cosmic noise 3.4 +/- 0.6 mas (1997) and 2.4 mas (2007), origin 'remains unknown'; 28 months of Hipparcos against 34 years of radio; mu_delta 11.30 +/- 0.65 against 9.37 +/- 0.28; Antares, same disc, 3.6 mas of cosmic noise; 'Betelgeuse is very bright, too bright for GAIA to observe'.
Joyce, M. et al. (2020). Standing on the Shoulders of Giants: New Mass and Distance Estimates for Betelgeuse. ApJ 902, 63 -- radius 764 (+116/-62) Rsun, distance 168 (+27/-15) pc, parallax 5.95 (+0.58/-0.85) mas, mass 16.5-19 Msun, from combined evolutionary, asteroseismic and hydrodynamic modelling with MESA. No angular measurement enters that distance.
Montarges, M. et al. (2026). VLT/SPHERE imaging of the candidate companion of Betelgeuse. A&A 711, L12; ESO release eso2611 (28 July 2026) -- companion detected December 2024 at 6.1 sigma, angular separation 52.32 mas, projected to about 8.8 AU using an adopted distance of 168 pc; estimated mass 2.6-3.1 Msun. The name Siwarha was approved for alpha Orionis B by the IAU Working Group on Star Names on 22 September 2025.
van Leeuwen, F. (2007). Hipparcos, the New Reduction of the Raw Data -- the 2007 solution and the intermediate data Harper et al. re-fit. ESA (1997), ESA SP-1200.
Gaia DR3 bright-end parallax bias, arXiv:2502.08068 -- Gaia's detectors saturate for sources brighter than G ~ 3.
Arithmetic shown on screen, recomputed: 44/4.51 = 9.8; 2.4/4.51 = 53%; the distance modulus moves 1.145 mag between 131 and 222 pc, a factor of 2.87 in light, matching (222/131)^2; (theta/2) x d gives 620, 795 and 1,050 Rsun; 52.32 mas x 168 pc = 8.79 AU.

ARCHIVE PLATES: all 15 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.

Corrections are pinned, not quietly edited.

#Betelgeuse #Parallax #Astronomy #Astrometry #AbsoluteMagnitude