Elongated Stars: Is It Your Tracking, or Your Optics?

Elongated stars are the most common complaint in astrophotography and the most commonly misdiagnosed. The reason is that two completely different faults produce a symptom that looks identical in a single frame — and the internet’s default answer is always “improve your guiding”, which is right about half the time and expensive when it is wrong.

There is a clean test. It takes one look at three frames.

The test

Look at the same corner of three frames taken an hour apart.

• If the stretch changes — different length, different angle, present in one and gone in another — the mount moved.
• If the stretch is identical in all three — same direction, same amount — it is the optics.

That is the whole test, and the logic behind it is simply this: your optics did not move between frames. Whatever they do to a star, they do to it in every frame of the night, the same way. A mount, by definition, is the thing that moves.

Why this distinction matters so much

Because the two fixes have almost nothing in common.

Tracking faults are addressed with polar alignment, guiding, cable management, balance, and not asking a mount to carry more than it can. Some of that is technique and free; some of it is a guide scope and a camera.

Optical faults are addressed with a coma corrector, a field flattener, backfocus spacing measured to the millimetre, tilt adjustment, collimation, and sometimes a sturdier focuser. None of that is guiding equipment, and a guide camera will not improve a single one of them.

Somebody who buys an autoguider to fix a tilted sensor has spent real money and will see no improvement whatsoever. That happens constantly.

Reading it off a single frame, if you must

You can often tell from one frame, if you look in the right place.

Optical faults are usually field-dependent. Coma, field curvature and tilt all get worse away from the optical axis. So compare the centre to the corners:

• Centre sharp, corners stretched radially (pointing outwards from the middle) — coma, or a missing corrector.
• Centre sharp, corners soft but round — field curvature. Everything is at a slightly different focus depending on where it sits.
• One corner or one side worse than the rest, asymmetrically — tilt. The sensor is not perpendicular to the light path, or the focuser is sagging under the camera’s weight.
• Everything soft, uniformly, corners no worse than the middle — that is not an optical aberration, that is focus or seeing.

Tracking faults are usually field-independent. The whole frame drifts together, so stars in the middle are stretched about as much as stars at the edge, and they are all stretched in the same direction — the direction the mount moved.

That last point is the single most useful thing to look for. Radial elongation, fanning outwards, is optics. Parallel elongation, all the trails pointing the same way, is motion.

Field rotation, the third answer

There is a case that is neither, and it catches people with alt-azimuth mounts and wedge-less setups: field rotation.

Stars near the centre are round, stars at the edge are stretched into short arcs, and the arcs curve around the centre of the frame. That is not tracking error and it is not an aberration — it is the field turning during the exposure because the mount is following the sky in altitude and azimuth rather than about a polar axis.

The fix is shorter subs and more of them, or a wedge.

Why software should not reject frames for this

Here is where automatic frame rejection goes badly wrong, and it is worth understanding if you use any stacking software.

Most quality scoring includes star shape, and reasonably so — a trailed frame is a worse frame. But if the elongation is optical, it is in every frame equally. A rule that rejects frames for elongation then either rejects all of them, leaving you with nothing, or rejects an arbitrary fraction of them, leaving you with a thinner stack and no explanation.

Neither is right, and the correct behaviour is to notice that the fault is shared:

Akastroid measures elongation across the whole session, not frame by frame. When the same stretch is present in all of them, it reports the eccentricity, says explicitly that this is the optics rather than the tracking, and rejects nothing for it — then goes ahead and stacks the session, because slightly oval stars in every frame is a picture and an empty session is not.

That is the difference between a measurement and a judgement. The measurement is “these stars are 0.97 eccentric”. The judgement is “and since that is true of all 200 frames, it is not a reason to throw any of them away.”

What to do tonight

1. Take three frames an hour apart. Compare the same corner.
2. Changing → mount. Start with polar alignment, then cables and balance, then guiding.
3. Identical → optics. Check spacing first, because it costs nothing; then tilt; then whether you need a corrector for this scope.
4. Curved arcs around the frame centre → field rotation. Shorter subs.
5. Round in the middle, radial at the edges → you need a coma corrector or a flattener, and no amount of guiding will substitute.

And measure rather than squint. Image Doctor reports eccentricity per frame and across the session, which is what makes the shared-versus-varying distinction visible without you flicking between frames at two in the morning.

Related: why are my stars trailing, a symptom-by-symptom guide to bad astrophotos, reading a quality score.