Your First Milky Way Photo Without a Star Tracker

Almost every astrophotographer starts here, and a good number never leave. A camera on a tripod, a dark field, and the galaxy going up behind something worth having in the foreground.

No telescope. No mount. No guiding. The barrier is genuinely low, and the results on a first attempt are usually good enough to be dangerous to your savings.

What you need

A camera with manual mode. Any interchangeable-lens camera from the last fifteen years. Recent phones with a night mode will produce something, but you cannot control the individual exposures, which rules out most of what follows.

A fast wide lens. 14mm to 35mm, f/2.8 or faster. The kit zoom at f/3.5 works and will simply need more frames. Fast beats wide if you have to choose — every stop is a halving of the time you need.

A tripod. Any tripod. This is the one piece where cheap is fine.

Raw, not JPEG. Non-negotiable. The processing later depends on having the sensor’s actual values and JPEG has already thrown them away.

That is the entire kit. The next upgrade after this is a star tracker, and it is worth delaying — the untracked technique teaches you things a tracker hides.

When and where

Moon phase. New moon, or a night when the moon sets early. A gibbous moon washes out the core completely and there is no processing fix.

Season. The bright galactic core — the part everyone photographs — sits in Sagittarius, so its season depends on where you are.

From the northern hemisphere it runs roughly March to October, best around June and July, and the core stays low to the south. From the southern hemisphere the same object passes close to overhead, which is dramatically better: less atmosphere, more of it visible, and a season running roughly February to October. If you are imaging from Australia, southern Africa or South America, the Milky Way you can photograph is the one northern imagers travel for.

The other half of the year, in either hemisphere, you are looking outward through the sparse arm. Still the Milky Way, much fainter and subtler.

Darkness. This is the one place where light pollution really does gate the result. A wide-field broadband shot has no narrowband trick available to it. Bortle 4 is a workable minimum; Bortle 2 is transformative. An hour’s drive is often worth more than any equipment purchase.

Foreground. A tree, a ridgeline, a building, a lake. A photograph of only sky is an astronomy record; a photograph with something terrestrial in it is a picture. Scout in daylight.

The 500 rule is wrong

Here is what everyone repeats: divide 500 by your focal length to get the longest exposure before stars trail. 20mm lens, 25 seconds.

It was a reasonable rule for 35mm film and for the six-megapixel DSLRs it was coined on. On a 45-megapixel sensor it is far too generous, because the pixels are much smaller and a trail that used to fall within one pixel now smears across four.

Use the NPF rule instead — most planning apps have it built in and will calculate it from your exact camera and lens. It accounts for pixel pitch and aperture, and it typically returns something between a third and two-thirds of what the 500 rule claims. On a modern full-frame body at 20mm you are often looking at 8 to 12 seconds, not 25.

If you have no app, halve the 500 rule figure and you will be roughly right.

And then: take more frames rather than longer ones. Which brings us to the actual technique.

Stacking without a tracker

Twenty 10-second frames contain the same light as one 200-second frame, but the stars stay as points. Averaging the frames cuts noise by roughly the square root of their number, which for twenty frames is a factor of about 4.5 — the difference between a grainy mess and something you can print.

There is a complication that does not exist in deep-sky imaging, and it is the whole reason nightscapes are their own discipline.

The sky moves. The landscape does not.

Align on the stars and the trees smear. Align on the trees and the stars trail. There is no single alignment that satisfies both, because the two halves of your photograph genuinely moved differently between frame 1 and frame 20.

The standard answer is to stack the sky and the foreground separately and recombine them, which done by hand means masking a horizon across twenty frames and is exactly as tedious as it sounds. Akastroid’s nightscape mode does this as one step: it registers the sky on the stars, keeps the foreground fixed, and blends them along a boundary it works out from the frames themselves rather than asking you to draw one.

The honest caveat: a horizon of trees against a bright sky is a hard boundary to find, and a complicated one — branches, wires, a fence — is harder still. It will not always be perfect. But it is the difference between a technique you use and a technique you read about.

Settings to start from

For a 20mm f/2.8 lens on full frame, in a dark sky:

• Aperture: wide open, or one stop down if your lens has soft corners at maximum. Coma in the corners is the usual reason to stop down.
• Exposure: whatever NPF says. Probably 10–13 seconds.
• ISO: 3200. Higher than feels comfortable. You are going to stack, and noise is the thing stacking fixes.
• Focus: manually, on a bright star, using live view zoomed all the way in. Autofocus will fail. Infinity on the lens barrel is usually not quite right.
• White balance: anything — you are shooting raw, so it is a preview setting, not a decision.
• Frames: 20 to 30 of the sky. Then, without touching the tripod, 5 to 10 more for the foreground at a longer exposure if the landscape is very dark.

Turn off in-camera long-exposure noise reduction. It doubles your capture time and it does badly what stacking does well.

The most common first-attempt problems

Everything is slightly soft. Focus. Almost always focus. Do it on a star, zoomed in, and check it again after you have moved the tripod.

Stars are round in the middle, comet-shaped in the corners. Coma. Stop down a stop.

One side of the sky is orange. A town over the horizon. Compose away from it, or accept it and remove the gradient in processing — a wide sky gradient is one of the more tractable problems.

Stars trailed even at the calculated exposure. Your lens is longer than you think, or you are pointed at the celestial equator, where the sky moves fastest. Stars near Polaris tolerate several times the exposure of stars near the equator; the rules give the worst case.

The whole thing looks flat and grey. That is a raw file, and it is correct. Nightscapes take real stretching. This is the same problem — and the same fix — as why a nebula comes out washed out.

The short version

• A camera, a fast wide lens, a tripod and raw files. That is genuinely all.
• Ignore the 500 rule on any recent sensor. Use NPF, or halve the 500 figure.
• ISO higher than feels right, exposures shorter than feels right, more frames than feels right.
• Focus manually on a star, and check it again after every move.
• Sky and foreground move differently and need stacking separately.
• Drive somewhere dark. It outperforms any purchase at this stage.