I can't really explain it so I'm finding it hard to Google for tutorials. I was wondering how to get this effect:

Instagram Video

By using something like Photoshop CS6 or a freeware program. I'm familiar with using the timeline to create normal timelapses, just not ones where the previous images linger on. Am I right in guessing a lighten blending mode is used?

  • 1
    \$\begingroup\$ Would this be better served on video.stackexchange.com? \$\endgroup\$
    – Calyth
    May 3, 2018 at 17:50
  • \$\begingroup\$ @Calyth Aha I never knew that was a thing! I'll ask there :) \$\endgroup\$
    – adam
    May 3, 2018 at 21:03

1 Answer 1


This is more of a photography question than a video question. If you attempt this with a video camera you will need to blend, if you use a still camera you will have much greater resolution and ramping options.

There are many ways to do it, here's the best and shortest video I could find with a half hours effort: "Turn Your Star Trails Photos into a Video".

On the right are numerous links to other methods. It's technical and lengthy to explain each method. Some general tips are available here: "Getting Started With Star Trail Photography", "Instructables - Star Trails - A Beginner's Guide" and "StarCircleAcademy - Star Trails".

A short explanation is to use an intervalometer, increase your bulb time for each shot in aperture priority mode. That will give you 'moving dots' that stretch into long star trails if you stack them in post - exactly like in your example. Just take all your photos and make them into a video or an animated .GIF file.

If you are in the northern hemisphere then pointing your camera towards (<1° off) the North Star (Polaris is a binary star, α UMi Aa is at a declination of +89° 15′ 50.8″) produces circular trails> In the southern hemisphere you would point towards Sigma Octantis. Pointing towards the equator provides long streaks instead of circular trails. A longer exposure will produce longer star trails, but will also usually dim the brightness of the trails.

The actual location of the celestial poles varies due to axial precession. Because of a phenomenon known as the precession of the equinoxes, the poles trace out circles on the celestial sphere, with a period of about 25,700 years. The Earth's axis is also subject to other complex motions which cause the celestial poles to shift slightly over cycles of varying lengths; see nutation, polar motion and axial tilt. Finally, over very long periods the positions of the stars themselves change, because of the stars' proper motions. The apparent positions of the stars also change slightly because of stellar parallax effects.

Axial Precession

When pointing at a celestial pole the stars complete one full rotation in less than 24 hours, and move almost 15 degrees every hour. In 24 Earth hours they travel 361°. Some easy math is available here: "Greg Boratyn - Night Photography".

If you want to shoot for more than 8 hours (sunset to sunrise) and for multiple days you might want to read: "Difference between sidereal day and solar day on Earth".

"Earth moves a little less than a degree around the Sun during the time it takes for 1 full axial rotation. So, for the Sun to appear on the same meridian in the sky again after 1 full axial rotation, the Earth has to rotate one extra degree to bring the Sun into the same apparent meridian in the sky. This is why the solar day is longer than the sidereal day by about 4 minutes.".

Solar vs. Sidereal Day

For a lengthy, almost easy to understand, explanation of nearly everything check out Guy Cook's webpage: "What we see in the Sky: Stars". To find magnetic north for your location try the "NGDC Mobile Declination Calculator", it uses the World Magnetic Model to adjust for anomalies in the Earth's magnetic field.

If you don't point towards the celestial pole but instead point towards Polaris you'll end up with a photo (from "Stellar Neophyte Astronomy Blog") like this:

Offset Error

The bright line is Polaris, the arrow points to the celestial pole.

Back to the simple explanation: Start a short bulb exposure, maybe 30 - 60 seconds, then ramp it up to an exposure time of 5 minutes and keep doubling it until you get a good result. Once you get close to the right exposure time, use your judgement and previous results to decide whether to increase or decrease the exposure time.

Some helpful software to cheat and blend what you have into what you want:

Use the "Exposure Triangle" to adjust between brightness, streak length and 'trail decay'.

Exposure Triangle

Need more info just ask, I'm willing to make a longer answer when I have more time available. Check out some of the videos and find an example of what you want.

Here's a video of one of the better effects - stars move slowly and then streak, sometimes they unstreak other times the scene fades. Looking at the background gives tips on the interval of the blends. Well done work translates into any language, he speaks English too, but the description text is in Chinese.

  • 1
    \$\begingroup\$ using this method, the starting point of the trails should move, but it doesn't in the video. \$\endgroup\$
    – ths
    May 4, 2018 at 9:35
  • \$\begingroup\$ Great answer - I'll try the Starstax 'comet-mode' when I get home, and when I next go out shooting I'll try increasing the exposure time with each shot. Thanks! \$\endgroup\$
    – adam
    May 4, 2018 at 9:57
  • \$\begingroup\$ Minor nit. The stars do NOT complete one full rotation in 24 hours. They complete one full rotation in one sidereal day, which is approximately 23 hours, 56 minutes, 4.0905 SI seconds. Thus the stars move ever so slightly more than 15° per hour. A year ( the length of time it takes the Earth to go one revolution around the sun) is ≈365.2422 solar days which equates to ≈366.2422 sidereal days. For every 365 times the sun rises, the stars on the celestial equator rise 366 times. \$\endgroup\$
    – Michael C
    May 6, 2018 at 2:06
  • \$\begingroup\$ P.S. The Moon only rises about 353 times per year, as it loses one "lap" to the sun every 29.5 days or so. \$\endgroup\$
    – Michael C
    May 6, 2018 at 2:14
  • \$\begingroup\$ @Rob thus the '≈' in front of most of the numbers in the comment. They're still several orders of magnitude more accurate than your original answer, and only very slightly less accurate than your updated answer. \$\endgroup\$
    – Michael C
    May 6, 2018 at 5:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.