Today I’m back with my first attempt at capturing star trails, from last week’s night-time visit to St Martha’s. Once again you really need to click on the image to see the larger version, as the rescaling done by the web server softens everything and makes the star trails look very faint. It’s also worth pointing out that even that larger version is itself scaled down from the full-size image I worked on, where the detail is obviously more visible. But realistically the only way to see that level of detail is on a print (until we start getting monitors with a notch more than 4k resolution, and enough money to actually buy one). I don’t usually state this – to me it’s rather obvious – but while it affects all images, in cases such as this the problem is even more significant.
Anyway, I digress. What you see is a composite of six exposures: one for the foreground, and a stack of five four-minute exposures for the star trails. This twenty-minute sequence was started about half an hour later than the previous image, and clouds started rolling in during this time. As you can see, the mood of this image is rather different from the previous one. The orange glow is from light pollution (there’s London about thirty miles in that general direction). This time I left the squiggly lines from low-flying aircraft. You can also see a long straight-ish horizontal line above them which is most likely a Perseid. I say ‘most likely’ because I know I saw a big one while these exposure were happening, and it was in that region. However, with all the aircraft flight paths there, I cannot be sure. As in the earlier image, too, you can see a few straight diagonal lines above the church from higher altitude flights.
At a technical level, the stack was achieved by linearly adding¹ the five frames, and compensating for the additional brightness with an exposure reduction of 2⅓ stops². Beyond that, processing was similar to the previous image, manually using a dark frame to identify hot pixel regions and cleaning these with a median filter, the usual selective to adjust contrast, and some extra sharpening in the sky to accentuate the star trails.
Click on the image for a larger version – really, please do, as the difference is substantial. As usual, comments welcome.
¹ In practice, the easiest way to achieve this is to add the layers using a linear dodge blend mode in 32-bit mode (which has a linear colour space and no clipping).
² If you care to know where that value comes from, remember that every time the exposure duration is doubled, this is equivalent to 1 stop. If there were four images, this would be the equivalent of 2 stops. Since there were five images, we can calculate the equivalent value in stops as log₂ 5 ≅ 2⅓.