Startrail ISO reference

[Nass]

I wanted to test an absolutely fantastic precise focus tool that I've made and I've come across various startrail & ISO questions so I constructed a comparison - may it help someone!

So one of the questions you always come across from people trying this for the first time is what settings to use - do you go superhigh ISO, use the widest aperture and what about the lens? All factors that come into play in deciding the number of stars you see.

Obviously you'll see less with 50mm than with 25mm because you cover less of the sky, but the real calculation to bear in mind is focal length / widest aperture. Because this decides your base sensitivity (but it is further decided by ISO). In this case I have used a 35mm f2.8 because for the purposes of this dem it gives me brighter stars than my 10-20mm f4. Also, with a 35mm I don't need to do continuous shots for 2 hrs like I do with 10-20 to get a decent sized arc because the Field of View is smaller.

So anyway, the shots at the bottom show various ISOs, commenting on each:

• 800ISO - theoretically you see the most stars BUT film grain (noise) increases, and the sky background also becomes brighter which drowns out the lightest stars
• 640ISO,500ISO - compromises
• 400ISO - sky brightness ceases to be a problem but the weakest stars don't get recorded

So I went with ISO500, which gave me the weakest stars but the darkest background. Now you'll notice that the RAWs I shot are pretty purpley compared to the final one.

This is deliberate. Urban light pollution tends to be orange, and I find that by using a custom colour temperature shoved far into the red/orange area of the colour wheel but controlling it back down to blue in the final image not only gives me more colour variation in the stars (they're meant to be different colours) than cranking the colour balance straight into 4000 degree territory in-camera but also minimises the ghastly orange glow from urban lighting. Bringing it back to blue is a trivial midtone colour balance adjustment towards cyan on the top and blue on the bottom.

Some other stuff worth talking about. This sequence put together was started 1 3/4 hrs post sunset with a half moon, facing SSE. 2 hrs would be better because there is a more consisted night light after 2 hrs. I had to face SSE because the moon in the west would mess stuff up totally; north gives you a circle in the sky and south gives you an ellipse arc (the top of a very large circle with that circle's centre behind the planet earth). To get the full ellipse you need to be more in 20-28mm territory but that comes with issues of its own because lenses tend to be slower there.

More star trail help here and if there's demand for it I'm quite happy to do a detailed tutorial, there are quite a few little things to remember if you're going to embark on this stuff but I'd encourage folk to try because you can do a lot of the neccessary experimentation just outside the back door!

If anyone has questions, feel free to pop them in this thread, I'll use this to create some sort of tutorial about this in the future online.

[The Blue Boy]

Good stuff.

I've a little question though. What about in-camera noise reduction time. I.E. does the time it takes the camera to perform this task make any difference to the final image? Or will it create gaps in the trails? Am I making sense?

[Nass]

Mark, you're making perfect sense.

In-camera noise reduction can be a problem, depending on the camera used and whether or not the option exists to switch it off. Bottom line, you need to switch it off because the time that it takes will introduce gaps when you come to finally add them together. If you don't have this option then you're better off considering one long exposure rather than blending together 30s exposures. If you do, then there are generally 3 ways of achieving consequent 30s shots:

• Intervalometer
• Timed shots
• Continuous shooting

I usually use the last because it has the smallest gaps between exposures. I use a manual 30s exposure and basically machine gun that until it's time to stop. My camera has the option to do this via IR remote; you'll need to understand your particular model to see the best way to achieve it. It might be through corded remote with a "keep going" type slider on the remote, or you may not have the option to do through remote and need to start the sequence off by pressing the shuttor on the body.

A little "gotcha" to bear in mind using an intervalometer, it is an intervalometer not a gapometer. In other words, if you set an interval of 1 second and the exposure is 5 seconds, this can make your equipment throw a wobbly because it wants to take the second shot 1 second after starting the first. But since the first is still exposing (we've set it for 5s), it can't. Your best bet is to experiment a bit beforehand.

[The Blue Boy]

Nass, cheers mate.

Yeah, I've got the option to turn it off. I've also got the intervalometer. I'll give it a go as soon as we get a clear sky. Only thing is we're on the Manchester/Salford border so light pollution is a real pain.

[Nass]

Rather than using the intervalometer, does your camera have a machine-gunning option, ie continuous shooting? Say you wanted to take 5 action shots in as short a time as possible pressing the trigger only once, how would you do it?

[The Blue Boy]

Yeah, I think so. I know I've got that option using the remote cord by locking the button. Will try that.

Edit; Just had a run through the manual and can do this using Continuous Low on the D300s.

[Nass]

Just to illustrate my point that gaps do exist with 30s intervals, this is 4 views from one image, 18mm focal length lens mounted. As you can see, at 400% and 100% the gaps are obvious, but by the time you get down to the sort of size you'd see at A3 or A4 they're not visible to the naked eye:

[Nass]

I also wanted to post this, which is the shot those 4 gap comparisons are made from, it's just a technique thing to make startrails more interesting. So I find that if I get to the location just around sunset and take some shots at that time, the ability to blend that lit foreground into a startrail starting 2 hrs after sunset generally makes the finished pic more interesting:

[Nass]

So a little bit about the shape that startrails make in the sky for this thread update.

Actually... stars don't move much. It's our planet that rotates at a bit of an angle, which produces the trails that you capture. Now, being at that bit of a tilt means that the shape in the sky will differ according to direction. What's worse, in the southern hemisphere it's the other way around. And it's different at the poles too.

So for where we are, Britain:





These are the approximate paths that you'll see your startrails make depending on the direction you shoot in, at out lattitude.

• Go higher (ie north pole), and instead of the angle to see the centrepoint north being about 55deg or so up, it's more like 85.. ie directly overhead.
• On the equator, the two centrepoints are on the horizon
• In the southern hemisphere the centrepoint is seen by looking south, not north

[Nass]

Just putting this up for reference, the first diagram I did yesterday to illustrate this concept. But I thought I could do nicer so made the above ones.

[Dave Humphries]

Hi Johan,

This is good stuff, I've never seen diagrams like that, makes it far more understandable.

Thanks,

[Peter Ryan]

Thanks Johan - very good work and thanks for sharing.

[RockNGoalStar]

This is turning into a very useful thread, thanks for putting it together.

I have only really had one go at doing star-trails and I used startrailes.exe to layer my 30 sec exposures. I wasn't particularly impressed with the results, which had noticable gaps between the stars. I used an intervalometer and I think the smallest gap between exposures is 1 second. I'd rather it was 0sec.

I want to have another go at this though, the only thing is that it is always cloudy when I go back to Wales for the weekend!

[Nass]

I wonder if anyone from this old thread is planning a shot of comet ISON in the weeks ahead? Not the same as a startrail obviously, but nevertheless now might be the time to prepare and practise before trying for the real thing in a few weeks!