Help with Milky Way photography

Messages
731
Name
chris
Edit My Images
Yes
Hello
I'd like to have a go at taking photos of the Milky Way but don't know where to look for it in the sky.

Iv got sky walk on my phone but it only seems to show the star consolations.

Any help would be appreciated
Thank you chris
 
Download a program called stellerium it shows all planetry objectslook for the constellation cassiopeia and you will see a light milky patch that is the milky way and you can follow it through the stars, It's not very promient at the moment you can see it better in summmer in a dark sky
 
Here's a screen grab of Stellarium showing the Milky Way

HdjTDXw.jpg


It runs from Cygnus through Cassiopeia

This screen shot is accurate as of the time of this post. It sets soon after. Early Autumn is best for Milky Way shots, but as you can see above, part of it is still visible even in Winter. To see it at it's best you really need to be in the southern hemisphere... but that's hardly practical :)

At this time of year, I recommend shooting NW at around 6-7pm in a sky free of light pollution. In a true, dark sky sight (or close to it) use f2.8 ISO6400 for 25 seconds with a lens around 24mm ish. If you have lots of light pollution you will need to either stop down, use a shorter speed, or lower the ISO. You'll need to experiment around those times.

Any longer than 25 seconds, or lens longer than 24mm and you'll start t get star trails.... or use the "500" rule. Divide 500 by your focal length to give max time in seconds. e.g, 20mm lens = 500/20 = 25 seconds.
 
Last edited:
I've never tried this type of photography, but I'm tempted. As David said, it's a lot more visible in the Southern Hemisphere. It's very conspicuous once you get beyond the light pollution of the cities, and I've often seen it as a great 'milky' swathe across the night sky from the Drakensberg, the Karoo and various game reserves. The stars are spectacular too, and seem so much closer.
 
Not to confuse things david, can i ask where you get the 500 rule from because from all the searching on the net about astrophotography I've only seen them mention divide by 600 or 400, i been using the 600 one next time i go out i will try 500 to see what it's like, like you say you need to experiment with these things
 
You've written your own answer. It's an average of the two contesting theories. I've tested "500" extensively and while it may not be scientific in it's accuracy, it MORE then suffices. Trust it... it will work.
 
Funny this should crop up - I am going out tonight and taking my chances on the Northern Lights appearing and it will be my first attempt at such things so have been Googling some knowledge all morning and have also come to the conclusion that 500 is the best starting point and im pretty excited to try it out tonight. Fingers crossed the Aurora is visible from my chosen location.
 
Last edited:
I've always been able to see it anywhere that's relatively dark as long as it's clear, although that could just be a benefit of living somewhere where there's a fair bit of dark night sky. You do need to allow your eyes some time to adjust, though.

I had my first attempt at photographing it over the summer - been waiting for a clear night to try it again recently. These were done at 30mm (on NEX 1.5x crop sensor) for around 30 seconds at ISO1600 IIRC.

533992_10151881298915695_797473299_n.jpg


537212_10151881299535695_1416490670_n.jpg
 
Last edited:
I've always been able to see it anywhere that's relatively dark as long as it's clear,

The Milky Way is visible all year round, but it's far more dramatic in summer/autumn... and the best part of it is unfortunately not visible from the northern hemisphere.

I made all of these images in the UK during winter/early spring though.. if I recall correctly... certainly not summer anyway.

SJMWYh1.jpg

VJzPo6f.jpg

81hMuLI.jpg
 
Last edited:
Great images David, especially like the first and the last one is very dramatic

Steve

The Milky Way is visible all year round, but it's far more dramatic in summer/autumn... and the best part of it is unfortunately not visible from the northern hemisphere.

I made all of these images in the UK during winter/early spring though.. if I recall correctly... certainly not summer anyway.

SJMWYh1.jpg

VJzPo6f.jpg

81hMuLI.jpg
 
Great phot Drew!

Just for the clarity of how to stop stars trailing etc. If you aren't sure, there is a website here (not mine!)
http://www.howardedin.com/articles/fov.html
which will give you the sampling in arcseconds of the pixels in most DSLRs, as a function of the focal length.
All you then need to remember is that stars near the equator seem to move across the sky at 15 arcseconds
per second of time (stars at the poles don't move at all, the ones in between move at an intermediate speed).
So in 25 seconds you'd need to have pixels not much bigger than 375 arcseconds on a side for stuff down in the
south above the equator to stop it appearing extended. Change the focal length by a factor of two, change the
sampling in arcseconds by the same (longer focal length, finer sampling, fewer arcseconds per pixel). Once you
have some idea of the magic numbers for your camera it's pretty straightforward.
 
thanks - some others, all one shot

Photographer-Drew-Buckley-shines-a-torch-into-as-the-Milky-Way-fills-the-southern-skies-over-the-Pembrokeshire-Coast-August-13th-2013.jpg


Milky-Way-streaks-across-the-southern-Pembrokeshire-sky-as-Elegug-Stacks-are-illuminated-by-a-hand-torch-August-13th-2013.jpg


StGovansChapel_MilkyWay.jpg


Milky_Way_Church_Rock_Broad_Haven.jpg


launching my astro photography workshops in march :)

cheers,
drew
 
Wow those are amazing Drew, I love the 2nd one and the last one, but they're all brilliant, thanks for sharing. Where are you going to run your courses?
 
Great phot Drew!

Just for the clarity of how to stop stars trailing etc. If you aren't sure, there is a website here (not mine!)
http://www.howardedin.com/articles/fov.html
which will give you the sampling in arcseconds of the pixels in most DSLRs, as a function of the focal length.
All you then need to remember is that stars near the equator seem to move across the sky at 15 arcseconds
per second of time (stars at the poles don't move at all, the ones in between move at an intermediate speed).
So in 25 seconds you'd need to have pixels not much bigger than 375 arcseconds on a side for stuff down in the
south above the equator to stop it appearing extended. Change the focal length by a factor of two, change the
sampling in arcseconds by the same (longer focal length, finer sampling, fewer arcseconds per pixel). Once you
have some idea of the magic numbers for your camera it's pretty straightforward.

So, I have a D7000 and say i'm shooting at 18mm, what does this mean to me? What exactly shall I do with the figures created here? I'm, at best, perplexed, but intrigued. :)
 
So, I have a D7000 and say i'm shooting at 18mm, what does this mean to me? What exactly shall I do with the figures created here? I'm, at best, perplexed, but intrigued. :)
OK, you enter your camera type and focal length in the first set of boxes and press 'Calculate'. This will tell you that the field of view measures 66.499° by 46.861°, but the key number is the arcseconds per pixel which is 54.775.

Now the advice about star trailing is:
All you then need to remember is that stars near the equator seem to move across the sky at 15 arcseconds
per second of time (stars at the poles don't move at all, the ones in between move at an intermediate speed).

So if your view includes any stars on the celestial equator, you'll start to see trailing (on a 100% crop of your image) if your shutter speed exceeds 54.775/15 which is roughly 4 seconds. With an 8-second exposure, stars on the equator will occupy 2 pixels; in 15 seconds they'll occupy 4 pixels; and so on.

But in the real world you're doing pretty well if your stars are single-pixel points to start with, so you've probably got a bit of latitude with these figures. Another real world consideration is whether you really want/need a 16-megapixel image. You only need about 6 MP for a good print. So if you plan to sample down from 16 MP to 6 MP, you can multiply all the exposure times by sqrt(16/6) which is 1.6.

Does that help?
 
OK, you enter your camera type and focal length in the first set of boxes and press 'Calculate'. This will tell you that the field of view measures 66.499° by 46.861°, but the key number is the arcseconds per pixel which is 54.775.

Now the advice about star trailing is:

So if your view includes any stars on the celestial equator, you'll start to see trailing (on a 100% crop of your image) if your shutter speed exceeds 54.775/15 which is roughly 4 seconds. With an 8-second exposure, stars on the equator will occupy 2 pixels; in 15 seconds they'll occupy 4 pixels; and so on.

But in the real world you're doing pretty well if your stars are single-pixel points to start with, so you've probably got a bit of latitude with these figures. Another real world consideration is whether you really want/need a 16-megapixel image. You only need about 6 MP for a good print. So if you plan to sample down from 16 MP to 6 MP, you can multiply all the exposure times by sqrt(16/6) which is 1.6.

Does that help?

Thanks very much, it does actually! I'll try putting it into practice and see what happens. Although 4 seconds doesn't seem so long! Do you need to consider crop factor when inputting the focal length on a crop sensor camera though?
 
OK, you enter your camera type and focal length in the first set of boxes and press 'Calculate'. This will tell you that the field of view measures 66.499° by 46.861°, but the key number is the arcseconds per pixel which is 54.775.

Now the advice about star trailing is:

So if your view includes any stars on the celestial equator, you'll start to see trailing (on a 100% crop of your image) if your shutter speed exceeds 54.775/15 which is roughly 4 seconds. With an 8-second exposure, stars on the equator will occupy 2 pixels; in 15 seconds they'll occupy 4 pixels; and so on.

Priority one with images like this is capturing as much light as possible to keep noise to a manageable minimum. If you start getting obsessed with stars being absolute single pixels (impossible actually) then you'll find yourself shaving more and more exposure time away in this quest, especially if you have a wide view with a 14mm lens or something, as there's always going to be stars near the horizon. You'll negate all that obsessive star sharpness in your noise floor anyway. I mean, look at Drew's images above... they're at 30 seconds. Going in close you will see some slight trailing on the stars near the horizon... but does it matter? As Milky way images go, they're good. You don't see him calculating how many arc seconds the stars move :) If trying to get the stars pin sharp compromises your ability to capture the much dimmer galactic cloud detail then you're wasting your time, surely. Do a rough "500" rule calculation and go for it. Who cares how any pixels the stars occupy :)



But in the real world you're doing pretty well if your stars are single-pixel points to start with, so you've probably got a bit of latitude with these figures.

I'm quoting this just to make it clear I wasn't having a go at Stewart here... he's right... don't get obsessed... there's a LOT of latitude. Brighter magnitude stars will flare out over around 5 pixels anyway... sometimes more. Look at Vega in that first image I posted... that's not post processing.... it just flared like that... it's a bright star. It looks like a golf ball!.... I think it looks cool though.



Another real world consideration is whether you really want/need a 16-megapixel image. You only need about 6 MP for a good print.

Depends how big you want to print, but yeah, reducing the res if you don't need anything massive, and specially if it's intended for screen will help a great deal.
 
Last edited:
  • Like
Reactions: Nod
I can never get mine to look 'cosmic-y' like that with the cloudy parts really showing up. I'm waiting for a clear moonless night to try again this year though.

You need a really true dark sky site for that. Drew, as he said, was on the Pembs coast.... it's really dark down there :)
 
You need a really true dark sky site for that. Drew, as he said, was on the Pembs coast.... it's really dark down there :)
I've got friends in Swansea. Maybe I'll have to see if they're up for being a stop over point for a trip out there.
:)
 
I might set a thread up for a night meet at the elan valleys once it gets a little warmer. Can do light painting, star trails and aim for Milky Way shots too if anyone's interested?
 
The very end of the North Wales Penninsula is great too... around Abadaron. That's a true dark sky site away from the villages. Galloway forest too, for those north of the border. Get away from towns like Newton Stewart and head into the forested areas and it's astonishingly dark. Most of the Lake District isn't dark enough IMO. Those shots I posted up were in the Lakes. Those cloudy parts of the galaxy are very, very faint, and if the light pollution is brighter than them, you'll never get them no matter how much processing you try to do on the images.

The truth is.... there aren't many places in the UK left that are true dark sky sites :(


I might set a thread up for a night meet at the elan valleys once it gets a little warmer. Can do light painting, star trails and aim for Milky Way shots too if anyone's interested?


Bit tricky doing Milky way shots if other photographers nearby are blasting coloured light/flash around because they're light painting! :)
 
Last edited:
The truth is.... there aren't many places in the UK left that are true dark sky sites :(





Bit tricky doing Milky way shots if other photographers nearby are blasting coloured light/flash around because they're light painting! :)

The elan valleys are in the process of applying for dark skies status and it looks like they're going to get it.

I'm sure we could do light painting early in the evening and star trails afterwards once its darker.
 
This is probably a bit nooby but I refer to the 500 rule above, about dividing 500 by the focal length to avoid start trails. Is this the ACTUAL focal length of the lens or the effective focal length? What I mean is, if I use a 50mm Nikon FX lens on a DX camera the effective focal length is 75mm, so do I divide 500 by 50, 75mm? My guess is 75, yes?
 
This is probably a bit nooby but I refer to the 500 rule above, about dividing 500 by the focal length to avoid start trails. Is this the ACTUAL focal length of the lens or the effective focal length? What I mean is, if I use a 50mm Nikon FX lens on a DX camera the effective focal length is 75mm, so do I divide 500 by 50, 75mm? My guess is 75, yes?

Yeah you use the effective focal length/35mm focal length
 
Back
Top