An accompanying image to my previous Dooralong startrails, this time it's a single exposure of 15 minutes. Do you want to do a startrails shot and are wondering if you should you do one long exposure or multiple short exposures and stack them?
Short exposures (~30 seconds) and stack them!
This image illustrates one of the reasons why - NOISE. Taken with the 40D, there are red and blue hot pixels covering this image as the noise builds up with the heat of the sensor over 15 minutes.
With multiple short exposures, the heat doesn't seem to build up as much. You can also use 'dark frames' to reduce noise, but I didn't with this shot. It would've meant an extra 15-minute exposure with the lens cap on, or turn on in-camera-noise-reduction.
Want more reasons?
- Individual frames can be used for nightscape images on their own, in addition to the resulting startrails image.
- The images used to create the startrails can also be used to create a timelapse.
- Planes or lights or dew won't ruin your whole image, just a few frames which can be excluded from the final stack and still get a usable image.
What are your thoughts? Do you have additional reasons or counters for whether to do short exposures or single long exposure for star trails?
You can see a less-compressed version of this image here, warts (noise) and all: http://goo.gl/fFnCS
Another aspect of choosing between short vs long subs for star trails is to balance the relative brightness of the stars versus background sky glow and airglow.
Shorter exposures, versus longer exposures with the same equivalent Exposure Value, will give you brighter / higher contrast star trails because the background stays constant but the stars won't smear out over long distances and become dim.
Interestingly enough, reciprocity failure back in the film days would have helped improve the star trails' contrast over the background.
I must admit that I don't like the slight disjointed effect of the short exposures. I think this is the one type of imaging that film still gives a more pleasing image.
I think that using a dark would be mandatory if you use long exposures with a digital camera. You would always do so with deep sky images. Why not with a star trail image?
I have to agree with Terry - I think this is probably the one astrophotography area that film still exceeds what can be achieved with digital. The smooth arcs that film gives are so much more pleasing and focus was always a snap
I always did an hour's expoisure without a drama with the old Minolta - sadly these days the lens has mould growing in it.
I must admit that I don't like the slight disjointed effect of the short exposures. I think this is the one type of imaging that film still gives a more pleasing image.
I think that using a dark would be mandatory if you use long exposures with a digital camera. You would always do so with deep sky images. Why not with a star trail image?
Yeh Terry it can certainly appear like there's still gaps with short exposures.
The software is getting better, and now there's "gap fill" algorithms to try and fill in gaps, but so far the results don't appear to be as good as they claim on the box. It's still noticeable.
The benefit of being able to use the single frames, or create a timelapse, can sometimes outweigh the downside.. also with a long exposure you're sometimes guessing on the right exposure length/f-stop/ISO and sitting around for 15-30 minutes hoping you've got it right is a test in patience
Doing a dark frame is definitely 'on the list', but I just forgot/didn't do it that night. This 15min image was the last one for the night before I packed up and went home, so didn't want to wait around for it.
I should've set it off and left it going in the car.
Another aspect of choosing between short vs long subs for star trails is to balance the relative brightness of the stars versus background sky glow and airglow.
Shorter exposures, versus longer exposures with the same equivalent Exposure Value, will give you brighter / higher contrast star trails because the background stays constant but the stars won't smear out over long distances and become dim.
Interestingly enough, reciprocity failure back in the film days would have helped improve the star trails' contrast over the background.
Thanks Dave.
Good point too - I guess light pollution might be more of a factor in long exposures?
To circumvent the gaps, I think it is quite easy to find where the pole is using simple Photoshop tools, then, set a centre point at the pole and rotate your stack of images until the arcs overlap any gaps between exposures. That should fill them in (using masks to hide the foreground, if there is one).
It's a little bit of messing about, but, should get rid of the jitter.
Nice image Mike. Each has it's own application. And I use both long and loooooooooong exposures for startrails equally, depending on the conditions and what my goals are from the imaging session
Quote:
Originally Posted by iceman
....I should've set it off and left it going in the car.
That's what I do. Stick the camera in the boot where the temp is closer to ambient, and let it click away.
Quote:
Originally Posted by iceman
.. also with a long exposure you're sometimes guessing on the right exposure length/f-stop/ISO and sitting around for 15-30 minutes hoping you've got it right is a test in patience
Being completely self taught- trying to determine the right exposure at night used to really confuse me and more often than not I would just guess (and more often than not get it completely wrong) until I read a great little ebook called "Seeing The Unseen" by Alister Benn. It was a lightbulb moment. Now I take a high ISO, 30 second test shot and just work it out from there based on the histogram of the test shot. As an example: Test done at ISO 6400, 30 Seconds and F/2.8. Histogram shows that I am underexposed by one stop. I want to shoot at F/5.6 to get more depth of field. F/2.8 to F/5.6 is two stops (running total 3). And I want to shoot at ISO 400 instead of 6400, which is 4 more stops for a total of 7. 30 seconds times 2, times 2, times 2, times 2, times 2, times 2, times 2 = 3840 seconds or 64 minutes to get the right exposure (I hope my math is right- doing this in my head)
Quote:
Originally Posted by Octane
To circumvent the gaps, I think it is quite easy to find where the pole is using simple Photoshop tools, then, set a centre point at the pole and rotate your stack of images until the arcs overlap any gaps between exposures. That should fill them in (using masks to hide the foreground, if there is one).
It's a little bit of messing about, but, should get rid of the jitter.
H
This works well if you have a not too wide of a field of view and have the celestial pole in frame or not too far out of frame. But if you are going really wide to encompass both the pole and equator, or you are shooting east or west with a wide angle and the stars are curving in both directions, then this method does not work (or it would work with an awful lot more messing about and a lot more masks).
Nice image Mike. Each has it's own application. And I use both long and loooooooooong exposures for startrails equally, depending on the conditions and what my goals are from the imaging session
Quite right! The best approach.
Quote:
Originally Posted by obsidianphotos
That's what I do. Stick the camera in the boot where the temp is closer to ambient, and let it click away.
I used to do that for deep space stuff years ago but I've just been forgetting to take darks lately.
Quote:
Originally Posted by obsidianphotos
Being completely self taught- trying to determine the right exposure at night used to really confuse me and more often than not I would just guess (and more often than not get it completely wrong) until I read a great little ebook called "Seeing The Unseen" by Alister Benn. It was a lightbulb moment. Now I take a high ISO, 30 second test shot and just work it out from there based on the histogram of the test shot. As an example: Test done at ISO 6400, 30 Seconds and F/2.8. Histogram shows that I am underexposed by one stop. I want to shoot at F/5.6 to get more depth of field. F/2.8 to F/5.6 is two stops (running total 3). And I want to shoot at ISO 400 instead of 6400, which is 4 more stops for a total of 7. 30 seconds times 2, times 2, times 2, times 2, times 2, times 2, times 2 = 3840 seconds or 64 minutes to get the right exposure (I hope my math is right- doing this in my head)
Ah thanks - good point. There's a method to the madness. I've seen Alister's book but not read it - will have to have a look.
Haha, this is /exactly/ what I do! Especially when I'm freezing my knackers off in NZ.
H
Quote:
Originally Posted by obsidianphotos
Now I take a high ISO, 30 second test shot and just work it out from there based on the histogram of the test shot. As an example: Test done at ISO 6400, 30 Seconds and F/2.8. Histogram shows that I am underexposed by one stop. I want to shoot at F/5.6 to get more depth of field. F/2.8 to F/5.6 is two stops (running total 3). And I want to shoot at ISO 400 instead of 6400, which is 4 more stops for a total of 7. 30 seconds times 2, times 2, times 2, times 2, times 2, times 2, times 2 = 3840 seconds or 64 minutes to get the right exposure (I hope my math is right- doing this in my head)
This works well if you have a not too wide of a field of view and have the celestial pole in frame or not too far out of frame. But if you are going really wide to encompass both the pole and equator, or you are shooting east or west with a wide angle and the stars are curving in both directions, then this method does not work (or it would work with an awful lot more messing about and a lot more masks).
Top show and Dig the compo Mike, dead tree allows the trails through not detaching from the theme.
Ive yet to use Starstax or similar software, but have seen plenty of inspiring results here. Apart from the technicalities of proceedure/processing ect, them mulitple short exposed stacks do look very surreal and powerful, if thats what one is wanting to be achieved ? Obviousely the long single result has a more realistic appearence with trails of only a few degrees !
So I think both have there place !
The bare tree in your image holds its place very well here, so composition is also a factor, foreground subject etc, if this same field was done with multiple shorts, I think the trails would smash that tree out of existance ? ? So be it, if thats whats one is after
Another pro of using shorter exposures is that if you have rapidly changing lighting conditions (moon set/rise, clouds, fog, airglow, etc) you can choose which parts of the timeline to stack in post processing - instead of risk losing the entire shot.
I do agree that the longer exposures look much smoother, plus it saves the wear and tear on the shutter. Perhaps medium length exposures - 3 to 5 mins - might be a good compromise?
I do agree that the longer exposures look much smoother, plus it saves the wear and tear on the shutter. Perhaps medium length exposures - 3 to 5 mins - might be a good compromise?
But then you can't use them for single shots or timelapses.
I have to agree with Terry - I think this is probably the one astrophotography area that film still exceeds what can be achieved with digital. The smooth arcs that film gives are so much more pleasing and focus was always a snap
I always did an hour's expoisure without a drama with the old Minolta - sadly these days the lens has mould growing in it.
I don't like the jaggie versions with short exposures either.
For long film like exposures with smooth trails I use a cooled CCD camera like the QHY8 fitted with a wide angle nikon lens via an adapter.
That way you avoid the noise of the long DSLR exposures and get the nice smooth trails.
Like Mike, I would recommend two camera setups, one DSLR taking short exposures and one low noise cooled CCD camera taking long ones.
I'm a big fan of longer exposures. Stacking the shorter exposures tends to make for very white stars that in a long sequence can start to blow out the whole image.. and it becomes very difficult to see the bright stars/constellations as they get drowned out by faint ones.
Hot pixels are a bit of a pain but they subtract out very well with dark frames as they are true 'hot pixels' rather than random noise.
My base setting that I remember is 15 minutes, f4, ISO200. Easy to scale up or down from there in the same way that Greg describes. Haven't been doing much (any) photography since the eclipse, but finally have some new shots from the last week at Camp Cooinda on the Gippsland Lakes, including some star trails. Hope to post soon..
I've only done a few star trails. I did one with I think it was either 15 minute or 30 minute exposures and another one that was 4 hours long.
This was using a 40D.
Short or long exposures for star trails
Each has it's own application. And I use both long and loooooooooong exposures for startrails equally, depending on the conditions and what my goals are from the imaging session
