[Skymist]

There has been no astrophotography thread for K-5 yet, so with your permission I'll start one. In forums without a specific forum for this kind of subject, I've seen a thread like this remain active for years. Let's see your K-5 sky photos, all kinds, including stars, planets, deep sky telescope objects, spacecraft, satellites, and atmospheric phenomena. Also advice, data, and ideas.

To start with, I want to post the data I took for the camera's built-in interval timer. It's a cool feature that the K-5 has the interval timer in the camera, so you can do without attaching an external one. In the menu, you can set the interval timer with a shutter interval, a total number of activations, and a time to begin the activations. The firmware acts like a finger pushing the shutter - that is, the interval you choose is the time between shutter activations, not the time from when the shutter last closed. That means, if the camera needs 1 second to process each shot, and you have the shutter open for 30s, you must set the interval to 31 seconds or more. That's because the camera must be ready for the next shot when the timer fires, or that shot will be skipped.

The issue is complicated by the Dark Frame Subtraction feature, DFS, which in the menu is called "Slow Shutter Speed NR". If DFS is on, after each shot the camera takes a reference frame with the shutter closed which is then processed with the image to remove hot pixels and other problems, at the cost of almost doubling the time per shot.

Using my K-5, I determined the minimum interval time for each of the long shutter times. Enter this value, or higher, into the interval time menu. Using too short a time will cause shutter activations to be skipped; using too long a time will unnecessarily waste time, and we don't want to do that if we are taking sky pictures. As you would expect, the values for DFS on are much higher than those with DFS off. You decide which you want to use.

Values in Bold are more efficient in time used, and have proportionally less dead time, than their neighbors. Use them if you can. For instance, 20s is a good choice for an exposure time - set the interval to 21s if you have DFS off, or 35s if you have DFS on.

Exp NoDFS DFSon
0.25 1. 1
0.30 1. 1
0.40 1. 1
0.50 1. 2
0.60 1. 2
0.80 2. 2
1.00 2. 3
1.30 2. 3
1.60 2. 4
2.00 3. 4
2.50 3. 5
3.00 4. 6
4.00 5. 9
5.00 6. 10
6.00 7. 12
8.00 9. 14
10.0 11 18
13.0 14 21
15.0 16 27
20.0 21 35
25.0 26 42
30.0 32 54


Some interval times are convenient divisors of an hour.
An interval of 6s gives 600 shots per hour.
An interval of 10s gives 360 shots per hour.
An interval of 12s gives 300 shots per hour.
An interval of 18s gives 200 shots per hour.
An interval of 36s gives 100 shots per hour.

[jbinpg]

Thank you for starting this thread. For timelapse wider angle astro shots, I have found the Remote Continuous Shooting drive mode better than using the interval timer for the K-5. One press of the remote locks the mirror up and begins the first exposure. The next frame will follow immediately after the first and while the 2nd frame is exposing, the first frame will be written to the card. As long as the exposure time is longer than the card write time (approx 3 sec for night shots), you will never miss any shots. The camera will continue to take frames until you decide you have enough. When you want to stop the sequence, just press the remote again or turn the camera off. The one disadvantage is that this process requires continuous power so it is best to have it on the AC inverter.

I have heard some say that there is no advantage to using high ISO for astro but this is not my experience. ISO1600 will necessarily capture fainter stars than ISO80 for the same exposure time. If you are taking multiple shots and stacking then I can see using base ISO but for timelapse I can easily capture stars down to magnitude 8 or 9 at ISO1600 and a couple magnitudes fainter at ISO3200.

Examples of astro timelapse are here:

Index of /

Some of them are quite blocky due to being animated gifs with the restricted gif palette.

Jack

[Skymist]

I have heard some say that there is no advantage to using high ISO for astro but this is not my experience. ISO1600 will necessarily capture fainter stars than ISO80 for the same exposure time. If you are taking multiple shots and stacking then I can see using base ISO but for timelapse I can easily capture stars down to magnitude 8 or 9 at ISO1600 and a couple magnitudes fainter at ISO3200.

For stacking, I've seen analyses which showed the ideal ISO values, which depend on the camera. Numbers for common Canon cameras are 400-1000 ISO. I've been shooting with 200 ISO, but since I process raw in Lightroom, I've sometimes used the entire +4 exposure boost it offers, so I think I'm going to try 800-1600 in the future to get more range and see how that goes. I think it's like most raw development - you set the camera as close to the ISO you want as you can, then you can make the smallest adjustment in the software.

[amaurosis]

Just found this terrific example of the K5's capabilities in this field. I find the result remarkable, considering that the kit lens was used! Can't wait to give this a try with a Limited lens or two

[TER-OR]

this is an idea I'd like to try the next time I visit my brother in Oregon and we go up on one of the dormant volcano parks. The caldera reflection of the striking night sky are really something to see. Thanks a lot for the post, I may do some practicing.

[DRabbit]

In some of my own research on this topic, the big key is avoiding star trails (unless you want them). The wider the lens the better. There's a formula...

600 ÷ focal length = exposure time (in seconds) to avoid star trails.

So for example... I shot some with the Samyang 14mm f/2.8 lens. It's effective focal length is 21mm on the Kx (the camera I was using at the time).

600 ÷ 21mm = 28.5 seconds.

So I shot at 30 seconds, wide open, ISO 1600 or 3200.

These aren't the most beautiful results, but it was my first attempt!









I'm looking forward to trying this again with the K-5... once I get a nice wide lens again

[Skymist]

This is one of my first K-5 astro photos, and I'm still intrigued by it. I've never seen a satellite move this slowly. To move this little in 10 seconds, it must be at an extremely high orbit. It was also very faint, and totally invisible to the naked eye, and in fact at first I missed it on the blown up shot entirely.

I'd like to figure out the math needed to determine the orbit height from the apparent speed. Perhaps I could then identify individual satellites easier.

This particular photo has about a 2 pixel track width. I found later that the infinity focus setting on my Zeiss 50mm lens is a tiny bit off, focusing "behind" infinity a tiny bit. In more recent shots I've been focusing by hand, and the track width is nearer 1.5 pixels. Nasty quantization shows up then.

This is the 100% crop of the shot, Pentax K-5, F/1.4, ISO 200, 10s, using a Zeiss T* F/1.4 50mm lens, converted to B/W in LR3, pushed about 4 stops. Some severe lens edge effects for the brighter stars, but then, I had no interest in the bright stars, only the faintest objects. Fairly poor sky conditions. Taken April 27th, just before midnight.

[Skymist]

I'd like to figure out the math needed to determine the orbit height from the apparent speed. Perhaps I could then identify individual satellites easier.

Thinking a bit, one can go this route to estimate it.

The satellite moved about twice as much as the stars in the exposure time.

Assuming the satellite was overhead, that means it would complete an orbit twice as fast as the stars, which means it would orbit in 12 hours.

Geostationary satellites are about 26000 miles high. This satellite orbits twice as fast, so it would not be that high. Let's start by guessing it would be half as high, or 12000 miles. The ISS is 190 miles high and orbits in 1.5 hours. If it were overhead, it would move in a track whose length was 24/1.5 = 16 times the length of the star trail.

All this disregards the earth's rotation though. If you look at the path of the object, and subtract out the vector of the star movement, the resulting vector shows that the true direction of the movement of the satellite is exactly perpendicular to the stars. The stars leave trails which are east-west - that is, they are perpendicular to the north-south axis. Since the true movement of the object is perpendicular to the stars, that identifies the satellite as being in a polar orbit. So now, we have a faint satellite, in a polar orbit, perhaps 10-12000 miles up. That might be enough info that I can start looking through lists of satellites. If it were for a military purpose, it would possibly be a lower altitude. So I'm guessing it is a scientific satellite.

[Skymist]

So now, we have a faint satellite, in a polar orbit, perhaps 10-12000 miles up. That might be enough info that I can start looking through lists of satellites. If it were for a military purpose, it would possibly be a lower altitude. So I'm guessing it is a scientific satellite.

Looking through the databases, it could possibly be one of ESA's Cluster satellites. They are scientific satellites which are in a very elliptical polar orbit. See link.

If so, it's pretty cool. This isn't one of the bright ones normally followed by satellite-watchers.

[Class A]

I think it's like most raw development - you set the camera as close to the ISO you want as you can, then you can make the smallest adjustment in the software.

I agree. The low noise levels of the K-5's sensor almost make the ISO choice irrelevant because pushing in post is almost as good. Still, there are slight advantages by using the analogue amplification that is employed by using higher ISO values.

Note, however, that going beyond ISO 1600 is not useful. Starting with ISO 3200, the K-5 just pushes the image numerically, i.e., you don't get any high analogue amplification but only run the risk of blowing out bright parts. For >ISO 1600 applications, it is better to underexpose in-camera and push in post.

The only reason I can see why one would choose lower ISO values (e.g., ISO 80 rather than ISO 800) is to expose longer and thus reduce the SNR.

So it seems to me -- I'm not an astrophotography expert though -- that one should expose as long as possible (avoiding unwanted star trails) and choose the highest ISO (<=1600) that corresponds to a desired exposure value.

Here's one of my very modest first attempts at capturing the night sky (with a lowly K100D). Featured is the Southern Cross as seen from the Church of the Good Shepherd at Lake Tekapo in New Zealand. In hindsight, the exposure was too long.

Since I uploaded this image, I used FocusMagic to reduce the star trails. While this is obviously not a perfect solution (FocusMagic assumes a linear motion blur) it considerably improved the impression of sharpness and more objects became visible. (I never uploaded the improved version that I obtained with FocusMagic because I don't want a duplicate and I cannot replace the existing one since I don't have a Flickr Pro account).

[Skymist]

Starting with ISO 3200, the K-5 just pushes the image numerically, i.e., you don't get any high analogue amplification but only run the risk of blowing out bright parts.

Of course, for star photos, blowing out the bright parts can be a virtue. That's how you see the dimmest objects. The only time you worry is when the dim object you want to see is next to a bright object.

That's must be a great constellation to have overhead. Lovely.

[Dr_who]

I'm currently looking at getting a wide angle lens, and can't decide between two. Either the 14mm 2.8 or the 15mm lmt. The 14mm opens up a bit more, but I've read the 15mm is quite sharp. 14mm could also add a few more seconds of exposure before stars start to trail. How much difference that would be I'm not sure. The 15mm being small as it is also makes a great portable everyday landscape lens as well, but I've seen amazing stuff from the 14mm as well. What to do?

[jbinpg]

I assume that formula to determine exposure time is for an equatorial zone shot which shows the highest relative star motion. One could likely get away with a much longer exposure at the celestial poles.

[oliver939]

@Drabbit,

your pics are very nice but exif says they're from Kx.
Do you have a K5 to try to do similar photos?

[Class A]

Of course, for star photos, blowing out the bright parts can be a virtue. That's how you see the dimmest objects.

OK, but you can blow out bright parts with post processing as well, if you wish to do so. At ISO1600 and above, pushing in post has no disadvantages, i.e., you can blow out bright parts, but you don't have to.

If you choose ISO 6400, for example, it may be that you blow out some bright parts, but unlike pushing in post from an underexposed ISO1600 image, you then don't have the option of not blowing them out.

That's must be a great constellation to have overhead.

Nearby where I took the shot, there is the Mt. John observatory. I took a tour and it was fascinating what constellations one could see with the naked eye because of the excellent conditions. The conditions are so good that they can actually search for extrasolar planets.