How to get long integration times

With astrophotography we don’t just take a single photo and call it a day. Instead, we take lots of photos (called subframes, or subs for short) and then integrate (combine) them together to make one image that we go onto edit. This process is sometimes called stacking. How much data we gather is the integration time. Combine 10 one-minute photos, and your final image’s integration time is 10 minutes.

When I started out in astrophotography, I’d be happy with getting just a few hours of integration time per target. Now I aim for around 16 to 24 and usually manage it, despite my cloudy skies in the South West UK. I’ve put this post together to hopefully explain why long integration times are worthwhile — especially for urban astrophotographers — and to give my top five tips to achieve them.

Why do I need long integration times?

We can break down every individual image we take into two categories: signal and noise.

• Signal is all the photons we want to record. Say we’re photographing a galaxy; all the light emitted by the galaxy that reaches our sensor, and is recorded, is signal. Signal is good. We like signal.
• Noise is everything our sensor records that we don’t want. Noise is bad. Our camera sensors themselves produce noise when in use. And I think of light pollution — which we have an overabundance of in our cities — as noise, too.

We want our images to have the best possible signal-to-noise ratio, or SNR for short. This means they look crisp and smooth with lots of detail. We do this by taking loads of photos and integrating them together. This improves the signal, while the noise (which by is more random by its nature) gets smoothed out. So, long integration times = better SNR.

How long is long enough? Well the longer the integration time the better, but it’s not practical (or fun) to spend a lifetime collecting data on a single image. I decided to conduct an experiment to find an answer for my location and equipment. I collected 20 hours of data on a single target, The Iris Nebula, and then made mini-images of various integration times to see the difference.

The first six hours make the biggest difference by far. After that there’s still improvement, and it’s worth keeping going, but we’re into the realm of diminishing returns.

I wanted to see the difference integration times make on background noise levels, so vlaiv from the Stargazers Lounge helped by taking data from my Bode and Cigar Galaxies photo and creating this comparison. In the left-half on the image you can see the noise level after two hours of integration. The right-half cycles in two-hour intervals all the way to 24 hours of integration. See how it gets much smoother with time?

I graphed the noise levels using PixInsight’s subframe selector function:

It matches the image quality of the Iris Nebula mini-pictures quite well.

After seeing these results I decided to aim for higher integration times. Nowadays I’m happy with anything in the range of 16 to 24 hours of good quality data. I can produce decent results with half that, but to be honest anything less and the SNR can be a bit too weak given how hard I push the data during processing. Any more and it’s not really worth it for the gain in SNR.

Also consider that the more subs you take, the pickier you can be about the quality of those you put into your final integration. (If you put rubbish in, you’ll get rubbish out). Take 24 hours of data and you can cut the worst 25% and still have 18 hours left.

I’m not saying that you should have the same target integration time, but it’s worth running a similar experiment for yourself to see how your local sky conditions affect your SNR. You may also like to check out this video and accompanying web app from Deep Sky Detail, which gives an indication of how long you should be aiming for.

With all that in mind, here are my top five tips to achieve long integration times.

1. Remote automation

Most of us start our astrophotography adventures with a DSLR and a remote shutter release, sat out in the cold snapping one picture after another. Remote automation is a huge leap forward. This allows you to control your system from afar — think telescope set up in the garden while you’re giving it commands from your living room sofa.

Remote automation allows your whole system to run much more efficiently and even pre-plan imaging runs that last an entire night. You go to bed while your camera keeps collecting photons. Sweet dreams!

There are different options available for remote automation. I used to hook up an old laptop loaded with Sequence Generator Pro to my telescope kit, and then TeamViewer to use the computer in my home to remote desktop onto the laptop. That worked fine, but now I use an ASIAIR Plus (review here), which I find to be more convenient and allows me to control everything from my phone.

Remote automation software often comes with plate solving that allows for automatic and very accurate target centering; i.e. your mount will move your telescope so it’s looking at exactly what you want to photograph. The first time you see a system plate-solving it just seems like magic. The really good thing is that you can use plate solving to get your telescope back imaging the exact same region of sky over many nights, which lets you rack up the integration time.

2. Plan your imaging sessions

If you want to crank up your integration times, plan your imaging sessions. The most important thing here is to choose a target that will be visible in your sky for as long as possible between sunset and sunrise. Ideally you want to be collecting your target’s photons as soon as it gets properly dark, right through until pre-dawn twilight. Avoid anything that’s going to dip below the horizon after a few hours of imaging. Likewise, you don’t want to be wasting precious hours of darkness waiting for your target to rise.

Planetarium software like Stellarium is really useful here. You can plug in your location and then see exactly where the best targets will be in your sky. Remember that you’ll likely be imaging the same target for several weeks if not longer, so ensure it’ll be well-positioned over that entire timeframe.

Here’s a useful trick. In Stellarium, go to Sky and viewing options (F4); Markings; then check the three boxes next to Meridian. You get the best quality data when your target’s close to the meridian line, so make sure you choose a target that will pass that area of the sky during your imaging session. You’ll also want to ensure that your remote automation software is enabled for meridian flips.

Advanced tip: you can create your own landscapes in Stellarium to know when that galaxy you want to image is going to move out from behind the trees. Plug that data into you remote automation software and you can actually image multiple targets over a night without any wasted time.

I also use Charles Bracken’s The Astrophotography Planner Custom Digital Edition (review here), which I recommend.

3. Optimise your set-up and pack-down

It’s critical to have an efficient set-up and pack-down procedure. The more hassle it is to get your kit ready to image, the less likely you are to actually do it. Think about how to trim time off every stage. Every minute helps!

The ultimate solution is an automated dome, or perhaps roll-off roof shed. Those are big and expensive, however. Personally, I use a DIY pier (sometimes called a “Todmorden pier”) in my garden. This is so good I don’t know how I coped without it. In short, a pier is a permanent fixture with your mount head on top, replacing your tripod. Simple but brilliant, and highly recommended.

Before, when I wanted to set up for an imaging run, I’d need to lug the tripod out from the shed and put it in the right position; fetch the telescope from the house and attach it; sort all the cabling; then polar align, which can be hair-tearingly frustrating to get just right. That entire process would take me about 20 – 25 minutes, and the polar alignment could only be done when it was dark.

Now, with my pier, there’s no tripod positioning or polar aligning. Most of what I need is outside already. I just take the telescope from my house and attach it to the mount. We’re talking literally five minutes from being in the house thinking “it’s clear skies tonight” to gathering photons. Plus, in the summer when it doesn’t get dark until very late, I don’t need to wait until I can actually see any stars; pier plus remote automation means I can plan the imaging run any time I want, and just set it to commence when it’s dark.

You might be thinking that the ~20 minutes saved on set-up time is good, but not a huge amount given that I’m espousing the virtues of clocking up 20 hours of integration time. You’d be right, but here’s the thing: the fact that it’s now so quick and easy to set up your telescope means that you’re far more likely to actually do it. Let’s say that there’s only an hour or two of clear skies forecast for the night. It’s a lot of hassle to set up and align the tripod for that. But five minutes because of the pier? Alright, let’s go for it! Hey presto, thanks to the pier you haven’t gained 20 minutes of integration time that night, but rather two hours. And all those one, two, three hours of bonus time add up quite quickly and before you know it you’re into double-digits.

If you’re renting your home, your landlord/lady might not be too happy about you digging up the garden to build a pier. I’ve heard about people filling tyres with concrete and using that as a base for a DIY pier. Sounds like quite a good idea. And sure, a proper dome would be even better — but then you need a lot more space, and about ten times the budget.

I’ve talked a lot about piers because that’s what I have, but of course that might not be a valid option for you. If you live somewhere that makes a pier completely impossible, like a tower block, then it’s still crucial to consider how you can make your set-up and pack-down as quick and easy as they can be. Even leaving your telescope set up with a cover (Cygnus Astro are recommended — review here) for a few nights on the trot will help a huge amount. Small improvements add up to big differences. Can you get your set-up routine down to 20 minutes? Great, now aim for 15. Then 10. Then 5…

For more info about my pier, see this post on my astrophotography kit.

4. Shoot short subs (if your camera has a modern CMOS sensor)

For many years astrocameras used a type of sensor called CCD. To be effective these needed sub lengths of many minutes — think 10, 20, even more. Nowadays we have a new type of sensor called CMOS. These have several benefits, particularly the fact that you can shoot much shorter subs. The important thing is the total integration time. How you get there doesn’t matter so much.

For the technical reasons why this is the case check out this video: Deep Sky Astrophotography With CMOS Cameras by Dr Robin Glover.

There are definite practical benefits to shooting short subs instead of lengthy ones:

• The shorter your sub is, the less accurate your mount’s tracking needs to be. Shooting a 10-minute sub requires far greater precision than a one-minute sub.
• If you do have any mount wobble or other issues, it’s better to have one short sub ruined rather than one long sub.
• Shorter subs mean you can take advantage of gaps in the cloud to keep gathering data.

The one main drawback to talking lots of short subs instead of fewer long ones is the amount of space they take up on your computer (and corresponding time to process the data). Every sub from my 2600MC is 50MB. That adds up quickly when taking lots of short subs, so you need a big harddrive and ideally lots of RAM in your computer.

In theory I could shoot subs as short as 10 seconds and get good results. I’d need an entire harddrive for each image though, so it’s not practical. I’ve settled on 120-second subs as standard, which seems to be a sweet spot for getting the benefits of short subs while not melting my computer.

5. Commit to it!

Finally, it takes a certain amount of dedication to achieve long integration times. First, there’s always the temptation to call it quits and move onto another target. Beginners in particular tend to get itchy feet; after all, there’s just so much out there to image! This feeling does fade after you get a few long integration images, and realise that the end product is much better as a result.

Second, we astroimagers can sometimes be quite good at thinking of excuses to not image. It might get cloudy later; the Moon is out; it’s Summer and the sky will only be dark for a few hours… the list goes on. My trick is to ditch all that and make it a binary decision: if the forecast is clear, I’ll be imaging. If not, I won’t. Simple but effective. Note that you still need to be careful about not letting low-quality subs through; you want quantity and quality!

So there you have it

I hope that if you’re currently getting integration times of just a few hours then you’ve now been given a nudge to collect more photons. It makes a huge difference to the quality of your final images. Remember that I’m in cloudy South-West UK, but using these tips I can produce a ~20-hour integrated image every three weeks or so in the Winter, or six weeks in the Summer.

Also check out my article on the Top 10 Upgrades, many of which will help you with this most noble of aims.

---

I’ve given you the gift of knowledge. Could you give me the gift of cash?