Beginner astrophotography help.

[Steenamaroo]

Hi all,

I don’t really have a specific issue but was hoping someone might help with general advice for astrophotography.

I’m using a 130pds with Canon 500D, shooting raw, on a belt driven guided Vixen-GP.
Guiding is SVBony SV-106 50mm with qhy5l-ii-c.

 

I did two test sessions - One on Monday night, one on Tuesday.
Both nights were remarkably clear. (I know...hard to believe!)

The first night I captured m81+m82 together, prime, with 1 minute 800 ISO images over about 90 minutes.
I know that’s not a huge amount, but enough to gain some experience+confidence, I think.

 

I learned that 1 minute at 800 probably isn’t enough but, regardless, the first subs looked usable.
I noticed, after the session, that subs became darker as time went on.
Shooting between around 22:00 and midnight that made some sense, as the sky was getting darker, but the subject (m81+m82) and stars got substantially darker too.

I expected the backdrop would darken over time but the target+stars should remain around the same brightness. Am I wrong?

The stack of these images really pleased me, though - Although the target galaxies weren’t particularly bright or well developed, the image was nice and clear and the backdrop smooth and dark.
A success, from my perspective.

 

For the second night I decided to improve by increasing the sub lengths and also decided to 2X Barlow and focus on just m82.
I’m aware Barlow would make the subs much darker than at prime so I did some tests at 3 minute exposure.
Not seeing a huge difference in target brightness I tried 4, 5, 6…and finally 8 minute.
Surprised at how little difference there was I tested increasing ISO bit by bit too.

I ended up taking 8 minute exposures at 3200 ISO, wondering why the target and stars still looked so faint.
I only got around 10 subs but, this time, I noticed they became brighter as the night went on - The exact opposite of what happened the night before, shooting at the same time (22:00 to midnight)
The stack this time was very disappointing - Very grainy/noisy, backdrop nowhere near as dark - Pretty much monochrome looking - It looked like something from the guide-cam but noisier!

I’m stacking with DSS, took 5 or more dark frames for each session, and used the same flat frames for both sessions.
I also tried stacking both sessions without flats, which didn’t account for the difference.

Environment was the same both nights - Same setup position, same neighbourhoods lights (or unusual lack thereof!), no rising moon…

I suppose my questions are

1. Why would my lights get darker (background, stars, and target) as the night went on on night one?
2. Why would my lights get lighter (background, stars, and target) as the night went on on night two?
3. Why would night one stack perfectly but night two stack looks terrible?
   Safe to say 3200ISO was pushing it but the stack should still look better than any sub, no?

Many thanks to anyone who read my wall o' text.

 

 

 

First session first sub

First session last sub

First session stack

Second session first sub

Second session last sub

 

First session stack

 

 

Edited March 3, 2022 by Steenamaroo

[Stefan73]

I'm a beginner too so this might be a bit blind leading the blind... but my understanding is beyond a certain ISO there's no point in increasing because you're just increasing noise too.  See http://dslr-astrophotography.com/iso-values-canon-cameras/

I think what you're aiming for is a histogram which isn't clipped to the left or right when you take a photo and not worry too much about how light or dark it looks.  The magic of post processing can sort all that out when applied to the stacked image.  The night 2 one looks quite noisy but I'm not sure if post processing could save it.

[happy-kat]

For the canon you have 800 or 1600 ISO I think, anything higher will be noisy.

Did your secondary mirror get dew on it?

Look for where the histogram is rather than just eyeballing the image when looking at how bright it looks

5 darks would just add noise, flats are good to use

Sky conditions can change throughout a night and be different between days even if it look clear

Just some thoughts

[Elp]

How high in altitude were the targets? What and where is your local light pollution like? Its strange that the lights were getting uniformly darker or brighter, if the targets start getting nearer to the horizon you will get gradient brightness (offset to one side of the frame) due to getting closer to light pollution, but your situation is very odd. 

[Steenamaroo]

Thanks for the thoughts, folks.
I'll try to cover everything.

@Stefan - I realise increasing ISO increases noise, but also that stacking reduces noise.
I wasn't necessarily expecting a good result from the second set, but certainly the stack should always be better than the subs, no?

Admittedly I didn't check the histogram, although clipping there would indicate overexposure?
Agreed - Checking histogram in future is a good idea, but I'm pretty sure I've seen people's subs as bright as/brighter than, the ones I've shown here.

 

@happy-kat - Agreed - 3200 is pushing it. Thank you though.
I suppose I was kind of baffled as to why it seemed impossible to overexpose.
Maybe that's just my lack of experience with fainter DSO's but gradually increasing exposure duration and ISO I expected to, at some point, look at a sub and think, "wow, that's way too far".
That didn't happen.
Comparing the first sub from night one and night two, I don't think it looks like I did an 8X on exposure and a 4X on ISO, even taking the Barlow into account.

"5 darks would just add noise, flats are good to use"
How's that?
Darks are pretty much essential since I've quite a few dead pixels, and my understanding is, just as with lights, more = better?

"Sky conditions can change throughout a night and be different between days even if it look clear"
I suppose that's true. That said, these were two ridiculously clear nights. Pretty much no twinkling - definitely no visible cloud or haze/mist...nothing like that.

"Did your secondary mirror get dew on it?"
I have no idea - Maybe that's something to watch. Thanks for the thought!

Local light pollution usually isn't great - There's usually someone with a security light on all night or whatever but for these two nights there was nothing.
It really was as dark as it gets here.
I think that's why I'm a bit bummed. I expected stack 2 to be more impressive but too noisy and then I'd do night three probably with 10 minute + 800 ISO, and a lot more stacks,
but the results from these two nights kind of threw me a bit.

 

@Elp - Targets were about as overhead as it gets. M81+M82 between 22:00 and 00:00 in UK.
Moving closer to, or farther from, the horizon or any source of light pollution would make sense but, unfortunately, that wasn't the case here.

[ONIKKINEN]

Drop the barlow, it only makes things worse for you. The target is emitting the same amount of light all the time, but with a barlow you are spreading this light thinner, over too many pixels and so thee frame looks darker. What looks darker (and is) would also be the target itself. Without the barlow you are shooting at 1.5'' per pixel resolution, which is pretty well matched. With the 2x barlow you are shooting at 0.75'' per pixel which is definitely without a shadow of doubt oversampled and not helpful. When you are oversampling you reduce the speed of the system by spreading light over too many pixels but dont actually capture extra detail from the object even if the resolution seems greater. The second stack looking noisier and worse is to be expected, unless you spend 4x the amount of time capturing it comparing to the no barlow one.

This sort of differential brightness during the night is not that strange, although ideally you would get rid of the outlier subs when stacking. You can have high cloud that is not visible to the naked eye and it will look like the image got a bit brighter, but lost some of the faintest stars. Since you have local light pollution it could also be that at some points during the night your scope is pointing closer to a local light source than others. Subs can also get darker during the night as the target rises higher, and so you shoot trough less atmosphere. There is a dramatic difference in light polluted skies between shooting lets say 35 and 70 degrees above the horizon. Some street lights and yard lights can also be automated to shut down at some point of the night, so the amount of light pollution is reduced. The darkest subs in your session are the ones least affected by your conditions, so would be the best. Subs getting darker as the night goes on is a good thing.

ISO speed and noise have some misconceptions around it, increasing your ISO will lower your read noise, not increase it. That said, there is a point of diminishing returns for most cameras where it makes little sense to keep increasing ISO. Dont have a 500D myself, but from what i have read the astrophotography recommended ISO would be 1600. You definitely also dont need to be taking 8 minute subs. 1 or 2 minute subs will do just fine (without the barlow that is. With barlow, dont expose and take the barlow out), even shorter if you have lots of light pollution. If your histogram is not touching the left edge the subexposure is long enough, no special trick to it. If your histogram is getting close to the right edge you would be clipping data, so best to avoid subs this long.

Edited March 3, 2022 by ONIKKINEN

[Swillis]

I'm by no means an expert at this, but have a couple of pointers I have picked up in my short time on here which may help. 

You said you are using the same set of flats for both sessions?  I think this is generally not the way to go as dust can shift between sessions, but also you have added a Barlow which will change the optical path quite significantly.

Also, it has been recommended on these forums that darks should not be used for canon DSLRs. Firstly, the temperature cannot be matched to the lights, but canons do some dark subtraction (even with that setting turned off), using a strip of masked off pixels down one edge. 

Also you didn't say if you used bias frames?

So, flats for each session without removing the camera from the scope. Don't use darks but do use bias frames. 

Happy to be corrected about this by someone more experienced!

 

[vlaiv]

15 hours ago, Steenamaroo said:

Why would my lights get darker (background, stars, and target) as the night went on on night one?

This is normal behavior.

Most of "brightness" of the sub comes from light pollution. As night progresses - traffic dies down, people go to sleep and turn off their lights - in general light pollution drops down a bit.

Another factor - but you'll be judge of that, is that target moves across the sky and you might start your session when target is lower in altitude as it moves towards zenith. If you are avoiding meridian flip, you are likely to shoot your target at one side of meridian only. It can be either climbing towards zenith or setting during session.

If it is "rising" - background will get darker because zenith is usually place with the least light pollution (but this is not generally the case and depends on your local LP sources).

15 hours ago, Steenamaroo said:

Why would my lights get lighter (background, stars, and target) as the night went on on night two?

When background is getting brighter as night progresses - that is usually case of issues with transparency.

It can be that target is setting - and as it moves, it goes into area closer to horizon that is brighter due to light pollution, but if that is not the case, then it is local transparency that is getting worse as night progresses.

For same level of light pollution - how bright the sky will be, depends on how thick atmosphere is. This is one of the reasons why there is more glow near the horizon, but also the reason why we can "play light sabres" with hand held torches in the fog. With thick atmosphere light scatters around much more.

If there are thin high altitude clouds (ones that you can't spot at night but during the day look like haze high above) or there is some general haze in the air - it will brighten your background.

15 hours ago, Steenamaroo said:

Why would night one stack perfectly but night two stack looks terrible?

Image will look good if there is sufficient SNR - or signal to noise ratio.

As name says - SNR depends on two things - signal and noise (and represents their ratio).

If on night two, you had poor transparency - it will affect SNR greatly. It affects both signal - by blocking light from the target, and noise - by scattering light pollution and making background brighter. More background signal means more associated noise.

It's all about signal to noise ratio.

[Steenamaroo]

@vlaiv - Thanks for the information, although your descriptions explain why the background would become darker, or lighter, over time.

If evening one's images started bright due to light pollution or transparency, and got darker because conditions improved, shouldn't the galaxy and stars have become brighter?
Conversely if evening two's images started dark and became brighter due to conditions worsening, wouldn't the galaxy and stars have become faint over time too?

As I say, folks, both nights were shooting the same target at the same time, and the target was more or directly overhead, far from the horizon.
Position in the sky was the same for both sessions.

@Swillis
My understanding with dark frames is that I don't really need bias frames if I have dark frames,
and I do need dark frames because without them hot pixels are a bit of an issue.

How do you mean `Firstly, the temperature cannot be matched to the lights`? The temperature of the sensor?
I take my darks on site with the same exposure length as the lights, so it should be the same?

[vlaiv]

13 minutes ago, Steenamaroo said:

If evening one's images started bright due to light pollution or transparency, and got darker because conditions improved, shouldn't the galaxy and stars have become brighter?
Conversely if evening two's images started dark and became brighter due to conditions worsening, wouldn't the galaxy and stars have become faint over time too?

If you really want to examine brightness of individual stars - you need raw data that is still linear.

As soon as you stretch data - you can't reliably tell if one stars is brighter than the other. This is because signal is additive and stretch is non linear.

Data is still stretched if you just converted it to .jpeg from raw in your favorite program because .jpeg uses sRGB color space which has gamma of 2.2 (which is non linear transform).

Post two .cr2 files - one from beginning and other from the end of session - for inspection of actual pixel values (both background and star) - so we can see if there is really dimming of the stars and objects.

There is one special case where you can have both darker background and darker stars (if that is indeed the case) - dew on primary mirror or front lens. It builds up and blocks the light and everything turns dark, however, it is not gradual process - it usually happens much faster than duration of the session.

If you happen to have sudden drop of brightness over say 5-10 subs, then it can be down to dew. In the end, you'll essentially end up with completely dark subs as dew completely blocks the light.

23 minutes ago, Steenamaroo said:

My understanding with dark frames is that I don't really need bias frames if I have dark frames,
and I do need dark frames because without them hot pixels are a bit of an issue.

How do you mean `Firstly, the temperature cannot be matched to the lights`? The temperature of the sensor?
I take my darks on site with the same exposure length as the lights, so it should be the same?

Sensor temperature dictates how much dark current there will be.

In general - if you take darks at the same exact temperature as your lights, then you should be fine, but that in practice never works because temperature is not constant over night. There is often drop of more than few degrees C.

People tend to shoot their darks at the end of the session - but that means that lights from the beginning won't be properly matched.

Dedicated astro cameras with set point cooling - deal with this easily as you can always set exact temperature on your sensor.

DSLR cameras (modern ones anyway) deal with this in different way. When you take your light exposure - some of the pixels surrounding the sensor (they are part of sensor but not exposed to the light because they have physical mask) record dark current. Their value is then subtracted from the final image.

This does not remove hot pixels, but there are other ways to deal with hot pixels. This also means that you don't need dark frames, and sometimes you don't even need bias frames.

Bias is needed with DSLR only if there is distinct bias signal. You can check this by taking two sets of bias subs - say 50-100 per batch and stack them and inspect them. If stretched they show the same pattern - then subtract one from another. If in resulting image you find that pattern is gone and you get smooth noise - its worth subtracting bias.

If you find that your original bias masters don't have any sort of repeating signal (this will usually be some sort of vertical or horizontal lines or something like that), then you don't really need to remove bias.

Here is example of Bias signal from Canon 750D

(you can see horizontal and vertical lines as well as darker and lighter regions)

Back on the hot pixels. There are several ways to deal with them:

1. Dither and use sigma reject.

2. Use hot pixel map created from master dark

3. Use cosmetic correction algorithm (it identifies single hot/cold pixels and removes them)

[Swillis]

Thanks @vlaiv I was just writing that you can explain it much better than me, when your reply popped up. Also that I picked those pointers from you in the first place.