Why is my M31 terrible?

[rnobleeddy]

Very new to this, but after a frustrating summer with clouds and getting used to equipment, I finally got everything working last week. I've captured short amounts of data on M13, M27 and M38 that aren't going to win any awards, but they look OK and reassure me that everything is working reasonably well and I'm stacking correctly. 

Then there's  M31!

Across 2 nights I have about 1hr20m of data spread across 48 subs consisting of some 4 min and 3 min subs and a lot of 1 min subs, with corresponding darks, flats and bias frames for each night and exposure length. I stacked these as 3 groups in DSS, but I've also tried individual sets of data and leaving out flats/darks/bias in various combinations to rule out issues with any individual aspect. I've taken out the obviously bad subs but a few with star trails appear to have made it to the final image below, and there's probably also a soft focus on at least one night. However, none of this would appear to explain the problem - which is there's almost no data on the outer part of the galaxy. 

Night 1 (about 1/3) of the data had a bright and almost full moon close to M31. Last night was about 60% moon but it wasn't very close at the time I was imaging. Kit is a 130PDS, EQ5, modded Canon 450D with bader cc, and Astronomik CLS CCD. Bortle 5 skies.

A couple of the stacks are below and one of the best individual shots. Whilst some of the stacks with all the data seem to skew towards red, none of the attempts I've made have any real detail in them, presumably because it isn't there.

I then see posts on Reddit like this and wonder what I'm doing wrong!

So wondering if anyone with more experience could advise:

- is this what you'd expect to see given my equipment or is something awry?

- if this is expected, what would have the biggest impact on making the image better? Clearly more subs at longer exposure can't hurt.

- Does the moonlight effectively kill the image? Or light pollution? I can drive a relatively short distance to a darker site but I'd rather work from my garden if at all possible, because that's possible most clear nights, whereas moving the kit is time consuming

- I tend to skimp on dark frames a little (I have 10x for the 1 minute subs, and maybe 4 each for the 3 minute and 4 minute subs) - is this an issue?

- are there any other tips/tricks to pull any more structure out? The images below are auto-scaled with the ASI fits viewer.

 

Stacked using a subset of the data with flats/bias/darks

All data stacked, just lights

One of the 3 minute lights

 

 

 

 

 

 

 

 

[MarkAR]

More darks will definitely help. Light pollution can be filtered depending on the source but the moon is a pig for killing anything but H-alpha narrowband.

[Seelive]

Have you applied any stretch to the images? If they are still linear the outer structure will be faint.

[Aramcheck]

I'd suggest getting a much longer total imaging time & to stick with one duration exposure, rather than a mix. If you can get away with 3 min subs then stick with that & get a total exposure of at least 1 hour (& preferably a lot longer).

For Darks and Biases you can take more (about 50 bias + at least 25 darks), but you only have to do that once to generate Master Bias / Master Dark images, which you can then reuse. Flats, of course, have to be done each night, or each time the imaging train is changed.

M31 is also a big image for the 130pds. I've only had one go at it & it took me 6 attempts at processing to bring out as much detail as I could in the outer edges, which are much fainter than the core.

The red colour cast in the top image will be eliminated during the processing (colour calibration).

Also worth looking at http://dslr-astrophotography.com/iso-values-canon-cameras/ which indicates that ISO 400 or 800 is optimum for the 450D.

I'm also not sure how suited the Astronomik CLS CCD is suited for galaxies?

Cheers
Ivor

[tomato]

I think the moon (even a partially illuminated one) has a big impact on capturing the fainter detail on M31, and any other DSO for that matter. I captured some 45 sec subs with an 8” F2 RASA and OSC camera on two nights recently, one with no moon and one with a 50% moon about 10 deg above the horizon. 
Based on star analysis in APP, the moonlit subs were about 60% of the quality of the moonless shots, and qualitatively the moonlit stack was inferior to the moonless stack, even though it had more integration time.

[Chefgage]

38 minutes ago, Aramcheck said:

 

For Darks and Biases you can take more (about 50 bias + at least 25 darks), but you only have to do that once to generate Master Bias / Master Dark images, which you can then reuse. Flats, of course, have to be done each night, or each time the imaging train is changed.

I though Dark frames are temperature dependent. So the dark frames need to be done at each session??

[Pryce]

46 minutes ago, Chefgage said:

I though Dark frames are temperature dependent. So the dark frames need to be done at each session??

You can build a dark library over time and match the exposure and tempratures of your darks to your subs.  Or so I hear. I haven't done it so I don't know if it actually works.

 

Anyways. For OP:

In the image from the reddit post it looks like you're using 1 and 2 second subs?   Are you operating unguided with the EQ5? If you can get nice round stars at 3 or 4 minute exposures, stick to that and add more dark frames! 

And obviously shooting on a moonless night is way better.  Try picking your targets depending the moon as well. Shooting M31 might not be optimal for you right now, so give something else a tey and see how it goes

Edited September 11, 2020 by Pryce

[rnobleeddy]

4 hours ago, Seelive said:

Have you applied any stretch to the images? If they are still linear the outer structure will be faint.

Those were probably linear. I'm still new to post processing (even unsure which tool to buy!) but I followed a tutorial using GIMP this morning and teased out a little more detail, at the expense of more noise.

 

 

 

 

Edited September 11, 2020 by rnobleeddy

[rnobleeddy]

9 hours ago, MarkAR said:

More darks will definitely help. Light pollution can be filtered depending on the source but the moon is a pig for killing anything but H-alpha narrowband.

Thanks - I'll have another go when the right combination of moon/weather/spare time allows!

 

[rnobleeddy]

1 hour ago, Aramcheck said:

I'd suggest getting a much longer total imaging time & to stick with one duration exposure, rather than a mix. If you can get away with 3 min subs then stick with that & get a total exposure of at least 1 hour (& preferably a lot longer).

For Darks and Biases you can take more (about 50 bias + at least 25 darks), but you only have to do that once to generate Master Bias / Master Dark images, which you can then reuse. Flats, of course, have to be done each night, or each time the imaging train is changed.

M31 is also a big image for the 130pds. I've only had one go at it & it took me 6 attempts at processing to bring out as much detail as I could in the outer edges, which are much fainter than the core.

The red colour cast in the top image will be eliminated during the processing (colour calibration).

Also worth looking at http://dslr-astrophotography.com/iso-values-canon-cameras/ which indicates that ISO 400 or 800 is optimum for the 450D.

I'm also not sure how suited the Astronomik CLS CCD is suited for galaxies?

Cheers
Ivor

Thanks,

The exposure length change was down to botched polar alignment on the second night. My mount guides poorly in DEC and probably needs a little more work on the DEC backlash. Assuming I get good PA it seems that 5 minute subs are fine, although I'm still working out the sweet spot for balancing exposure time and reject rate. I also have an EQ6 mount that I'm working on tuning, which will hopefully just be better.

I was at ISO800. I was hoping that I could just leave the CLS CCD in my camera for everything - are there different filter requirements for galaxies vs nebula (assuming I'm sticking with OSC)?

 

[x6gas]

As others have said, your results are pretty much what would be expected with your kit, data and conditions.

We are all trying to dig out faint details and in broadband images like this dealing with light pollution is a big issue.  The data you collected with a near full moon close to the target will mean that data will be heavily compromised - the glow from the moon will be washing out the fainter details of the galaxy - and if you include those subs in your stack it will reduce the overall signal to noise ratio.  This is why many imagers will use moonlit nights to check their kit or image narrowband targets.  At the very least you should aim for a target as far from the moon as possible but even then the results will be compromised.  Better skies may help a bit if there is less air pollution to scatter the moonlight, but it won't make a huge difference and it really is better to avoid broadband imaging of deep sky objects if the moon is a factor.  You can always image the moon, of course, or even widefield shots of the brighter planets...

As I say you really want to be able to process the data (stretch it) to bring out the fainter details and the bigger the difference between those details and the background sky the better.  As you stretch the data, though, you'll be bringing out the noise in the image as well as the faint details.  As others have said, stick to one exposure time and more data is the key here, so long as the quality is good.  The stacking software can easily take care of satellite trails and even the odd subexposure with less than perfect stars but it is important to exclude subs that have a poor signal to noise ratio - i.e. that the target is getting washed out against the background.

You don't have enough dark frames and that will mean that you are introducing noise to each light frame during the calibration process.  You are better off without them unless you have at least 16-20 in my opinion.  These should be taken at a similar temperature to your light frames - within 2 or 3 degrees won't make much difference.  If you keep a note of the temperature when you take your lights you can build a library of darks at, say, 0, 5, and 10 degrees C over time that would cover you for lights taken at a temperature of -3 to +12 or so.  This is another good reason to keep your light frames to specific times as each dark needs to be the same duration as the light.

All in all I think your results are pretty good considering the conditions!

Clear skies and stay safe, Ian

Edited September 11, 2020 by x6gas
corrected light pollution to air pollution

[Seelive]

7 minutes ago, rnobleeddy said:

Those were probably linear. I'm still new to post processing (even unsure which tool to buy!) but I followed a tutorial using GIMP this morning and teased out a little more detail, at the expense of more noise.

 

 

 

 

When you stacked the different image durations in DSS did you use the Entropy Weighted Average stacking mode? That will preserve the fainter detail captured in the longer exposure whereas if you used any of the other methods, the larger quantity of the short exposures will tend to 'wipe out' any of the fainter detail captured in the longer exposures.

Stretching will increase the noise. I tend to carry out noise filtering/reduction prior to stretching.

[rnobleeddy]

1 hour ago, Pryce said:

You can build a dark library over time and match the exposure and tempratures of your darks to your subs.  Or so I hear. I haven't done it so I don't know if it actually works.

 

Anyways. For OP:

In the image from the reddit post it looks like you're using 1 and 2 second subs?   Are you operating unguided with the EQ5? If you can get nice round stars at 3 or 4 minute exposures, stick to that and add more dark frames! 

And obviously shooting on a moonless night is way better.  Try picking your targets depending the moon as well. Shooting M31 might not be optimal for you right now, so give something else a tey and see how it goes

Sorry for the confusion - the Reddit post was someone else's work I stumbled upon when I was waiting for DSS to stack my attempt, and I couldn't understand how far less data resulted in better results!

For my subs, the EQ5 is guided. The mount has terrible DEC performance and so if I get good PA I tend to be able get up to 5 minute subs. The 1 minute subs in this example were because I accidentally moved the PA, so I'd get 3-4 minutes of solid guiding (less than 1" RMS) and then the DEC axis would jump by 10" and ruin the sub. Bizarrely, rather than fix the PA I decided to go with 1 minute subs and throw away every 4th one. 

 

 

[Pryce]

Oooh!   Thats my bad for not reading properly.  I blame dysclecszia! 😂

In that case I'd say more darks, moonless night and fine tuning of your post processing 

[Seelive]

2 hours ago, rnobleeddy said:

 

For my subs, the EQ5 is guided. The mount has terrible DEC performance and so if I get good PA I tend to be able get up to 5 minute subs. The 1 minute subs in this example were because I accidentally moved the PA, so I'd get 3-4 minutes of solid guiding (less than 1" RMS) and then the DEC axis would jump by 10" and ruin the sub. Bizarrely, rather than fix the PA I decided to go with 1 minute subs and throw away every 4th one. 

 

 

Have you tried loosening the dec axis. On delivery, the dec axis on my Skywatcher mount was far too stiff. PHD would start firing correction pulses to the dec axis but because of the stiction, initially it wouldn't move but eventually it would suddenly jump. Slackening of the dec axis retaining nut slightly solved the problem.

[alacant]

Hi

The 450d is noisy so best to lose the dark frames, dither between each exposure, use IS0 400 and stack using a clipping algorithm.

HTH 

[rnobleeddy]

4 hours ago, x6gas said:

As others have said, your results are pretty much what would be expected with your kit, data and conditions.

We are all trying to dig out faint details and in broadband images like this dealing with light pollution is a big issue.  The data you collected with a near full moon close to the target will mean that data will be heavily compromised - the glow from the moon will be washing out the fainter details of the galaxy - and if you include those subs in your stack it will reduce the overall signal to noise ratio.  This is why many imagers will use moonlit nights to check their kit or image narrowband targets.  At the very least you should aim for a target as far from the moon as possible but even then the results will be compromised.  Better skies may help a bit if there is less air pollution to scatter the moonlight, but it won't make a huge difference and it really is better to avoid broadband imaging of deep sky objects if the moon is a factor.  You can always image the moon, of course, or even widefield shots of the brighter planets...

As I say you really want to be able to process the data (stretch it) to bring out the fainter details and the bigger the difference between those details and the background sky the better.  As you stretch the data, though, you'll be bringing out the noise in the image as well as the faint details.  As others have said, stick to one exposure time and more data is the key here, so long as the quality is good.  The stacking software can easily take care of satellite trails and even the odd subexposure with less than perfect stars but it is important to exclude subs that have a poor signal to noise ratio - i.e. that the target is getting washed out against the background.

You don't have enough dark frames and that will mean that you are introducing noise to each light frame during the calibration process.  You are better off without them unless you have at least 16-20 in my opinion.  These should be taken at a similar temperature to your light frames - within 2 or 3 degrees won't make much difference.  If you keep a note of the temperature when you take your lights you can build a library of darks at, say, 0, 5, and 10 degrees C over time that would cover you for lights taken at a temperature of -3 to +12 or so.  This is another good reason to keep your light frames to specific times as each dark needs to be the same duration as the light.

All in all I think your results are pretty good considering the conditions!

Clear skies and stay safe, Ian

Thanks - this was exactly what I was hoping was the case. I guess the correct way of thinking about it is that I'm lucky to have got a decent image of M27 given the moonlight, rather than worry why M31 wasn't as good.

Once I've had a few more good runs I'll work out the best exposure length is and work on a dark library.

[rnobleeddy]

1 hour ago, Seelive said:

Have you tried loosening the dec axis. On delivery, the dec axis on my Skywatcher mount was far too stiff. PHD would start firing correction pulses to the dec axis but because of the stiction, initially it wouldn't move but eventually it would suddenly jump. Slackening of the dec axis retaining nut slightly solved the problem.

Mine was actually loose to start with and needed tightening. I've since followed a few of the steps in the video below, including swapping the worm shaft bearings. It's hard to know what helped, but RA now guides at less than 1" which I'll take for this mount. I've only had two proper sessions but DEC was fine the first time with good PA. The second time, I accidentally moved the alt bolt after I'd started imagine, and PHD2 later reported a 1 degree PA offset, so I'll probably let the mount off in that case!

 

 

[rnobleeddy]

1 hour ago, alacant said:

Hi

The 450d is noisy so best to lose the dark frames, dither between each exposure, use IS0 400 and stack using a clipping algorithm.

HTH 

Thanks - I'm aware of dithering but couldn't seem to find that option in EKOS when I looked. It definitely exists though. 

I can probably do the research myself, but is it dithering or dark frames, or would both still be worthwhile? I've skimped on darks because it's usually been quite late before I've moved on to them - this should be less of an issue with the nights getting longer.

[Seelive]

Have you checked the PHD guide logs with the viewer? Often they can help in diagnosing any guiding problems. I normally get under 1" in both RA and Dec even with a large PA, but Dec is always slightly worse which I assume is due to the large backlash as it takes PHD2 longer to correct.

Providing you use a sigma stacking algorithm you shouldn't need darks if you dither - using both adds little or no benefit. Darks are intended to get rid of coherent noise which sigma stacking should also do if you dither. But too few darks will add random noise to your final image that (in theory) sigma stacking won't. Having said that, due to my setup, I don't (can't) dither so normally rely on a darks library (although imaging in a Bortle 7 sky, the sky background random noise normally exceeds the random noise generated by the camera so using darks makes little difference when I sigma stack).

[alacant]

On 11/09/2020 at 16:10, rnobleeddy said:

couldn't seem to find that option in EKOS

Ah, ok. It's:

Guide tab -> options

Using EKOS you also have the option of multistar guiding which will help with your star shapes and sizes. I see PHD2 have multistar at the 'coming soon' stage too. 

If you can get your hands on one of the newer eos cameras, especially the ones with the 18mp sensor, so much the better.

HTH

Edited September 16, 2020 by alacant

[Chris33benoit]

Question.  I have tried to shoot M31 with a DSLR and a barn door tracker.  Here are my settings:  600 lights (5 sec exposure), 60 darks, 60 bias.  What would you consider for a good ratio for lights/darks/bias.  I use Deep Sky Stacker for the image (and finish up with Photoshop).  Thanks in advance!  Chris

[alacant]

1 hour ago, Chris33benoit said:

lights/darks/bias

2000/0/64 with 16 flat frames. 

HTH

[Chris33benoit]

Thanks alacant, I will tried that!

[AstroRookie]

On 11/09/2020 at 01:33, rnobleeddy said:

Very new to this, but after a frustrating summer with clouds and getting used to equipment, I finally got everything working last week. I've captured short amounts of data on M13, M27 and M38 that aren't going to win any awards, but they look OK and reassure me that everything is working reasonably well and I'm stacking correctly. 

Then there's  M31!

Across 2 nights I have about 1hr20m of data spread across 48 subs consisting of some 4 min and 3 min subs and a lot of 1 min subs, with corresponding darks, flats and bias frames for each night and exposure length. I stacked these as 3 groups in DSS, but I've also tried individual sets of data and leaving out flats/darks/bias in various combinations to rule out issues with any individual aspect. I've taken out the obviously bad subs but a few with star trails appear to have made it to the final image below, and there's probably also a soft focus on at least one night. However, none of this would appear to explain the problem - which is there's almost no data on the outer part of the galaxy. 

Night 1 (about 1/3) of the data had a bright and almost full moon close to M31. Last night was about 60% moon but it wasn't very close at the time I was imaging. Kit is a 130PDS, EQ5, modded Canon 450D with bader cc, and Astronomik CLS CCD. Bortle 5 skies.

A couple of the stacks are below and one of the best individual shots. Whilst some of the stacks with all the data seem to skew towards red, none of the attempts I've made have any real detail in them, presumably because it isn't there.

I then see posts on Reddit like this and wonder what I'm doing wrong!

So wondering if anyone with more experience could advise:

- is this what you'd expect to see given my equipment or is something awry?

- if this is expected, what would have the biggest impact on making the image better? Clearly more subs at longer exposure can't hurt.

- Does the moonlight effectively kill the image? Or light pollution? I can drive a relatively short distance to a darker site but I'd rather work from my garden if at all possible, because that's possible most clear nights, whereas moving the kit is time consuming

- I tend to skimp on dark frames a little (I have 10x for the 1 minute subs, and maybe 4 each for the 3 minute and 4 minute subs) - is this an issue?

- are there any other tips/tricks to pull any more structure out? The images below are auto-scaled with the ASI fits viewer.

 

Stacked using a subset of the data with flats/bias/darks

All data stacked, just lights

One of the 3 minute lights

 

 

 

 

 

 

 

 

 

On 11/09/2020 at 01:33, rnobleeddy said:

Very new to this, but after a frustrating summer with clouds and getting used to equipment, I finally got everything working last week. I've captured short amounts of data on M13, M27 and M38 that aren't going to win any awards, but they look OK and reassure me that everything is working reasonably well and I'm stacking correctly. 

Then there's  M31!

Across 2 nights I have about 1hr20m of data spread across 48 subs consisting of some 4 min and 3 min subs and a lot of 1 min subs, with corresponding darks, flats and bias frames for each night and exposure length. I stacked these as 3 groups in DSS, but I've also tried individual sets of data and leaving out flats/darks/bias in various combinations to rule out issues with any individual aspect. I've taken out the obviously bad subs but a few with star trails appear to have made it to the final image below, and there's probably also a soft focus on at least one night. However, none of this would appear to explain the problem - which is there's almost no data on the outer part of the galaxy. 

Night 1 (about 1/3) of the data had a bright and almost full moon close to M31. Last night was about 60% moon but it wasn't very close at the time I was imaging. Kit is a 130PDS, EQ5, modded Canon 450D with bader cc, and Astronomik CLS CCD. Bortle 5 skies.

A couple of the stacks are below and one of the best individual shots. Whilst some of the stacks with all the data seem to skew towards red, none of the attempts I've made have any real detail in them, presumably because it isn't there.

I then see posts on Reddit like this and wonder what I'm doing wrong!

So wondering if anyone with more experience could advise:

- is this what you'd expect to see given my equipment or is something awry?

- if this is expected, what would have the biggest impact on making the image better? Clearly more subs at longer exposure can't hurt.

- Does the moonlight effectively kill the image? Or light pollution? I can drive a relatively short distance to a darker site but I'd rather work from my garden if at all possible, because that's possible most clear nights, whereas moving the kit is time consuming

- I tend to skimp on dark frames a little (I have 10x for the 1 minute subs, and maybe 4 each for the 3 minute and 4 minute subs) - is this an issue?

- are there any other tips/tricks to pull any more structure out? The images below are auto-scaled with the ASI fits viewer.

 

Stacked using a subset of the data with flats/bias/darks

All data stacked, just lights

One of the 3 minute lights

 

 

 

 

 

 

 

 

Hello,

as for the ideal ISO of a canon eos this info was very helpful to me: https://www.astrophotography.app/EOS.php

Kind regards,

AstroRookie