first attempt at long exposures - Noisey and blury

[vertigo262]

I have been trying to figure out how to get long exposures for a while, and it turned out, as embarrassing as it is. That I thought the Star sense guided.

After setting up a guide scope and learning phd, I went out tonight for my first attempt.

i could swear, I had extremely good focus, and even the single frames seem in focus. But after putting them togethor with DSS, they look horrible. One thing to note is that I didn't do any calibration frames for these, however I did use the PHD dithering in BYE. But it looks horrible and I am trying to figure out why. it is extremely blurry, and horrible noise. I don't know if the calibration frames would fix this bad of noise.

My first attempt without a guide scope. 1600 iSO, 100 x 2 min exposures and calibration frames. 

this is 3 x 10  minute exposures  IS0 800, without calibration and dithering on.

single frame from tonight

Any advice as to what I am doing wrong would be very helpful

[glowingturnip]

I would say firstly, you've got quite a few dustbunnies there, even on the single frame, so you should do flats to take care of those.

It looks to me as though the stacking has been done without any sigma clipping ?  Your dustbunnies seem to have multiplied  (insert joke about proclivities of rabbits), and you've got noisy streaks that shouldn't be there after stacking. 

Sigma clipping will find an average and standard deviation of a pixel in a given location, reject any outliers, and average the remainder for the stack.  Each dustbunny should occur in the same place in only a handful of frames, especially after dithering, so should get rejected out of the stacking.  I think from memory DSS has kappa sigma stacking, or an adaptve something sigma stacking algorithm - try those.

edit: just seen that second image was only 3 lights ?  sigma clipping will need several more frames than that I'm afraid - more data needed

[steppenwolf]

I wouldn't beat myself up too much over that image - 2 minutes is a very short exposure for the Horsehead Nebula so even your multiple frames will require severe stretching to release the detail but this will also make the image appear noisy despite the positive effects on signal to noise ratio of stacking your 100 frames. There is no substitute for longer exposures and there's the rub, you need to get autoguiding running to achieve those exposures.

Even with good focus, the tracking errors will tend accumulate and bloat the stars as you stack them as you will not actually be stacking 'pinpoints' and, of course, that savage stretching will also bloat the stars as they become 'over-stretched'. You can process the stars separately to obviate some of this though.

You could do with attacking the source of the dust motes in your system -  you have a pretty comprehensive collection including a hair fragment so a good blow with a bulb blower (brush removed and discarded!).

Your light pollution looks pretty bad so you could try to attack that with a light pollution filter - again, relying on removing it purely in post processing will add to the noise.

For just 2 minute exposures, I think you have done remarkably well - just think was longer tracked exposures and LP defences could achieve for you!

[vertigo262]

Hi Steve and Stuart.  I will try sigma clipping. I don't know what that is but will research it. I never heard that term, Dust Bunnies, that is funny. I keep cleaning my cmos but they keep coming back for some reason.

Staurt, I was comparing my 2 minute to 10 minute exposures and my 2 minute ones look way better stacked. 

But why does it look so blury, and the single frame looks fairly clear. 

[vertigo262]

This is with the Sigma Kappa at 2 in DSS. the Data seems a little more usable, but still seems really blurry to me. My focus seemed dead on, but maybe it was off. I could do a little more with this data.

[glowingturnip]

not sure if you saw my edit to my first post, sigma clipping won't work with just three frames I'm afraid, the algorithm can't tell what is real signal and what is outlier, so just averages them.  Needs at least about 12 to start working well.

I don't think it's blurry or out of focus as such, if you look at the smaller stars, they look ok, but you have got a lot of noise going on, and the only cure is going to be more data.  Your first pic, the unguided one, has a total of 200 minutes of data, this one only has 30 minutes.

[vertigo262]

so then, if I am doing 10 minute exposures, how many would you suggest? I was pulling a number out of the air, but I was thinking about 18 frames, 3 hours. although it started to get windy, so I had to pack up and I got 6, and half are bad because the wind. But I was assuming because the exposures were 10 minutes, I woudn't need nearly as many because the amount of data in 1 exposure should be a lot more

In this, I am using the firsst 3 out of 6 before the winds started

[glowingturnip]

18 frames, 3hrs, sounds decent to me

[dph1nm]

 But I was assuming because the exposures were 10 minutes, I woudn't need nearly as many because the amount of data in 1 exposure should be a lot more

You would need the same total exposure with 2min or 10min subs, apart from the effect of the extra read-noise with the shorter subs. This extra read noise means that you can get away with less total exposure with the longer subs, but only if read noise is a significant contributor to the 2min subs.  If it isn't, then you won't see any advantage from longer subs.

NigelM

[vertigo262]

NigelM,

but then the question is.

100 x 2 min = 200 mins

20 x 10 mins = 200 mins

they come out to the same in time. But I was under the assumption that because I am exposing for a longer period of time, i have more data in the 10 min exposures. Because 2 minutes doesn't get a lot of information

So the 10 minute exposures should look a lot better

at least that is what i am assuming

[vertigo262]

This was interesting.

I went out tonight, and made sure the exposures were right. Because they were so short, I decided to go down to ISO 200. 

I took 22 x 400 second shots.

This is what came from it. I haven't heard of anyone going down to that low of an ISO, but I thoght I'd give it a try.

I think since this isn't over exposed like the last one, is why it is so much better. I'm thinking I want to go out again and get more frames for this shot for clarity

also, I went out and got a uv ir cut filter. solved the star bloating issue

[Starlight 1]

I have put some post in on  trying low ISO I got down to 900s @ 100 and 1200s@ at 100 data was down but made a better pic .

[vertigo262]

basic photography, the lower the iso, the less noise, but I'm not sure if that applies to astro photography 

[Starlight 1]

Yes as you said,  I was thinking back to the old days of 35mm on a sunny day hand hold I was useing 25 and at night 64 , 25 in the dark room I blow that photo up to wall size with 6400 post card. So I see it its not hand hold  wee are guiding so with darks ect to remove noisy ,so Iso 100 is going to lower them 6400 taking time up wee back to the start in noisy .

But some how it work for me doing low ISO better photon .

If you are lost in this post so am I.

[sharkmelley]

It would be useful to have some information on the camera and scope combination you are using and whether or not the camera is modified. The single frame you posted seems to indicate you have a fair amount of light pollution (the Flame Nebula is only just visible above the background) - this light pollution, even after background subtraction, will be the main source of noise in the final image.  If you are not already using a light pollution filter you will find it helps considerably.  The background streaking is typical of Canon cameras (so I guess you are using Canon!) and sigma stacking will not remove it.  It is most noticeable when a strong light pollution signal has been subtracted from the image - another good reason to use a light pollution filter.

In general, low ISOs will have more read noise than higher ISOs but with the strong light pollution you are seeing, the read noise will be swamped by noise from the light pollution and it really won't matter what ISO you use.

Mark

[dph1nm]

they come out to the same in time. But I was under the assumption that because I am exposing for a longer period of time, i have more data in the 10 min exposures. Because 2 minutes doesn't get a lot of information

So the 10 minute exposures should look a lot better

No - you get exactly the same number of photons from your DSO in 2min x 100 as 20min x 10 - the only difference is 100 doses of  read-noise from one and only 10 from the other (which means a sqrt(10) difference in read-noise contribution to the final image).

NigelM

[vertigo262]

It would be useful to have some information on the camera and scope combination you are using and whether or not the camera is modified. The single frame you posted seems to indicate you have a fair amount of light pollution

this image is 20 x 400 seconds at ISO 200

My setup is:

Explore Scientific ED 127mm 7.5

Canon 7D converted to full spectrum

Celestron Advanced vx mount

a uv cut ir filter. no other filters