Ongoing noise problems

Gavin1234 · #1 09-10-2018, 01:04 PM

I started imaging sculpture and Orion last night. Orion is turning out reasonably well but sculpture is giving me grief.

Some people would have seen my posts when I was having problems with Helix. It seems like I’m having the same issues with sculpture. Basically way too much noise. I’m not getting anywhere near as much with Orion.

I’m thinking it might be because both helix and sculpture are similar in that they’re feint objects in the middle of a dark background? I don’t have the same issue with Orion or horsehead or m20 etc.

I’m hoping you can see in the attached pictures. Both are stacked in PI with the same flats, darks and bias. Orion has 18 x 5 min lights and sculpture has 26 x 5 min lights. Both given exactly the same treatment in PI just to make the visible, auto background extraction, scnr and a very quick and rough stretch. Orion gets better from this point whereas sculpture just gets more and more noise from here on regardless of what I do.

RyanJones · #2 09-10-2018, 01:13 PM

Hi Gavin,

I saw your noise issue on helix but I left it to others to comment. As for Sculptor, I have the same issue. I was expecting it though. I had some self imposed tracking issues so out of a planned 3-4 hr intergration, I got 1 1/2 hrs, pretty much the same as you. I know in my light polluted skies, anything upwards of magnitude 6, I need a good 500 subs ( mine are 22sec) to get anything with half decent SNR. Sculptor is in an easy spot and is up at a good time so my advice would be double the subs and see what happens.

Atmos (Colin) · #3 09-10-2018, 02:00 PM

I would suggest that what you’re running into is limited exposure and depending on your location, light pollution.
Even under dark skies, two hours of integration isn’t a lot and it becomes even more problematic when imaging under light pollution.

My suggestion would be to add another 5 hours of data to NGC 253 and see if you get an improvement

You’re on the right track so far though.

xelasnave · #4 09-10-2018, 02:19 PM

Hi Gavin
Have you tried dithering?
I think the idea is to move the camera frame a little between each capture say by the width of a couple of stars in order to "distribute" the noise.

Also I dont know if what I suggest next is ok but mask your main object and stars and darken the background...I imagine there will be better ways we will discover later but that may hide the noise that annoys you☺
In any event try to focus somewhat on everything that is going great for you and dont let a little noise worry you...I suppose if you have the object masked you could experiment with blur as well.

Anyways keep on turning out these great images.
Alex

Gavin1234 · #5 09-10-2018, 06:39 PM

Thanks guys, hopefully it is just a lack of exposure time. It’s just that it’s fairly similar to the noise I had with helix which persisted even with 30 hours. It’s also strange that it isn’t there to anywhere near the same extent on the nebula images.

The path sculpture is taking is pretty much perfect for me, its in range from sunset and goes straight over the top of me. It starts off close to the direction of Sydney but I’d be surprised if it’s light pollution.

Anyway I’m going to just focus on Orion and sculpture for while to build up lots of data. Hopefully they turn out well.

Thanks Alex, I dither after each frame. My posts probably seem a bit negative but I’m really enjoying this new hobby. I just tend to post about the parts I’m struggling with. I’m just looking into printing a large copy of my horsehead image onto acrylic at the moment.

CalvinKlein (Kelvin) · #6 10-10-2018, 09:28 AM

Good images Gavin. Looks to me like your subs are way too long. I'd suggest taking many more but at shorter duration (something I wish I learned a year ago). All your stars appear saturated which is a tell-tale sign. 5 min is a long time unless your target is in very un-light polluted region of sky. Instead of 26 x 5 try 130 x 1min. You only need to expose each frame so that your background sky level is 10 to 20 times your read noise. As others have said you ideally need more integration time too. to increase your signal to noise ratio. Have you read the CloudyNights forum threads on the optimum exposure settings for CMOS cameras ? Very worthwhile studying. (oh and colour cameras need MUCH more exposure time than mono cameras unless the sky is very, very dark where you are imaging)

RyanJones · #7 10-10-2018, 09:54 AM

Quote:

Originally Posted by CalvinKlein

Good images Gavin. Looks to me like your subs are way too long. I'd suggest taking many more but at shorter duration (something I wish I learned a year ago). All your stars appear saturated which is a tell-tale sign. 5 min is a long time unless your target is in very un-light polluted region of sky. Instead of 26 x 5 try 130 x 1min. You only need to expose each frame so that your background sky level is 10 to 20 times your read noise. As others have said you ideally need more integration time too. to increase your signal to noise ratio. Have you read the CloudyNights forum threads on the optimum exposure settings for CMOS cameras ? Very worthwhile studying. (oh and colour cameras need MUCH more exposure time than mono cameras unless the sky is very, very dark where you are imaging)

That's really intersting to hear Kalvin. I think I'll also have a look at that CN thread but what you're saying makes a lot of sense. I have persevered with my unsmooth mount that can only give me 25 odd second frames consistently basically for that reason with light pollution. Sometimes I take long ( 1.5 min ) exposures with star movement just to check the framing of my target but the sky glow is horrible. I learned over time that I could get some satisfying images with 25 second subs, I just need to take a bare minimum of 500 of them even on mag 6 objects. I'm now aiming at mag 8 targets so I imagine I'll need maybe 7-800 which I'm fine with. Jeez this game is fun

ChrisV (Chris) · #8 10-10-2018, 09:54 AM

I would agree with Kelvin - assuming this is with the asi071. In my outer suburban Sydney skies subs of 90-120sec are way more than enough to swamp read noise. At 5 min you are just going to start saturating things. Shiraz had a good thread and a spreadsheet to help figure it all out for this camera

Atmos (Colin) · #9 10-10-2018, 10:07 AM

I use 300s exposures at ISO 400 with my Nikon D810 with a F/5.2 refractor and find that works perfectly on all except for the brightest of parts.

Burnt Orion core with 300s exposures but your shot is burnt out a bit more than what mine is so dropping it to 180s might be better?

CalvinKlein (Kelvin) · #10 10-10-2018, 02:16 PM

There's a great thread on Cloudy Nights titled "Astrophotgraphy in Light Pollution" - google that and it will come up in the first few search results. It's incredible what some photographers are doing with just very cheap DSLR's (though they are modified) in heavily light polluted zones. In particular look at posts from Jon Rista and jgraham who both image from heavy light pollution.

Oh and with regards the blown out Orion core just take a few short exposures of different lengths (usually only a few seconds) and blend them into your final image with a layer mask in your preferred imaging editor. Very simple to do and very effective. (lots of youtube tutorials on it). It's a perfectly acceptable technique.