M101

[Rodd]

This is 40 300 sec red subs using TOA 130 and asi 1600.  Due to a host of reasons I am working on several projects at once.  None f them are likley to be completed any time soon.  For over 3 hours of data i expected the background to be a bit less terrible.  I have spent far to long battling it in this stack---the reason it is passable at normal vieiwng.  Full resolution revels the illusion.

I am pleased with the disc and probably will finish this image before the others I started simply becuase with the Moon now threatening I can collect a lot of Ha for M101 while M104 and NGC 3953 don't have as much Ha.  Its amazing to e how the tiniest of mifferences in placement in the FOV can radically change an image.  I centered the galaxy, not realizing that much of t would extend upward tword the top edge.  This is a bit of a crop to center.  I have been in the habit of collecting 5 hours each of RGB for my images, thinking that this was sufficient.  I am no longer confident of this.  I am remembering why I was trying to find a faster system......I believe I will eventaully come up with a decent HaLRGB image.  Its the eventually that is killing me

 

 

[wimvb]

Looks like a good start already. 

[Rodd]

30 minutes ago, wimvb said:

Looks like a good start already. 

Thanks Wim.  I am pleased with the resolution—the details in the disc. The background is my normal mess.  Would that I could say that it will improve with the addition of data.   Sadly, that is not how it typically goes for me.   Maybe THIS time huh?  I just wish it wouldn’t take a month to find out 

[gorann]

Very promising I would say. If you worry about framing you could just crop off a bit at the bottom.

[Laurin Dave]

Nice one Rodd, M101 is low surface brightness so to get the outer arms to  show requires a big stretch ..  hence the background...  I got 12 hrs lum last year on it which ended up ok..    the Ha is worth getting as there are loads of Hii regions in the arms ...  all the bright bits that show on your lum are Hii

Dave

Edited March 1, 2020 by Laurin Dave

[Rodd]

1 hour ago, gorann said:

Very promising I would say. If you worry about framing you could just crop off a bit at the bottom.

Thanks Gorann.  Yes, I cropped it more than I wanted.  Not too bad though. 

[wimvb]

5 hours ago, Rodd said:

For over 3 hours of data i expected the background to be a bit less terrible

I have found that with my ASI174MM-COOL, and MN190 (f/5.3), I need at least 3 - 4 hours of luminance to get a decent background, with 120 s subs. Ie, at least some 100 - 120 subs. Colour is more forgiving, but 40 subs (@240 s)/channel is a minimum. Colour can always be blurred before lrgb combination to suppress colour noise. But all this is very individual. The total integration time that you need not only depends on optics, but also on sky quality. 

[Rodd]

49 minutes ago, Laurin Dave said:

Nice one Rodd, M101 is low surface brightness so to get the outer arms to  show requires a big stretch ..  hence the background...  I got 12 hrs lum last year on it which ended up ok..    the Ha is worth getting as there are loads of Hii regions in the arms ...  all the bright bits that show on your lum are Hii

Dave

Thanks Dave.  I know.  But it seems no matter how much time I put into an image I can only stretch it a tiny bit before the background shows nasties.  You’d think that 24 hours would be enough.  Not for me.  

[Rodd]

1 minute ago, wimvb said:

I have found that with my ASI174MM-COOL, and MN190 (f/5.3), I need at least 3 - 4 hours of luminance to get a decent background, with 120 s subs. Ie, at least some 100 - 120 subs. Colour is more forgiving, but 40 subs (@240 s)/channel is a minimum. Colour can always be blurred before lrgb combination to suppress colour noise. But all this is very individual. The total integration time that you need not only depends on optics, but also on sky quality. 

And, it seems, processing skill!🙂. This image has 40 300 sec subs and it is nowhere near enough.  Not even half.   And, I might add, conditions were above average.  The entire stack has a Fwhm of 2.3.  Some subs came in at 1.8.   For me that is like a booyah moment!  

[Laurin Dave]

Hi Rodd,  maybe it’s your flats that are the issue.  I recently changed from using an led tracing pad to an ElectroLuminescent panel and my flats now work much better giving a much smoother background particularly on lum 

Dave

[david_taurus83]

The background does look very noisy. Dont think I ever noticed it like that on my ASI1600. Theres always a good iteration of MLT on the shadows?

[Rodd]

4 hours ago, Laurin Dave said:

Hi Rodd,  maybe it’s your flats that are the issue.  I recently changed from using an led tracing pad to an ElectroLuminescent panel and my flats now work much better giving a much smoother background particularly on lum 

Dave

I use one too.  I don’t think the pixel scale issues are from flats. I could be wrong 

 

[Rodd]

5 hours ago, david_taurus83 said:

The background does look very noisy. Dont think I ever noticed it like that on my ASI1600. Theres always a good iteration of MLT on the shadows?

Noise suppression of any kind doesn’t solve it.   There is a pixel math formula for mottled backgrounds that is the best thing I have found.  But what it does is brings all pixels up to a certain brightness, then you can reduce them all together.   But this is just too bad for even that.....which I used. More data is the only solution that I know

Edited March 2, 2020 by Rodd

[wornish]

Have a go with the latest Topaz DeNoise AI on your full size image .  I have found it works surprisingly well.

 

 

[Rodd]

4 minutes ago, wornish said:

Have a go with the latest Topaz DeNoise AI on your full size image .  I have found it works surprisingly well.

 

 

Is that with the latest PI release.  I think I have to pay for the update now, right?  Its been a couple of years since I got PI. 

Rodd

[wornish]

5 minutes ago, Rodd said:

Is that with the latest PI release.  I think I have to pay for the update now, right?  Its been a couple of years since I got PI. 

Rodd

No unfortunately not.  I runs it as a standalone app, or as a PS plugin.

[Rodd]

Oh well

[Laurin Dave]

If this was mine I'd suspect the calibration files, might be worth stacking the lights registered but a) uncalibrated, b) calibrated with just darks and c) calibrated with just flats to see whether noise is being introduced 

[Rodd]

53 minutes ago, Laurin Dave said:

If this was mine I'd suspect the calibration files, might be worth stacking the lights registered but a) uncalibrated, b) calibrated with just darks and c) calibrated with just flats to see whether noise is being introduced 

I'll give it a shot.  But I have tried it in the past when my filter wheel does not align and the dust bunnies are not completely removed.  In those cases, the only difference is the dust bunnies.  Flats wont take out light pollution, or the effect of clouds or haze.   My sky is bad.   I use a flatman panel--if that doesn't work, no sense doing this anymore.

[Rodd]

2 hours ago, Laurin Dave said:

If this was mine I'd suspect the calibration files, might be worth stacking the lights registered but a) uncalibrated, b) calibrated with just darks and c) calibrated with just flats to see whether noise is being introduced 

Dave--lets assume for a minute that the flats are the culprit.  What can be done?   For broadband flats I use 50-100  and they are either 1.3, 2.2, 2.8 or 3.2 second long, depending on filter.  I use flat darks of the same duration.  In Pixinsight out of 1 (1 being 65,500), I try for flats between .35 and .45.  

Any ideas?

[Laurin Dave]

What you are doing all sounds good Rodd and the same as I do ..  so assuming the EL panel is ok I'm not really sure, the only thing I could think would be to try sky flats.   Do the integrations from the other filters have similarly noisy background?   

[Rodd]

1 minute ago, Laurin Dave said:

What you are doing all sounds good Rodd and the same as I do ..  so assuming the EL panel is ok I'm not really sure, the only thing I could think would be to try sky flats.   Do the integrations from the other filters have similarly noisy background?   

Yes--I have been complaining about this for years.  There have been numerous threads of me trying to get to the bottom of the "speckled" background as I call it.  A lot of data makes it disappear--as expected.  But it takes so much, I am not sure I can go on like this.  I don't mind capturing 25 hours of data for an image--but when that 25 hours takes a month and a half to collect, it gets old real quick.  That's why I got the FSQ 106 and .6x reducer for the ASI 1600.  Like a fool I stepped right into the reducer trap (a.k.a.myth) thinking that F3 on the FSQ 106 would be like an Epsilon, or other F3 system.   I thought it would reduce my exposure times to 4-8 hours--possible to do in 1-2 nights.   Now, of course, I understand that without a corresponding increase in aperture, a reduction in focal ratio is not  the golden road to blissful speed.  Does it help?  Most would say yes.....I say....maybe a  bit.  For a real increase in speed you need more photons.  Without getting into the whole debate about SNR of the whole frame and extended objects  and on and on, I will say....it did not solve my problem.  I still need 15-20 hours minimum.  So 4-5 weeks instead of 6-8.  

I am sort of out of ideas.

[gorann]

Although I don't think your background is very noisy, it is a bit noisy for being 3 hours, and I expect it has to do with your light pollution, which you complain about all the time - and still end up with so many great images.

Three solutions: (1) more data, (2) settle for what you have, or (3) sell your house and move to a darker place😉

Or sell your 5" Tak and buy a 6"Esprit - 33 % more photons and cheaper

Edited March 2, 2020 by gorann

[Rodd]

Just now, gorann said:

Although I don't think your background is very noisy, it is a bit noisy for being 3 hours, and I expect it has to do with your light pollution, which you complain about all the time - and still end up with so many great images.

Three solutions: (1) more data, (2) settle for what you have, or (3) sell your house and move to a darker place😉

Or sell your Tak and buy an Esprit

Very clever.....though the last suggestion has me a bit stumped.    I could throw a reducer on the TOA and get down to F5.38...which is respectable, more in line with the Esprit, if that is what you mean.  But then I lose scale, and for galaxies, that is important.   And, and this is key, it doesn't really matter what scope I use or at what focal ratio I image....my background looks the same.  So if you mean sell the TOA 130 and get an Esprit 150--I don't think that would help with the background.

I would like a Esprit 150 though

[gorann]

Now you are falling into the deep hollow of the f-ratio myth. A 6" refractor collects 33% more photons than a 5" one, whatever reducer you throw at it (and you can throw one on the 6" also). It still has a bigger front lens!

Edited March 2, 2020 by gorann