Another Binning question

[bluesilver]

Hi, first sorry about yet another Binning question, I have the basic understanding but i am just trying to figure out how to best get my processing right.
So basically i have been using the astronomy.tools site to work out my arcseconds per pixel

The figures say that Ideal resolution is 0.67" - 2" arcseconds
With my scope and Camera, ( Scope is a Skywatcher Evostar 150ED and camera is a ZWO ASI2600MC
My resolution comes out as 0.65" arcseconds per pixel and tells me it is Slight Oversampling

Now if i select 2X2 Binning i get a figure of 1.29" arcseconds per pixel and this tells me is good.

So reading and researching a bit more, all the information i am getting is that they are say with a CMOS sensor that you should only bin in processing after you have done the stacking, not before.

So all up i am trying to figure out how to get my image resolution to this 1.29" arcseconds per pixel figure,
Please correct be if i am wrong here, but is this as simple as just clicking on the Bin icon in Startools?
Only reason i am asking is when i click on that i don't get the 2X2 bin option, just the percentages, 25% 50% 70%

I am sure i am overthinking things, i am just trying to perfectly match up the camera to the scope.

Any information would be greatly appreciated.
Thanks.

[Stefan B]

Hi!

According to astronomy.tools your resolution is cut in half if you bin 2x2. StarTools gives you the resolution percentage of the unbinned image. Thus, 50% should correspond to a 2x2 bin. You can also do fractional binning and are not limited to 2x2 or 3x3. One way to determine a suitable amount of binning might be to have a look at the stars in 100% and see when they appear blocky when reducing the resolution. The right amount might be just before they are getting blocky or details in the image appear blurry. Thus you gain SNR without losing detail, I think. At least that's my understanding of the binning module.

Regards
Stefan

PS. I have around 1.4 ''/px with my setup unbinned, which is usually fine with images containing lots of data and fine details. When I only have images with "low" exposure times and a fair amount of noise, I usually use 0.71% binning giving me around 2''/px if I am not mistaken. This helps with SNR and noise reduction during processing without making the image appear blurry.

[admin]

Further to @Stefan B's great explanation, in a majority of cases it is the actual atmosphere ("seeing") that is the limiting factor, in terms of what you can resolve. Indeed, the "blur" test is a good way to visually gauge whether your dataset from a particular night is oversampled.

That said, the SVDecon module is starting to overcome atmospheric (and optical) limitations in resolving power in significant ways now.

[bluesilver]

Thanks for the replies and advice, appreciated.
I wasn't aware that you could do a custom bin other than the 25% , 50% , 75%
I will have to take a bit more of a look at that.

I take it that you still only bin before you start the processing work like Auto Devo?

Thanks again.

[admin]

I take it that you still only bin before you start the processing work like Auto Devo?

Ideally, yes. That's because the dataset will be cleaner and less noisy due to the binning, allowing all algorithms to achieve cleaner/better results as a consequence.

[bluesilver]

Appreciated,
Will have to put all this into practice and see how things go.
Cheers.

[Mike in Rancho]

Interesting about the blur test, I will have to try that out. Another reason the continuously variable binning could be so useful.

However, a couple things I noticed, or maybe did, during some recent processing. Possibly though because I don't know what I'm doing.

When I processed M106, it seemed that doing HDR in full resolution gave me better core detail. I actually ran two passes, one default at only a tick or two above the 1.2 starting strength, which set the stage for a reveal core with a strong setting. Then I binned for the remainder of processing and the detail survived the scaling down. Any other order of doing it did not give me the same core detail.

And just now doing the Fireworks Galaxy, I kept a high resolution of about 4000x2500 (71% bin), and I seemed to have better stars for SVD in the sampling selection phase. I did have to increase the blue box size a few notches, of course. When I had tried it earlier following a bigger bin down, the good stars seemed harder to come by. That said, all things may not have been equal (difficult as ST is so often auto-sensing the state of things and acting accordingly), so I'm not sure this is a scientific finding. Will just have to keep trying it out and see if patterns arise.

[admin]

it seemed that doing HDR in full resolution gave me better core detail.

Though we're pretty spoiled in StarTools for predictable behaviour no matter the characteristics of the dataset, some modules behave differently depending on the resolved detail.

The HDR module is such a module - you may have to tweak the Detail Size Range for it to allow the module to start affecting areas of high frequency detail.

Hope that helps/make sense!

[Mike in Rancho]

Though we're pretty spoiled in StarTools for predictable behaviour no matter the characteristics of the dataset, some modules behave differently depending on the resolved detail.

The HDR module is such a module - you may have to tweak the Detail Size Range for it to allow the module to start affecting areas of high frequency detail.

Hope that helps/make sense!

Yes, it now clicks. I've only started altering the detail size range and noting the effects within the last few images. So back then I likely just made up a convoluted, multi-step workflow to achieve results I could have obtained just by setting the parameters correctly.