#1  
Old 13-03-2019, 04:52 PM
The_bluester's Avatar
The_bluester (Paul)
Registered User

The_bluester is offline
 
Join Date: Feb 2011
Location: Kilmore, Australia
Posts: 3,342
Effectiveness of drizzle integration

A quick bit of background, I am currently imaging at quite a small image scale, using a ZWO ASI294 (4.63 micron pixels) on a C925 with a focal reducer with the spacing arranged to give me about 0.68 reduction ratio for a reasonable compromise between field of view, vignetting, coma and speed.

I am looking at branching into much more widefield stuff, still a the budget end of the spectrum. Something like the Skywatcher Evostar ED72 might fit the bill and most calculators put that in the "Slight undersampling" category with my camera.

For similar money though there is the likes of the William Optics Redcat 51mm, but that pushes into "Significant undersamplling" territory and would be around 3.8 arcseconds per pixel.

Just how much data can be retrieved via drizzle integration if images are well dithered? Amd I likely to end up with a disappointing result with that sort of combination? A second camera is not on the cards.

While it costs a lot more, the Celestron 8" RASA almost looks like it was made with my camera in mind and at F2 imaging would be seriously fast! But with how tiny the critical focus zone is going to be, I would not even consider that without a focus motor with very fine steps so focus could be automated and run regularly as temperature changed, the SCT is bad enough for focus changes as it is and I have seen talk of refocusing the RASA scopes as often as every 15 minutes to keep them sharp if the temperature is changing.

That is by the by, my real question is how undersampled can you go and still extract good information with drizzle?
Reply With Quote
  #2  
Old 13-03-2019, 06:27 PM
NorthernLight's Avatar
NorthernLight (Max)
Settled

NorthernLight is offline
 
Join Date: Sep 2008
Location: Auckland, NZ
Posts: 343
Hey Paul,

I haven’t run the numbers but I think you’re mixing over and under sampling up. The pixel dimensions parred with a 460 ish fl results in quite a bit of oversampling. No need to recover any data with drizzle. It’s all there and each star should occupy quite a bunch of pixels.

Cheers
Reply With Quote
  #3  
Old 13-03-2019, 06:45 PM
The_bluester's Avatar
The_bluester (Paul)
Registered User

The_bluester is offline
 
Join Date: Feb 2011
Location: Kilmore, Australia
Posts: 3,342
Bintels calculator calls my C925 and reducer (About 1500mm FL) slightly over sampled, the ED72 slightly under sampled and the Redcat significantly undersampled. Though I am not sure what sort of seeing they base that on.
Reply With Quote
  #4  
Old 13-03-2019, 08:45 PM
gb44 (Glenn)
Registered User

gb44 is offline
 
Join Date: Jul 2012
Location: Gold Coast
Posts: 273
Its got nothing to do with seeing.
It relates to the number of pixels that record light of any particular star.

Craig Stark has explained it here
http://www.stark-labs.com/craig/reso...rizzle_API.pdf

GlennB
Reply With Quote
  #5  
Old 13-03-2019, 11:03 PM
Camelopardalis's Avatar
Camelopardalis (Dunk)
Drifting from the pole

Camelopardalis is offline
 
Join Date: Feb 2013
Location: Brisbane
Posts: 5,425
The end result has everything to do with seeing, as it’s the turbulence in the atmosphere that convolves the airy disc and bloats the stars in our images.

With a scope of around 500mm focal length, I think you’d be pleasantly surprised at just how much detail you can pick up, before trawling with drizzle. The sub-pixel information leveraged in drizzle is not going to give you a 2x increase in detail...just a bigger image.
Reply With Quote
  #6  
Old 14-03-2019, 05:56 AM
The_bluester's Avatar
The_bluester (Paul)
Registered User

The_bluester is offline
 
Join Date: Feb 2011
Location: Kilmore, Australia
Posts: 3,342
Well, the PDF by Craig Stark is interesting, it suggests you might be able to extract data to something like half the arc seconds per pixel that is in the initial image, assuming it is under sampled to begin with.

That would take the 3.8 arc seconds I would see out of the Redcat and extract data more like I could expect to see at 2 arc seconds/pix, which sounds pretty interesting. I already dither between images so that is a non issue.

Unfortunately it does not make a choice between the redcat and ED72 any easier to know that tricky image processing could probably bring the level of detail recorded by the redcat up near that I would see out of the 72mm! But fitting the whole SMC in one frame would be pretty nice.

Last edited by The_bluester; 14-03-2019 at 08:07 AM.
Reply With Quote
  #7  
Old 14-03-2019, 07:40 AM
Atmos's Avatar
Atmos (Colin)
Ultimate Noob

Atmos is offline
 
Join Date: Aug 2011
Location: Melbourne
Posts: 6,980
My experience is that it is all to do with the FWHM in your image rather than image scale. For instance, when shooting M16 in Ha with a QHY163M and 130mm refractor I get an image scale of 1.16”/pixel and under good seeing I was getting FWHM values of 1.7 pixels in the stack from that night.

Did drizzle integrating help? Yes. What it does is enlarge the image in a way that approximates the values of the missing pixels. It doesn’t actually show new information that isn’t in the original non-drizzle stack but it enlarges it and makes it easier to see and process; especially with later deconvolution.

With a OSC you don’t get extra resolution as much as you get better colour rendition. I have bayered images with FWHM values as low as 1.5 pixels but once it is debayered and interpolated that rises to 2.2-2.6 pixels. Bayer Drizzle doesn’t increase resolution in the same way that you can with a mono camera because images still have to be interpolated but you do get better more accurate colour and better PSFs on stars as drizzled OSC images are still going to have FWHMs of 4+ pixels as a minimum.
Reply With Quote
  #8  
Old 14-03-2019, 08:17 AM
The_bluester's Avatar
The_bluester (Paul)
Registered User

The_bluester is offline
 
Join Date: Feb 2011
Location: Kilmore, Australia
Posts: 3,342
I suppose this is really morphing into a discussion on if a cam/OTA combination producing about 3.8 acrcseconds/pix is going to produce nice results or disappointing ones in reasonable seeing with the processing tricks that are reasonably available.

I could cobble stuff together to test it with my old 350D camera as it would produce a similar image scale but given it does not do bulb over USB and does not talk to Windows 10 (Or even Win7 in my experience, I only have 64bit versions of that) it would be a lot of work to get meaningful length subs and darks, and dithering. I would more or less have to have two PC's running the show plus my mount itself (For the SNAP port to get subs longer than 30 seconds) and manually commanding a dither after every sub! with one PC running the mount and guiding through my SCT with the cam piggybacked.
Reply With Quote
  #9  
Old 14-03-2019, 08:28 AM
multiweb's Avatar
multiweb (Marc)
ze frogginator

multiweb is offline
 
Join Date: Oct 2007
Location: Sydney
Posts: 22,060
I have an image scale of 3.04asp with my old Q (530mm FL) and QHY8 and 2.10asp with my QHY9. In both cases drizzle integration makes a noticeable difference especially with the QHY8 as the bayer matrix just adds to the sampling issue. At 3.8asp I reckon dithering and drizzle integrating will show a significant improvement if your guiding is tight enough.
Reply With Quote
  #10  
Old 14-03-2019, 08:32 AM
Atmos's Avatar
Atmos (Colin)
Ultimate Noob

Atmos is offline
 
Join Date: Aug 2011
Location: Melbourne
Posts: 6,980
3.8”/pixel can produce some exceptional results. A FSQ106 with KAF-16803 works at 3.5”/pixel and it is a very good performing system. At that image scale you’ll rarely if ever had seeing conditions where you won’t be under sampling. Seeing can get pretty bad but not that bad.
Reply With Quote
  #11  
Old 14-03-2019, 08:36 AM
The_bluester's Avatar
The_bluester (Paul)
Registered User

The_bluester is offline
 
Join Date: Feb 2011
Location: Kilmore, Australia
Posts: 3,342
Lol, you guys are not making the decision any easier!

It would be easier if there was a significant price difference between the two approaches I am looking at. It might come down to if the backfocus of one of them does not allow the use of my OAG, but even that looks like it is likely to be OK in both cases. 55mm from the corrector on the ED72 is just about perfect, and the Redcat looks pretty close to the mark as well. The redcat even has the right thread at the rear to screw straight to my OAG while removing an adapter I have to use now!
Reply With Quote
  #12  
Old 14-03-2019, 04:51 PM
Camelopardalis's Avatar
Camelopardalis (Dunk)
Drifting from the pole

Camelopardalis is offline
 
Join Date: Feb 2013
Location: Brisbane
Posts: 5,425
I think the moral of the story is...don’t equate resolution with a great image

The ED72 and RedCat are different enough in focal length that you should consider which targets you would go after with each and let that steer you...just my opinion of course
Reply With Quote
  #13  
Old 14-03-2019, 05:14 PM
The_bluester's Avatar
The_bluester (Paul)
Registered User

The_bluester is offline
 
Join Date: Feb 2011
Location: Kilmore, Australia
Posts: 3,342
Yep, that is the problem, two scopes with similar effective price points but different image scales, both will appealing points and limitations in terms of targets. Nothing I have read here makes the choice any easier. If one or the other was a significantly faster ratio then that would be a bit of a pull factor.

Last edited by The_bluester; 14-03-2019 at 05:45 PM.
Reply With Quote
  #14  
Old 15-03-2019, 08:18 AM
Camelopardalis's Avatar
Camelopardalis (Dunk)
Drifting from the pole

Camelopardalis is offline
 
Join Date: Feb 2013
Location: Brisbane
Posts: 5,425
How about...the RedCat looks prettier
Reply With Quote
  #15  
Old 15-03-2019, 09:12 AM
The_bluester's Avatar
The_bluester (Paul)
Registered User

The_bluester is offline
 
Join Date: Feb 2011
Location: Kilmore, Australia
Posts: 3,342
It does!

I need to spend a few hours cruising around in Stellarium to have a look at suitable sort of targets at both image scales and see if I can decide what appeals more.
Reply With Quote
  #16  
Old 19-03-2019, 10:29 AM
avandonk's Avatar
avandonk
avandonk

avandonk is offline
 
Join Date: Aug 2005
Location: Melbourne
Posts: 4,786
The mathematics are quite straight forward.

If you are undersampling and your optics resolution is better than what you are recording then drizzling and many dithered images will get most of it back.

I have an FLI1803 behind my RH200 and even at native resolution of 9 micron pixels I am undersampling.

Here is a comparison of drizzled vs the original x2 binned. The drizzled image is very close to unbinned x1 image as far as star sizes etc but the faint nebular detail is far better.

https://www.flickr.com/photos/343863...n/photostream/

Hope this helps.
Reply With Quote
Reply

Bookmarks

Thread Tools
Rate This Thread
Rate This Thread:

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +10. The time is now 08:38 AM.

Powered by vBulletin Version 3.8.7 | Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Advertisement
Bintel
Advertisement
Testar
Advertisement