Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

High Resolution Imaging - Is This Real?

astrophotography CMOS EAA equipment imaging optics SCT solar
  • Please log in to reply
11 replies to this topic

#1 MeteorBoy

MeteorBoy

    Sputnik

  • -----
  • topic starter
  • Posts: 32
  • Joined: 19 Jul 2016
  • Loc: Toronto, Canada

Posted 10 September 2020 - 12:17 PM

I need some help please.  I’ve just processed a solar photo that seems to produce a high resolution that seems quite unlikely.

Here’s the photo that I’m having some trouble believing.

 

Please note the very small sunspot group at the bottom of the arrow, then the enlargement (the posted photo is not at full resolution).

The photo in question yields an apparent image resolution of ~0.15 arc-seconds. That seems most excessive.  Dawes’ limit for my ‘scope is 0.57 arc-seconds.  To make it even less believable the photo was (carefully) shot through an open window (my balcony didn’t face the sun at the time).

 

Equipment Used

  • ZWO ASI178mm camera (2.4 microns).
  • 20 cm SCT ‘scope operating at F5.6 resulting in an effective focal length of 5.5 meters (including the camera).

The nearly whole solar image was the result of stacking 540 frames then drizzled at 3x and sharpened several times.  The insert enlargement was shot separately using the same equipment but used 14-bit ROI and then selecting the best 10% of 3,400 frames. 

 

The insert was heavily processed by…

  • Drizzling at 3x, then re-sampling upwards by 3x more.
  • 2-3 layers of sharpening tools separated by light blurring.
  • Applying a variation of HDR to expand the dynamic range.

My questions are…

  1. I calculated my image resolution manually by taking the fraction of the image that is occupied by the smallest feature that I can resolve and then multiplying that fraction by the photo’s dimension in arc-seconds. I know this is not the same as resolving a double star used for calculating Dawes’ limit as it’s subjective.  Is this sound?
  2. Is there a web page somewhere where I can upload a photo and have it measure the photo’s image resolution (not plate-scale)?
  3. I processed the insert photo for maximum resolution but it comes at the expense of a less natural rendition.  Might there a way where I can do both?

I’d appreciate some enlightenment please.

 

Aug 18 2020 - Stacked and Processed (CN).jpg



#2 Xyrus

Xyrus

    Mariner 2

  • -----
  • Posts: 215
  • Joined: 20 Jun 2017

Posted 10 September 2020 - 12:29 PM

That's not resolution you're seeing. It's artifacting caused by over processing. You can easily tell by looking at the sunspots themselves. Sunspots do NOT look like that, even high res in h-alpha.

 

You applied a 3X drizzle. You've basically dithered your image into clip art.


  • RedLionNJ, TOMDEY and t-ara-fan like this

#3 Tapio

Tapio

    Gemini

  • -----
  • Posts: 3,146
  • Joined: 24 Sep 2006
  • Loc: Tampere, Finland

Posted 10 September 2020 - 12:49 PM

When was this taken ?

There hasn't been a sunspot for a while now.

 

"20 cm SCT ‘scope operating at F5.6 resulting in an effective focal length of 5.5 meters (including the camera)"

Don't understand this math. Did you have a focal reducer ?


  • RedLionNJ likes this

#4 TOMDEY

TOMDEY

    Cosmos

  • *****
  • Posts: 7,676
  • Joined: 10 Feb 2014
  • Loc: Springwater, NY

Posted 10 September 2020 - 12:51 PM

That's not resolution you're seeing. It's artifacting caused by over processing. You can easily tell by looking at the sunspots themselves. Sunspots do NOT look like that, even high res in h-alpha.

 

You applied a 3X drizzle. You've basically dithered your image into clip art.

Good call! The base image is actually respectable... Also that MB suspected something was ~too good to be true~ good suspicion, and reaching out for consul.

 

There are just a few contributors here who indeed manage to achieve spectacular resolution on moon, planets, sun... and (not surprisingly) are highly experienced with magnificent equipment, conditions, skills and persistence.    Tom



#5 BinoGuy

BinoGuy

    Viking 1

  • *****
  • Posts: 639
  • Joined: 11 Jul 2014
  • Loc: Michigan

Posted 10 September 2020 - 01:27 PM

What was the solar filter used?  At that size OTA presumably solar film, so this is a white light image (not something in one of the small bandwidths) and color was added.



#6 MeteorBoy

MeteorBoy

    Sputnik

  • -----
  • topic starter
  • Posts: 32
  • Joined: 19 Jul 2016
  • Loc: Toronto, Canada

Posted 10 September 2020 - 01:43 PM

Yes my image is highly processed.  However, both of my two images are in agreement with each other as to sunspot shapes.

 

The image was taken Aug 18, 2020, when a very small sunspot group formed.  I have attached Space Weather's solar image of the same group.  It too is in approximate agreement with the shapes and sizes of the sunspots (but was taken about 20 hours after mine and so the group has evolved somewhat).

 

I used Celestron's Eclipsart polymer solar filter.  Yes I was using a FR operating at 0.56x resulting in the 'scope's F10 to become F5.6.  The camera is mono and colored was added.

 

 

Attached Thumbnails

  • Space Weather Image - Aug 19, 2020.jpg


#7 t-ara-fan

t-ara-fan

    Surveyor 1

  • -----
  • Posts: 1,520
  • Joined: 20 Sep 2017
  • Loc: 50° 13' N

Posted 10 September 2020 - 03:03 PM

 
  • ZWO ASI178mm camera (2.4 microns).
  • 20 cm SCT ‘scope operating at F5.6 resulting in an effective focal length of 5.5 meters (including the camera).

Your suspicions were correct.

 

20cm @ f/5.6 ==> 1,120mm FL.  How do you get 5,500mm?

Image scale is 0.44"/pixel. Unless you have a PowerMate in there, which I would doubt since you have a reducer.

3x drizzle is fairly optimistic, and gives 0.15"/pixel. 

Upsampling 3x more gives 0.05"/ pixel or as I say: artifacts and nothing but artifacts.

 

For reference the Hubble telescope operates at 0.1"/pixel.



#8 RSX11M+

RSX11M+

    Sputnik

  • -----
  • Posts: 39
  • Joined: 27 Dec 2019

Posted 10 September 2020 - 07:14 PM

You are playing at the fun edge of astrophotography, pushing the limits of enhancement. It's natural that you are questioning results and conclusions. You may well be alone in this, as not many push this far. Beyond Dawes, there be dragons! So be it.

 

 

I too get an Image scale of 0.44" per pixel in the original images. The Dawes of that aperture is indeed 0.6"-ish.

 

Conventional wisdom has it that your camera is well matched to that scope, and that any smaller pixel size would be wasted. Maybe so, but it's a fun place to play. I've been experimenting with 1.4µ pixels myself, but that's another story.

 

It is very interesting (to me) that you measure 0.15" details in your final result after processing. Generic wisdom is that these are photographic or processing aberrations.  Can we set about proving that? Maybe.

 

What you need is a known yardstick. An object who's actual structure we know from close up observation, and who's images you could take and process as you did those in question. Then compare the output with what we know the object does look like. By inference, we could then judge the accuracy of your 0.15" results - no?

 

I suggest you target a known feature on the lunar surface for this testing.

 

We could continue to debate and conjecture, but it would be much more fun to really know.


Edited by RSX11M+, 10 September 2020 - 10:53 PM.


#9 Xyrus

Xyrus

    Mariner 2

  • -----
  • Posts: 215
  • Joined: 20 Jun 2017

Posted 10 September 2020 - 09:06 PM

Yes my image is highly processed.  However, both of my two images are in agreement with each other as to sunspot shapes.

 

The image was taken Aug 18, 2020, when a very small sunspot group formed.  I have attached Space Weather's solar image of the same group.  It too is in approximate agreement with the shapes and sizes of the sunspots (but was taken about 20 hours after mine and so the group has evolved somewhat).

 

I used Celestron's Eclipsart polymer solar filter.  Yes I was using a FR operating at 0.56x resulting in the 'scope's F10 to become F5.6.  The camera is mono and colored was added.

It's neither the size nor the shape of the sunspots that demonstrate you have over processed your image.

 

Do you understand what the drizzle operation actually does? Do you understand what is happening when you apply all those operations? Sunspots do not look like a speckled mess, regardless of magnification. Neither does the solar surface. The effect your seeing isn't resolution.

 

Here is what high res natural light sun spots look like:

Attached Thumbnails

  • Solar_Archipelago_-_Flickr_-_NASA_Goddard_Photo_and_Video.jpg

  • TOMDEY likes this

#10 alan.dang

alan.dang

    Mariner 2

  • *****
  • Posts: 265
  • Joined: 15 Dec 2011

Posted 11 September 2020 - 01:27 AM

Those are essentially sharpening halos that have transformed your photo into astro art.
  • TOMDEY likes this

#11 TOMDEY

TOMDEY

    Cosmos

  • *****
  • Posts: 7,676
  • Joined: 10 Feb 2014
  • Loc: Springwater, NY

Posted 12 September 2020 - 01:23 AM

Enthusiasm is wonderful. But astronomers start out by assuming that unlikely or impossible data are artifactual, until proven otherwise. Comet/planet discoveries, ET candidate radio signals, interacting galaxies, Martian Canals. And very high on the list as unbelievable is... when the equipment performance seems to defy the laws of nature. Very closely related to your ultra-fine structure this extremely common one >>> thinking a single hot pixel or cosmic ray hit is a star... on an image that has otherwise soft star images. The other protocol is that the onus is on the originator to prove that his discovery is real, not for others to prove that it is not.

 

Your opening salvo is prudent. And when you see that the ~structure~ seems to be four times better than Dawe's... the hope must end right there. At that point, all that remains is to investigate how the artifactual structure has contaminated the presentation. Professionals somewhat, and we amateurs especially... are very prone to over-processing. I'm sure we all have had these episodes... where our raw images look disappointingly ~meh~. Then we try a little contrast enhancement, unsharp masking, etc. etc. and is looks better (probably not more real, but prettier)... so we try a bit more and a bit more and it just keeps looking sharper and sharper. It's like when the kids get into mom's makeup kit for the first time. Pretty soon clownish is judged most appealing. Next morning second thoughts, "what on earth was I thinking?!"

 

A good technique is to process not at all or lightly, saving the renditions... continue until it's obvious we've gone too far. Then back off to gentle as best. Indeed, on the very best images under the very best conditions... sometimes almost no processing is best. And those rare images are quite astounding! Like the personable, young, fit model... who looks best with nothing on at all!    Tom



#12 MeteorBoy

MeteorBoy

    Sputnik

  • -----
  • topic starter
  • Posts: 32
  • Joined: 19 Jul 2016
  • Loc: Toronto, Canada

Posted 12 September 2020 - 06:13 PM

As the thread starter, I want to sincerely thank those who left comments.  They are appreciated.

 

As originally stated the sunspot photo is over-processed.  My intent was to question the resulting image resolution which seems extreme.  Is it real, and how can I maintain it with a more natural-looking appearance?

 

I have now re-processed the photo with a more realistic appearance.  This reduces the apparent resolution while appearing more natural like.  It still has a lot of "detail" for a 1.8 arc-min wide photo taken with a 20cm SCT.  These are very small sunspots.  What do you think of the photo now?

 

 

Taken August 18, 2020, with a 20cm SCT.  The insert is just 1.8 arc-mins wide.

Attached Thumbnails

  • Aug 18 2020 - Sunspot Detail a.jpg



CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics





Also tagged with one or more of these keywords: astrophotography, CMOS, EAA, equipment, imaging, optics, SCT, solar



Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics