Processing FULL Disks with Flats

[Macavity]

With little (obvious) action on the Solar Disk at the moment
(And a break from DIY decorating!) I thought I'd post this
update on experiment re. processing FULL Disks via Flats. 

As ever, no rocket science and stolen ideas acknowledged!
Comments welcome, but in the interests of continuity, this
will be in a few bits, so please hold off for the moment...   

Hope someone finds this useful!

[Macavity]

PART ONE:

Some of you may be familiar with the sort of Full Disk Image
I post on this forum! They usually looks something like this:



Neither the best or the worst? Rather typical for a Lunt50 + DMK41 combination?

But not without problems either! My DMK41 camera is a splendidly stable camera...
It banishes Newtons Rings at a stroke! But it does have some downsides. Notably,
if you look closely at the TOP of the image there are some dodgy (black) pixels!

As most of you are aware, buying a NEW camera may not solve present
issues and may (particularly for H-Alpha!) open a new Can o' Worms!

Following up ideas of "Cling Film Flats" (Michael, Alex and many others!) 
I have had significant success with PART (Barlowed) disks using Flats, but
FULL Disk have always proved "challenging"! (I use AS!2 for processing)!

With a SINGLE layer of Cling Film and (sensible default) AS!2 settings:



The FLAT is on the left and the stacked image on the right! 

The obvious issues is that the image is DARK and the limb
shows brightening where the flat doesn't cover Solar disk.

Also, the FLAT is decidedly asymmetric, so I added a second
layer of Cling Film *perpendicular* to the first! I then get:

This looks rather better! But the basic problems still exist...

Encouraged... I kept added further layers of Cling Film in a
"Union Jack" orientation (To allow for uneven scattering!)

Eventually with EIGHT layers I got something like this:



N.B. This corresponded to a limit where the eposure
began to slow down my (slow!) data acquisition rate!

But this looks much more hopeful! I then processed
the image via Registax 6 (as an example!) to give:



This is Pretty Good, right? Hey, C'mon Folks...  

BUT it now has other issues! *Principal* amongst 
these is that *flattening* the image has removed
the "Life" of the original e.g. limb-darkend image?

In the next part, I'll show how I addressed this...

[Macavity]

PART TWO!

Starting from the Final image above, I thought... If the image
is already *almost* flat, why not Flatten it properly? Then try
to reconstruct a properly "rounded" Sun from the result... 

Here's the above image (again!):



The above was processed using *Artificial Flat* techniques:

I then generated an *Artificial Sun* using the IRIS Software!
The following shows the effect of multiplying the above,
by the (IRIS) generated Sun:



Oops! It's gone a tad DARK again... But with a bit of tweaking:



That's pretty much the END of this! As ever, you can add the
"Bells and Whistles"... Colour, Proms etc. Ultimately this...

OK, That's it Folks! Not sure I'd want to do this EVERY time!
But maybe on special occasions...

*** THE END ***

[Macavity]

P.S. Forgot this bit! To demonstrate the removal of various
"crud" via the application of Flats, here's a comparison of
the enlarged images "before & after". It really does work!



I'm definitely DONE now!  

[Peter Drew]

Well worth the effort. Should boost LS50 sales!   

[Montana]

That is great! rather than 8 rounds of cling film how about trying the supermarket fruit bag instead, I have great results with that. It doesn't matter about exposure time, I usually ramp it up to get around 2 frames per second and match the histogram to the original by upping the gain instead. It doesn't matter how high the gain is so long as you have a good number of frames to create your flat with (evens out the noise). I find 200 is plenty.

Don't worry about limb darkening in Halpha, if you look by eye there really isn't that much, it is quite subtle, not as much as white light, in Halpha an etalon sweet spot can cause more, my PST is terrible for this compared to the Solarscope. 

Many folks find using the drift method easier and simpler for full discs, if you have the space just let the Sun drift from one side of the field to the other during capture, then Autostakkert will even out any defects during stacking.

Alexandra

[Macavity]

Thanks Alex - GLAD you dropped by - To allow me to credit you
with the inspiration for this! Yes, I did try Morrisons(!) fruit bags,
with some success, but found them a bit dense for the purpose.
Or DMK cameras don't have a lot of additional gain to play with? 
But worth while experimenting with a *variety* of materials! 

Certainly true that you don't need to get "too exotic" for FLATs
re. event number, gain etc. Within reason the end result can be 
surprisingly effective? Amazing what you can drag out of a FULL
disk with a bit of *judicious* processing (and better seeing)! 

Solar Limb Darkening is an interesting thing, Eh? Basically,
the Sun is not quite a "Billiard Ball in the Sky" after all! 

With limited aperture and restricted possibilities re. Barlowing
the Lunt 50, Flattening  (real & artificial) can help a great deal
re. "book-matching" the Quadrants for e.g. 2x2 matrix images.
Whatever else, you "Can't see the join" (as much) anyway! 

I lost a bit of contrast on the 2x2 disk, but I'm working on it...

[Montana]

Ah well, if you must try those cheap Morrison's bags, you need the posh Waitrose ones, they are much thinner! in fact the one I got the other week was so thin I had to double it over to get the right density  I guess they must be cost cutting.

A great shot above, beautifully done

Alexandra

[Macavity]

On 7/28/2017 at 10:02, Montana said:

Ah well, if you must try those cheap Morrison's bags, you need the posh Waitrose ones

Brilliant! lol. Exactly what I was thinking re. that... 

Don't get me wrong. I sense FLAT production is fairly
amenable to cam settings as you were suggesting? 
As long as they are "reasonably grey"! I may need to
think about the quoted noise figures versus "gain"...

It is clear that it's a good thing to get these FLATS 
"as flat as possible" and to reduce the "black bits"!
Obviously using the square *ROI* is a good start!

I have tentatively tryed a combination of 50% grey
image combined with a real world flat! One of the
GIMP mask modes selects the maximum of the two
at any point! This might become a lifetime study...
(I doubt it really merits that)

[Ibbo!]

Lidl bag user here I have tried others.

my other half is a craft type so.....

 

 

 

She says she will make me a better holder once I am happy with the flat results.