"Flattened" Image Recovery (Synthesised Sun!)

[Macavity]

You can tell it's raining... Eh, Readers? 

But continuing on with my fruitless and endless quest for "perfection"...

I suspect I will be stuck with using SOME sort of "Flat" in WL processing.
And it's certainly no great hassle to de-focus the image a bit! As ever edge
effects are present when a flat is "divided out" during Registax processing.

Having chopped out the worst stuff (Using GIMP to "cut out" the disk! lol)
the resultant image is certainly minus any DUST and camera artifacts... But 
looks rather lifeless!  (You lose the nice "modeling" of Limb Darkening!) 

The "Flattened" Image:

I was convinced that there must be some way of wrapping the image around
a "sphere", but success proved decidedly elusive!!! It just didn't LOOK right!

After some Googling, I came across a method in IRIS software to generate a
"Synthetic Sun" -- With the right sort of brightness variation across the disk!

SYNTHE_SUN [X_CENTER] [Y-CENTER] [RADIUS] [WAVELENGTH] [INTENSITY]

Produce a synthetic image of the sun disk taking into account a realistic the limb darkening. 
The algorithm is based on the H. Neckel model see H. Neckel, Solar Physics, 229, 13-33, 2005. 

http://www.astrosurf.com/buil/iris/new552/new552_us.htm

After the canonical messing around, I was able to produce a nice BLANK disk
of the same size as the above (full disk) image! Using "Multiply" image mode
I was able to combine the Synthesized Sun with the Flattened Image to give:

The "Rounded" Image:

Not perfect! Due to a slighty dodgy flat, there is still the above "polar brightening"!
Hopefully I can put some of the experience of taking flats to better use next time.

Just for fun, I have attached a SYNTHETIC SUN (I guess we could all use one!?!) 

[johnfosteruk]

Ooh. I've lots of data I can rescue with this. Nice work Chris. 

[Peter Drew]

Congratulations Chris, the result was well worth the effort!.  

[Iainp]

Fascinating! And a very convincing result

 

[Pete Presland]

Nicely captured.

[Ewan]

Looks good Chris.

Not sure if anyone has already mentioned this but I shall explain how I take flats using the defocused method.

I don't suffer from NR's anymore because I use an Eliminator but flats are essential for good images either way.

When you select your area of interest I take it you have all your capture settings ready to capture a run ?

What you may find (as I have) if you set the settings up, say for a limb capture, then do your capture(s) you then go to the centre of the disc, defocus & take 300+ frames for the flat file, when you process & use the flat file it could be too bright in one area ?

The best thing to do is decide on your area of capture, do select your settings then slew to the centre of the disc, do this BEFORE you capture any data, if your settings are around 80-90% of the Histogram they look correct but when you slew to the centre they may rise & the flat will be too bright.

Best bet is select are, select settings, slew to centre, adjust settings to fill 80-90% of Histogram then slew back to the target area then grab your data, or indeed grab the flat first as this also prevents 'forgetting' to take one.

I know it looks like I have waffled a bit but wanted to make it as straightforward as possible.

Have you not thought about using PIPP, Autostakkert 2 & then IMPPG for the whole process Chris ?

I find IMPPG is a lot better for processing & Autostakkert is great for stacking any images, just a thought mate.

Atb & if I can help at all please get in touch. Good hunting.

[Davey-T]

You could have used an image from a few days ago, it looked just like your synthetic Sun 

Dave

[Macavity]

On 3/31/2017 at 20:54, Ewan said:

I know it looks like I have waffled a bit but wanted to make it as straightforward as possible.
Have you not thought about using PIPP, Autostakkert 2 & then IMPPG for the whole process Chris ?
I find IMPPG is a lot better for processing & Autostakkert is great for stacking any images, just a thought mate.
Atb & if I can help at all please get in touch. Good hunting.

FEEL FREE to Waffle... As far as I'm concerned!

First reaction... Oh Noes... He wants me to change my whole "work flow". lol
But Registax 5 is showing modest age and seems to struggle reading LARGE
mono files?!? I have to pre-process them with "Virtual Dub" for some reason!
Something to do with "I8" versus "RGB888" format?!? All a bit of a pain...

BUT I decided to try a head-to-head between Autostakkert and Registax5!
I tried to make the contest as fair as possible! I still used Registax5 Wavelets
and "level" cutting to post-process... In the way I would normally do so...

I selected the best 100/1800 of a recent Barlowed ED66 White Light effort!
PLEASE IGNORE the rather "naff" attempts a FLAT application. It removes
"chip non-uniformity", but I have yet to optimise flats for "dust removal"!

As every, I would strongly urge others to do their OWN experiments and
take the below with a big pinch of salt! But superficially AS!2 > REG5???

Perhaps it would be fairer to pit Registax 6 versus AS!2
I only use ONE alignment point with Registax 5 etc. etc.
Whatever, the result is CLOSE! (Within operator error?)

Registax5 Image Stacking:

Autostakkert Image Stacking:
 

[Ewan]

Maybe it's the processing Chris but the lower image clearly is sharper then the other. The size of the Alignment points in Auto does make a difference though, 50 is around the best size from what I have seen so far.

Nice images btw.

[Macavity]

On 3/31/2017 at 20:54, Ewan said:

Not sure if anyone has already mentioned this but I shall explain how I take flats using the defocused method.
I don't suffer from NR's anymore because I use an Eliminator but flats are essential for good images either way.

Rreturning briefly (or not) to your post. Yes, I do sense ridding oneself of Newton's Rings
is a necessary prerequisite to other things? Sounds like your "Eliminator" does the BIZ!

Previously I succeeded in removing NRs via Drifting (Reconfig EQmod to track at 90%!)
But it was ever a bit "hit or miss", if patterning is very noticeable? Plus I forget to drift. 
With the DMK41, now NR Free, I still get noticeable CCD patterns plus (inevitable) dust. 
So I continue to work on some sort of overall strategy re. producing / including FLATS! 

TO FIRST APPROXIMATION: I don't need FLATS with H-Alpha images tho! (Mostly...)

But White Light remains a major challenge. Did more experiments yesterday. Mostly
re. "Drifting". Again the latter didn't help much! Dust blobs became streaks and CCD
patterning (mostly) remained. This being parallel to both drift + scan lines anyway!

Random experiments with EXTREME defocussing (to focusser limit!) gave rise to the
above images. Chip variations eliminated, but dust motes badly addressed. Clearly
typical systems are not well collimated - When you BIGLY depart from nominal focus!  

But I have HOPE regarding de-focussing / or diffusing (or a combination thereof).
Experiments with all of this stuff to be continued (when/if the SUN reappears)! 

P.S. With but few dust motes, it is far easier to "Snopake" out the odd blob? lol
Most seem rather amenable to GIMP's "Clone Stamp" techniques or whatever.