Help with Solar (partial Eclipse) Image

[decay]

Hi all,

this is from the (here in Western Europe partial) solar eclipse of June 10, 2021. Since I don’t have a solar filter for my newton, I captured the data with the finder scope (50/~165). I removed the eyepiece and attached my ASI 120 MC with adhesive tape. I know, that seems pretty ridiculous, but the quality of the small scope is surprisingly good.

FireCapture was used to capture the data and PIPP and AS!3 to create this image (a bit post processed, I don’t remember exactly):

Sun_122111_l6_ap271_P60_1.png (109.23 KiB) Viewed 3811 times

I was quite happy with this image, but lately I saw that ST has a special colour matrix for solar images:

viewtopic.php?p=12647#p12647

And of course I tried it out!

solar_eclipse_test-st-2.jpg (51.56 KiB) Viewed 3811 times

I was absolutely blown away as I saw what ST did with my data! The HDR and the SVDecon modules revealed a lot of structures. Unfortunately I’m not sure if these are real structures? Shouldn’t solar granules be smaller and invisible at this resolution? At least the structures at about 7:00 and 11:00 o’clock are real, they can be seen in other images as well. And some kind of banding is visible ...

What I did in ST:

Open TIFF image from AS!3 as Non-linear sRGB source (Is this right?)
- reverse stretch, activate Tracking

Crop

AutoDev

Wipe - with defaults

AutoDev – final stretch, no ROI, defaults (what about dropping Detector Gamma a bit?)

Contrast – dropped Locality down to 5% (maybe it is better not to use this module?)

HDR
– Context Size 10x10, Shadows Detail Boost up to 70%
- I had no success to get the details/structures to reach the border of the sun – is this due to missing resolution?

Sharp – Structure Size: small, Planetary preset

SVDecon - Planetary preset, played around a lot, but the defaults seem to work best!

Color – Matrix: False Color: Solar, Green Bias Reduce to 1.57, Dark Saturation 2.2, this gives a nice transition from red/orange (border) to yellow (centre of the Sun)

Super Structure – not used, maybe use Saturate preset to adjust colour afterwards?

Noise Reduction – without effect, as expected (?)

I searched the ST forum, but I only found a few hints/threads about solar/lunar/planetary processing and so I would be interested to hear if this is a reasonable processing? Any hints/advice/help would be much appreciated.

(And if anyone has any hints concerning the processing in AS!3 or PIPP (in this special case) I would like to hear them too.)

Thanks!
Best regards, Dietmar.

P.S.: If someone would like to take a look, the TIFF I used as input for ST may be found here:

https://c.web.de/@334960167135216273/oy ... Vd6fPyMRLw

Edit: corrected the type of this solar eclipse: of course it was partial as seen from Europe.

[admin]

Hi Dietmar,

It is commendable to question your results for veracity - if only more people did this!

The local variations in brightness appear to be real features of your dataset, though they have indeed been amplified a lot (chiefly using the HDR module) in your image. The Sun's surface is a very turbulent place, with lots of activity happening at all sorts of scales.

As for what's "correct" or "useful", it totally depends on your goals. Strongly amplified detail can be quite useful for examining an object (see also this technique) and can help reveal interesting processes.

The behaviour of the Denoise module does, however, gives a clue on how "sure" it is about the introduced detail; it wants to reduce the detail quite a bit, which is an indication that the data was pushed fairly heavily.

The Sun is a sphere, so it stands to reason that the detail, from out point of view gets "compressed" towards the edges of the sphere, as it no longer faces us directly;

(Image from Paul Sutcliffe)
Indeed, this appears to be happening. The "chunk" that is missing due to the eclipse is very helpful in this sense, as we know that detail visibility should NOT fall off like it does for the edges of the sphere. It doesn't appear to do so, so this too speaks to the veracity of the detail.

Hope that helps,

[decay]

Hi Ivo,

thank you very much for the detailed answer.

It was my assumption as well, that these structures are real features. It looks a bit like an orange, the structures are changing with the angle the surface has towards us (as you explained). It's the same thing we know from images of Jupiter or Mars.

The behaviour of the Denoise module does, however, gives a clue on how "sure" it is about the introduced detail; it wants to reduce the detail quite a bit, which is an indication that the data was pushed fairly heavily.

Thanks, I guess that's a valuable hint in general. I processed the image several times and in the case of that one I posted, Denoise had almost no effect. But on previous runs I saw the detail reduction you described. I don't know what the difference was ... never mind.

Thanks as well for the link to the paper on fractional differentiation. I will read it later, right now I am on vacation and my family doesn't allow to spend to much time on astronomy

Could you please be so kind and give me a hint, if loading the image as Non-linear sRGB source and activating tracking was the right way in this case? Thanks!

Best regards, Dietmar.

[admin]

Could you please be so kind and give me a hint, if loading the image as Non-linear sRGB source and activating tracking was the right way in this case? Thanks!

Hi Dietmar,

I don't have enough information to give you guidance on this. It totally depends on how this was preprocessed and how the camera that acquired the data was configured at the time...

[decay]

I don't have enough information to give you guidance on this.

Yes, I was afraid you would say that, Ivo. And that I will have to use my own brain . Thanks for your reply anyway
Best regards, Dietmar.

[decay]

Hi Ivo,

I took some time and read the paper about fractional differentiation you mentioned.

Strongly amplified detail can be quite useful for examining an object (see also this technique) and can help reveal interesting processes.

I have to admit, I never heard about fractional differentiation before. I learned about partial derivatives, total differentials and much more … but it’s a long time ago (and it was not ‘in a galaxy far far away’ )

The application of fractional differentiation for image processing seems to yield amazing results, as shown in the paper. But unfortunately I do not understand how or why it works and the paper only describes the application and the results. The meaning of ‘normal’ differentiation is a quite easy to understand – in mathematics, physics and of course in image processing. But fractional differentiation seems to be not that easy to understand. I read, that already Leibnitz and Euler worked on it – but it took 200 years until the first real life applications came …

I saw, that there is a fractional differentiation filter in ST’s Layer module … maybe I should give it a try ...

Anyway, thanks for expanding my horizon at least a little bit.

(I do not expect any response; I just wanted to write down my thoughts and give some feedback.)

Best regards, Dietmar.