M42 first attempt

[DaveHKent]

My first attempt at the Orion nebula. I'd appreciate any comments about my image processing technique.

Taken on a night with a lot of humidity in the air and a half moon giving a nice haze over everything.

Skywatcher StarAdventurer, Canon EOS 700D 75-300mm lens at 300mm f5.6. 75x30 second shots. Stacked in DeepSkyStacker and processed in Gimp. I didn't take any flats or darks, but clearly should have done.

Processed mainly by dragging levels. This brought out a green halo around the nebula, so I used curves to remove the green. I think I've perhaps stretched it too much because at higher magnifications the gradients between the colours are very visible. I think when I selected just the background to darken that, I've given too harder edge to the nebula and should have taken some shorter stills as the core is over exposed. I'm not sure what all the small red and green flecks across it are. Perhaps something that could have been removed with darks.

Should I have gathered more data? Would it have been better to select around the nebula for a lot of the stretching, so the stars don't seem blown out?

It's my first attempt at any DSO so I am pleased with what I've got, but want the next one to be better!

Attached, the processed image and the original TIFF.

2021-01-09-m42_stacked.TIF

[AKB]

This is very good.  The red/green spots are from hot pixels stacked in slightly different places after the stars have been aligned.  Bias/darks may help and/or a more sophisticated stacking process (sigma clip, etc.)

A quick look at your raw TIFF shows more there than you have shown – you have the 'Running Man' nicely too:

Great stuff!  My early ones (or even recent) are often not so good!

Tony

[DaveHKent]

Cheers, I do seem to have lost the running man in my processing.

 

What did you do to bring that out?

[AKB]

Just this:

• background subtraction (DBE)
• colour correction
• noise reduction (TGV)
• arcsin / histogram / curve stretching

[DaveHKent]

Thanks. That looks like the opposite of what I did, which was stretching first.

 

More googling and learning required. Having only done planetary and lunar imaging before, it is amazing just how much of this is in the post processing.

[Swillis]

@DaveHKent Worth trying Siril for processing. 

I'm a beginner myself and have a lot to learn about processing, but I kept hearing about Siril and how it is better to do certain bits before stretching. So far I'm very pleased with the results even if I know there is more detail there than I have brought out. 

Also it's recommended not to take darks for Canon DSLRs. Just take bias and use that for bias and dark frames. Apparently Canons have part of the sensor masked off which is used to subtract the dark current, so darks frames are not required. 

[rickwayne]

In general, you want to do as much processing as you can on linear data, that is, before stretching. One of the advantages to astro-specific processing programs such as PixInsight, Astro Pixel Processor, and Siril is that they can present a preview of what the data would look like stretched, but on still-linear data. I haven't used The GIMP in a long time but in Photoshop, you can simulate the same thing with a Curves layer on top of everything else, and then don't ever Stamp Visible or save the result unless the layer is switched off.

For stretching in non-astro programs, my best advice is to patiently apply incremental stretches instead of doing it all in one go, and not to treat the whole range equally. What you want to do is bring up the dimmer areas but touch the brighter ones as little as possible. Your TIF is already pretty blown-out in the Trapezium area, so any stretch applied at all to that region will just give unrecoverable white. Levels are just a simplified version of curves, but curves give you finer control over that sort of thing.

Best forty bucks you can spend in astrophotography: Charles Bracken's The Deep-Sky Imaging Primer. Steve Richards' Making Every Photon Count is also well-thought-of, though I've not read it myself.

M42 is actually quite a challenging target for advanced imagers, primarily because of its extremely high dynamic range -- dim dust all around, but the bright Trapezium too. Nice job on the capture, there are a lot of things to like about these data!

Quote

 

 

[DaveHKent]

Thanks for all the advice. I've tried Siril and it does looks pretty good. Following a couple of videos, the background removal was nowhere near what they got, but I could see on the histogram view that I had a lot of noise there. I'm annoyed at myself for not capturing darks, flats and bias, which would have made a lot of difference and got rid of all those red and green flecks.

I've have nothing to learn if the first one was perfect. I think this is the best I'm going to get it. Most of the work done in Siril, as per the advice here, but I did use gimp at the end just to bring down the brightness of the background a bit. I'm reasonably pleased with the final result.