Second attempt at orion w/Star Adventurer

[pipnina]

I went down from f8 to f2.8 (3.3 stops brighter!) and managed to get the flame showing in single subs, despite the nearly full moon! However, I am struggling to process the image. Vignetting is massively increased and GIMP doesn't seem to be cutting it (or I'm doing it wrong). Some of my attempts have brought out M78, flame and horse head really well, but left most of the image with glowing green patches, multicolored noise and not even removed the gradients fully. A lot of people recommend PixInsight for things like gradient removal but I was wondering if other software exists for that purpose specifically. I use Linux so support for that platform is preferred but not (necessarily) required.

 

I've included my stacked TIFF if anyone wants to do better than I have with this one . My shoddy attempt below...

Hopefully you guys got clear skies as well!

Thanks and good night

Orion mk3.tif

[SilverAstro]

1 hour ago, pipnina said:

I've included my stacked TIFF 

Thanks, I do like having a fiddle in GIMP !

Here is my humble offering.

[JemC]

I think you have done well with this,

i have just acquired a SA, i'm finding it is taking a little getting used to, bit more fiddly with just a DSLR and lens, 

i was out for the first time last night with it on Orion, going to post up my first image in another thread,

 

Regards 

[Mognet]

Vignetting does become more apparent with wider apertures on camera lenses. Image stacking programs will have the option to use flats and will handle the processing automatically. It should be possible to do the equivalent in GIMP subtracting a flat from the original image using the layers tool, and then do the processing from that blended image

[SilverAstro]

There was no flat available in @pipnina 's tiff file so I had to resort to a Gaussian blur mask/subtraction and a couple of 'curves'. It was a late night fiddle and much enjoyed doing, thanks pip for providing the tiff for us to play. (usually only jpgs are posted which make for mixed results,,, and to my shame I did also post a jpg , it was late !! )  if wanted I could post an uncompressed png at 6Mb, but my modified tiff is about 45Mb ! dunno why yet.

[pipnina]

1 hour ago, SilverAstro said:

There was no flat available in @pipnina 's tiff file so I had to resort to a Gaussian blur mask/subtraction and a couple of 'curves'. It was a late night fiddle and much enjoyed doing, thanks pip for providing the tiff for us to play. (usually only jpgs are posted which make for mixed results,,, and to my shame I did also post a jpg , it was late !! )  if wanted I could post an uncompressed png at 6Mb, but my modified tiff is about 45Mb ! dunno why yet.

17 hours ago, Mognet said:

Vignetting does become more apparent with wider apertures on camera lenses. Image stacking programs will have the option to use flats and will handle the processing automatically. It should be possible to do the equivalent in GIMP subtracting a flat from the original image using the layers tool, and then do the processing from that blended image

I took 100 bias frames and 15 flats (pointed my camera at my computer monitor displaying a white image) and restacked the image. Still gradients but possibly not as bad? I don't know if DSS will remove bias from light in absence of darks but the noise pattern in the image looks similar... Wondering if my flat files are too strong as starting to glow around corners more, but moon might be playing a role in strange ways.

TIFF provided again

 

Orion MK3.3.1 16.tif

[wimvb]

Here's my attempt, using PixInsight

Besides the left-to-right gradient, there was also a circular gradient. It's may still barely visible as a circular colour pattern. You might want to check your flat calibration.

You also just about captured the Horse head nebula, but it's too close to the noise floor to bring it out.

 

(click image for full size)

[pipnina]

10 minutes ago, wimvb said:

Here's my attempt, using PixInsight

Besides the left-to-right gradient, there was also a circular gradient. It's still barely visible as a circular colour pattern. You might want to check your flat calibration.

You also just about captured the Horse head nebula, but it's too close to the noise floor to bring it out.

I think I see little bits of barnard's loop next to M78!? Thanks for the incredible process!

I did only point my camera at my computer monitor, so perhaps it's just that, my screen being a TN panel, there are some effects going on? Maybe I took my flats incorrectly? (I used 100iso instead of 800 for the lights & bias, exposure at 0ev).

An interesting discovery is the prevalence of the banding, even in pixinsight (edit: I always thought that it was a problem with GIMP). My camera only outputs 12-bit colour in its raw files and DSS assumes the existence of the next 4 bits, so maybe I'm doomed to keep that banding so long as I don't have a 14/16 bit camera?

 

P.S. when I tried to access the first image, it gave me a 404 error and now the image doesn't show up in the post?

[wimvb]

I edited my previous post. Try reloading the page.

I think the band is caused by flats rather than Barnard. Here's an extreme stretch before applying the third iteration of DBE. The circular gradient is an optical issue, which should have been corrected by flats.

[pipnina]

1 minute ago, wimvb said:

I edited my previous post. Try reloading the page.

I think the band is caused by flats rather than Barnard. Here's an extreme stretch before applying the third iteration of DBE. The circular gradient is an optical issue, which should have been corrected by flats. You could try

Ah. Maybe my laptop screen (an IPS panel) would be a better candidate for flats. Am I supposed to take flats at the same ISO as bias? I have heard that bias is a requirement for good flats, but I don't know much about flat calibration beyond the fact they are important.

Thanks again!

[wimvb]

To calibrate your flats, you need bias frames at the same ISO, yes. If you shoot the lights at another ISO than the flats, then you basically need two set of bias frames.

Otoh, if you always shoot your flats at the same ISO, you can reuse the master bias. The bias signal of a camera doesn't change (much) over time, so you don't need to reshoot bias frames often.

To improve the master flat, try with a t-shirt over the lens/scope. When I imaged with my DSLR, I used to take flats by pointing the camera (with 135 mm lens) at a white, evenly illuminated panel. Then I made sure that the image histogram was between 1/4 and 1/2 from the left hand side in the camera preview screen. The only thing that was tricky, was trying to avoid imaging the camera's shadow on the white panel.