Melotte 15 / IC1805 ZWO ASI 1600 Narrowband

[JSeaman]

Thanks to Vlaiv for a little encouragment to get this one across the line!

This is my first DSO with the ZWO ASI 1600mm, I feel like I'm not quite there yet but it's certainly an improvement for my pictures. This is a bit of an unconventional set up, waving a 12 inch Newotonian around as an imaging scope but it works and it is good fun. I have moved my Lakeside focuser onto the stock 300PDS focuser and it performed very well too which is a bonus.

I was using 300 second/5 minute subs with 1x1 binning at unity gain (139) with an offset of 15 which gathered plenty of detail and meant fewer wasted minutes with any wobbles etc. Processing details are as follows:

Hydrogen Alpha

I got a little over 7.5 hours of good Ha data over 4 nights, this was stretched with a simple levels and curves adjustment intially and showed some good detail. I then masked off a few regions of interest (ROI) such as the bok globules, Mel15 itself etc. in Photoshop before using PixInsight to apply a light MultiscaleLinearTransform (2.5 to 0 in 0.5 steps over 5 layers) which smoothed out the really fine speckles on the background. 

I applied a 0.44 MorphologicalTransformation star reduction with no star mask and then created a star mask for the larger stars using PI's tool (scale of 9 and 0.26 noise threshold) before customising the mask in PS to the big hitters only before applying a .040 MorpholocialTransformation. I pulled the black point a little with levels and called that my Green channel.

Luminance

I took the Ha data further with a minimum pass filter at 1px in PS and then pushed the contrast up. I used shadows/highlights to change the midpoint contrast and pull down some of the bright spots in the middle and then decided to do something a bit whacky which I hadn't done before, I dropped the image into Registax and ran a Wavelet sequence on it. I used all 6 layers:

All Sharpen were set to 0.100

    Denoise    Preview
    0.20    8.9
    0.30    20.2
    0.40    43.5
    0.25    55.8
    0.00    72.4
    0.00    87.1

This was applied to the whole image and brought out the details well. A small amount of speckling was introduced so I just popped my ROI mask back over it and ran another light MultiscaleLinearTransform to smooth out the background. 

I then did a more agressive process as follows:

    Denoise    Sharpen    Preview
    0.40    0.100    100
    0.55    0.110    100
    0.40    0.130    46.6
    0.25    0.130    49.0
    0.00    0.100    60.7
    0.00    0.100    79.1

I masked this off for the centre of Mel 15 to really make the details pop. I used these two luminance masks together and it really did transform the sharpness of the image.

OIII + SII

I imaged both OIII and SII during a full moon after reading a number of threads about narrowband in moonlight. Indeed I have tried narrowband in moonlight before and got passable results but it always felt like the moon was washing things out. This target let me know for sure it isn't worth gathering anything other than Ha when there is a full moon. There was about 10 hours of data and just enough to pull out the result below. I probably haven't done it justive in post.

So this is it, my 33rd attempt at processing it (yes, really) and I'll probably hate it by this afternoon but it's where I'm at currently...

 

[Stuf1978]

Looks lovely to me, great work

[JSeaman]

Thank you very much, I'm a little disappointed but was really pleased with the Ha data. 5 years ago I would think this was the best image in the world but I guess that's the nature of the hobby, the more you know the worse you are

 

[Stuf1978]

Lol isn't that the truth. Nice image non the less 😄

[simmo39]

That is stunning, looks so 3d!

[glafnazur]

I agree with simmo39, it looks stunning.

[JSeaman]

Thanks very much indeed, helps make all those frustrating nights worthwhile!

[ultranova]

Wall hanger for sure.

Nicely processed, with loads of detail coming through,

well done 

Paul

[JSeaman]

Thanks Paul, appreciated

[Adam1234]

Very nice image

[JSeaman]

Thank you, it looks like you have a very similar set up!

[Allinthehead]

Very dramatic looking image. 

[MartinB]

I love this structure in the middle of the Heart, it is so photogenic.  This is magnificent, it is so vibrant.  The "in your face" process might not be for the die hard purists but who cares, if you are wanting an image to impress "normal people" this is it!!!

[JSeaman]

Thank you, that's definitely my target audience rather than pixel peepers who will easily find fault with it!!

 

[Snoani]

This is a beautiful image.  Well done.  

I am generally not a big fan of narrowband, preferring the result of RBG images.  However, this one is stunning.

[AndysAstroPix]

Hi JSeaman - A well composed image with plenty of bold impact! 👍

However the sharpening processes have left some artifacts that could be more handled by a more subtle method.

If it's ok with you, may I suggest using Topaz NR Ai on your masters? If you play with the sliders you can seamlessly smooth out noise and recover considerable detail with added sharpness.

However, don't fall into the trap of pushing it too hard as the Ai may interpret data to create details that don't exist! 😂

Anyway, love the colour palette and your perseverance is truly admirable! 🤩

cheers, Andy

[JSeaman]

I'm not sure which artefacts you are referring to but I decided to try it. I downloaded the 30MB installer, it told me I was already registered (which I wasn't) and then eventually managed to recover my non-existant account with a password reset. 

I ran the denoise function and it appears to be worse than PixInsight (bad ringing around stars and corruption of star shape amongst other things) and costs $80. I don't think I'll be recommending this. And this is what 'AIClear' does with it's lowest settings

 

Edited March 13, 2021 by JSeaman

[AndysAstroPix]

16 hours ago, JSeaman said:

I'm not sure which artefacts you are referring to but I decided to try it. I downloaded the 30MB installer, it told me I was already registered (which I wasn't) and then eventually managed to recover my non-existant account with a password reset. 

I ran the denoise function and it appears to be worse than PixInsight (bad ringing around stars and corruption of star shape amongst other things) and costs $80. I don't think I'll be recommending this. And this is what 'AIClear' does with it's lowest settings

 

I wouldn't be too hasty to write Topaz off just yet although I agree it does not perform well on images with stars. (I don't use the Ai clear function.)

However once stars are removed with Starnet++ magic can and does happen! 🙂 Each to their own of course.

FYI - In the examples attached...

1) Ha stack before & after Starnet++

2) Ha stack now starless before & after Topaz NR ai

3) Finished tricolour NB  image from my TAK TOA 130 vs. Planewave 20"

4) Topaz Nr Settings used

5) Details recovered by Topaz NR from M8 core

6) Core Details from my Tak compared to Hubble for purposes of determining whether details were "real". 

Final versions of these images can be seen on my Astrobin - N206 M8

If interested, I gave a presentation on The Astro Imaging Channel last year, I discuss using Topaz NR fro Astro from around the 40min mark.

Hope that's useful! 🙂

Cheers

Andy 

 

 

Edited March 13, 2021 by AndysAstroPix

[JSeaman]

No I'm afraid it's not adding anything to PI imho and especially at that price but thanks anyway