M42 - my first attempt at a HDR process

[Filroden]

With a really clear night on Friday I decided to continue collecting data for an existing project (NGC1333) but also start a few more. I managed to gather a small amount of data on three new targets, the final of which was M42 which had just cleared the fence. Having tried this last year with the 9.25" SCT I already knew I would be in with a challenge. That, and I was quickly learning that the ZWO ASI 1600MM-C really needs short exposures in LRGB as it quickly over exposes star cores, even with 45s exposures.

So for M42 I set up a new sequence at -20C, 300 gain and a 50 offset and shot:

30 x 1s L

2 x 45s L (quickly stopped as the image was too over exposed)

58 x 30s L

15 x 30s R, G & B

I processes the 45s and 30s L frames together and blended them with the 1s frames using the HDR tool in Pixinsight. I was surprised how easy it was, though the resulting image after stretch was flatter than I'd like. I've tried to expand the contrast within the nebula but having read a little more overnight I realise now I probably could have done this better. I also realised afterwards that I didn't capture 1s RGB data so I have probably lost a little of the colour detail in the core around the Trapezium. However, the beauty of now thinking of targets as projects is that the next time I get a clear night it won't take much time to add this to the next sequence before collecting more data.

Anyway, here's my attempt. Unlike last year with the SCT, I haven't resolved the Trapezium with the Esprit 80ED but I prefer the subtlety of the colour in this image. I find myself zooming in and defocusing my eyes a little to just see the colours blending. I would welcome feedback

Version 1

Version 2 - added another run of 30s subs and also included a series of 5s subs in LRGB to improve detail in the core

Version 2.5 - same data as version 2 but I've (hopefully) improved the colour balance and made a few other finishing touches in Lightroom

[Filroden]

Digging back through my archive I found an image I took last year with the 235mm f10 SCT using the DSLR. If I remember, it was a stack of short subs, no more that 15s and I don't think there was more than 5-10mins worth. However, it was at a much higher resolution so I tried to blend it with the one above to bring out more in the core. Other than one residual line I think it's worked well.

Version 1

[Taff]

I do like your work nice details in your capture.Mark

[Filroden]

Thank you. The change to the cooled camera made it all that much easier. 

[PatrickGilliland]

Hi Ken

Nice work - may need to tweak setting on the combine or watch start post combine as some star are very bright Vs local context and has caused some halo's - that aside for a first try look nice.

Paddy

[Filroden]

25 minutes ago, PatrickGilliland said:

Hi Ken

Nice work - may need to tweak setting on the combine or watch start post combine as some star are very bright Vs local context and has caused some halo's - that aside for a first try look nice.

Paddy

Thank you. I'm hoping to get more data tonight and this time I'll remember to capture RGB data at the short exposure so I can improve the blend. Also, I've just seen a link Olly put into another post that should help me combine the images a little better. I'd noticed the halos but I didn't have the Photoshop skills to address them.

[ollypenrice]

This link? http://www.astropix.com/HTML/J_DIGIT/LAYMASK.HTM

For my money this technique knocks the others into a cocked hat. You use the saturated regions to mask themselves, in effect. (This was the original darkroom 'unsharp masking' as I understand it.) If you try it pay attention to the saturation in the short subs because bright signal tends to reduce colour saturation, so give your short subs layer a good lift in saturation before flatttening.

Olly

[Filroden]

16 minutes ago, ollypenrice said:

This link? http://www.astropix.com/HTML/J_DIGIT/LAYMASK.HTM

For my money this technique knocks the others into a cocked hat. You use the saturated regions to mask themselves, in effect. (This was the original darkroom 'unsharp masking' as I understand it.) If you try it pay attention to the saturation in the short subs because bright signal tends to reduce colour saturation, so give your short subs layer a good lift in saturation before flatttening.

Olly

That was the one. I'd seen it before months ago but lost my favourites. And thank you for the tip on saturation. 

[Filroden]

I finally processed the 405 additional lights I took on the 29 November. I had to discard the 1s RGB subs as they would not calibrate with any combination of bias, darks or flats and leave me something I could work with. Thankfully I took 30 x 5s LRGB subs so I think they've given me enough (plus the original 30 x 1s L) to preserve core details. Here's the capture details:

Although I tried to replicate my processing steps I have ended up with a very different colour balance and I'm not sure if I prefer it. It feels a lot warmer and it grows on me the more I look at it. I haven't blended in the older SCT data and I still need to apply the star halo reduction technique (and probably heal out the star cores).

Version 2

[Filroden]

And with a little more work to better balance the colours in Photoshop and some more finishing touches in Lightroom, here's the same data as version 2 but closer to a final image.

[PatrickGilliland]

20 hours ago, Filroden said:

I finally processed the 405 additional lights I took on the 29 November. I had to discard the 1s RGB subs as they would not calibrate with any combination of bias, darks or flats and leave me something I could work with. Thankfully I took 30 x 5s LRGB subs so I think they've given me enough (plus the original 30 x 1s L) to preserve core details. Here's the capture details:

Although I tried to replicate my processing steps I have ended up with a very different colour balance and I'm not sure if I prefer it. It feels a lot warmer and it grows on me the more I look at it. I haven't blended in the older SCT data and I still need to apply the star halo reduction technique (and probably heal out the star cores).

Version 2

Hi Ken

 

I personally prefer the colours in this one rather than the last version but the data is taking shape nicely.  Did you layer stars back in at some point? Just noticing the black halo which i would not expect from HDR but it can happen.

Great progress though and a nice rendition of this old fav.

Paddy

[Filroden]

4 minutes ago, PatrickGilliland said:

Hi Ken

 

I personally prefer the colours in this one rather than the last version but the data is taking shape nicely.  Did you layer stars back in at some point? Just noticing the black halo which i would not expect from HDR but it can happen.

Great progress though and a nice rendition of this old fav.

Paddy

I didn't remove the stars to process this. I wish I'd kept a processing log. However, looking at the original at full size I think it was the HDR combination. Version 1 did have a blend of a second, larger scale image from a 235mm SCT, which isn't in either v2 or v2.5 but I don't think it's that. However, I'm pleased with all three images and other than some clean up of halos and cores which I will do later in Photoshop, I think I will probably have this printed.

[PatrickGilliland]

1 minute ago, Filroden said:

I didn't remove the stars to process this. I wish I'd kept a processing log. However, looking at the original at full size I think it was the HDR combination. Version 1 did have a blend of a second, larger scale image from a 235mm SCT, which isn't in either v2 or v2.5 but I don't think it's that. However, I'm pleased with all three images and other than some clean up of halos and cores which I will do later in Photoshop, I think I will probably have this printed.

You could take just a star layer (which i do sometimes at a lower stretch and mask it back in with some careful blending as a lightness layer.  Only a few minutes to fix.  

Don't talk to me about printing!  I have 3 prints here from various sites (to metal) and none match the colours of the image i have sent.  If you wait until Monday I can update you on an interesting option though, went to UKPOS who make the illuminated backlit poster you have in shops.  Ordered a backlight panel and an image (you can order images an insert as required).  If they get the colours right may be and interesting display method.  

But a nice print is always good when you have produced something nice.

Paddy

[Filroden]

I use theprintspace.co.uk for my daytime prints and I have the ICC profiles for their paper/printers so I can soft proof in Lightroom. I have four test strips coming this week on matt, gloss, flex and metal to see which looks best.

[Barry-Wilson]

Interesting thread Ken and fascinating to see the development of your image.

My personal favourite is version 2 with its slightly softer feel - great work.

[Filroden]

1 minute ago, Barry-Wilson said:

Interesting thread Ken and fascinating to see the development of your image.

My personal favourite is version 2 with its slightly softer feel - great work.

Thank you. I also prefer v2 on screen for that same reason but I don't think it will work so well in print, so I created v2.5 to stretch the contrast a little more.