M42 and Running Man

[MarkS]

I took this on 22 Dec.  There was still snow on the ground and the background sky glow was the brightest I've measured in Sidcup.

Peltier-cooled and Ha-modified Canon EOS 350D on Skywatcher ED80 with WO MkII 0.8x reducer with CLS filter.

The subs were as follows, all at ISO 800:
39 x 5min
4 x 1min
15 x 4sec

Instead of the "traditional" layering technique to deal with the huge dynamic range I tried a slightly different approach: I first created an HDR (high dynamic range) image from the different lengths of sub and then applied asinh dynamic range compression.  Finally it was 2x2 binned and cropped.



Larger size here:
http://gallery.orpington-astronomy.org.uk/albums/userpics/10046/m42_22122009final.jpg

Mark

[Carole]

Fabulous image as usual Mark, you never fail to impress.  I don;t remember seeing all that wispyness around these nubulae before.

"traditional" layering technique to deal with the huge dynamic range

I would like to find out more about this technique.

Carole

[Fay]

EXtremely nice Mark, I am thinking in PS, only certain editions offer HDR

[JohnP]

Nice wide field Mark. Still pretty noisey for length of exposures but considering conditions it was taken under very nice. It looks kinda 'milky' but I am guessing this is due to snow reflections lowering contrast?

John

[MarkS]

Thanks for the comments.

Carole, 
You must have seen the wispiness before, the recent M42 images from both Daniel and Chris had better wispiness than mine!

Fay,
You are right about HDR, my version of Photoshop doesn't support it.  Even if it did, I don't have any other application that can import the resulting 32 bit (HDR) TIFF file. 

John,
This is still a work in progress - you're right about the high noise levels and the overall milky appearance.  I wrote my own software to create the HDR image and then to apply the asinh scaling, because I don't have any application that handles the wide dynamic range in the combined subs (it actually required 25 bits per RGB channel).  The milkiness is a result of the severe scaling required to preserve both the outer wispiness and the bright trapezium in a single image - it has severely impacted the contrast.  On the other hand, the Photoshop "layering" approach preserves contrast but is unscientific in the sense that it does not preserve the correct relationships in relative brightness for all the elements of the image.  I'll play around with the scaling but, in the end, I think I'll have to sacrifice scientific accuracy for artistic merit.  But I'll make it "sound" scientific by calling it locally adaptive range scaling (or similar).

Mark

[Daniel]

Hi Mark, great work, I've often wondered if HDR techniques woulkd work well on things like M42 and M31, did you have to mess with levels at all before combining as an HDR?  Might have a bash at this myself as soon as I get another night out imaging (I hear Friday and Saturday may be good)


Daniel
:O)

[Rocket Pooch]

Hi Mark,

What are you writing in?

Chris

[MarkS]

Daniel,
No manipulation of levels was required before combining as an HDR but I did precalculate the relative brightness between exposures  - essentially by dividing one exposure by the next to derive the multiplier required for all the pixels in that exposure.  I ended up with pixel values in the range 1 - 21,000,000

Reading the Wiki on HDR images it appears that the asinh scaling I later applied would be classed as a global (i.e. spatially uniform) tone mapping operator.  Some kind of local (i.e. spatially varying) tone mapping operator might well give a better looking overall result.

Chris,
I'm using C++ with MFC


Mark

[mickw]

 
I need a lie down 

[RobertM]

Lovely image Mark, even better that it been in part produced by your own software.  It's especially impressive considering that it was taken in Sidcup.  Doesn't the arcing caused by the trains on the track behind your garden cause problems?

[MarkS]

They don't arc outside my house - luckily.

Mark

[MarkS]

This version applies a local operator when converting the HDR to preserve more contrast.



Larger version(2/3 scaled and cropped) here:
http://gallery.orpington-astronomy.org.uk/albums/userpics/10046/m42_22122009finalv2.jpg

Mark

[JohnP]

Looking good Mark - definitely less noisey & core looks better however overall contrast is lower. If you could preserve contrast in first with lower noise in the second it would be excellent.

John

[MarkS]

If you could preserve contrast in first with lower noise in the second it would be excellent.

I'm going to have another go because I'm still experimenting.   I'm beginning to think it is much easier to use the traditional "Photoshop layer" approach to add the bright core later - the final result ends up being more or less the same.

I also had a bit of trouble with star colours because the red layer is slightly out of focus - dunno exactly why.  However, it means that the star light has spread by different amounts. 

Mark