Crux Coalsack Mosaic

[avandonk]

The HDR method I am working on is finally doing what I want. That is to get vastly different luminosities in the same eight bit image. As you can see the main stars of Crux are not blown out and have colour. The faint or dim dust in the Coalsack is showing it's underlying structure. All other faint stars have colour. A faint PN is also well defined.
I have to post large images otherwise jpg artefacts are all that you will see.

Full res image here 10MB.
http://d1355990.i49.quadrahosting.com.au/2009_09/cruxcoalmos08.jpg

The method is not perfect but I am getting closer. If I can get this sort of result with a twelve bit DSLR, imagine what a top astro CCD would produce.

Bert

[sheeny]

That looks superb, Bert!

Al.

[bojan]

Very nice :-)
How did you match the luminance of individual images, so they blended so smoothly?
I tried to do something similar earlier this year, but the results I obtained were.. well, way below any acceptable level :mad2:
So I simply gave up.

[avandonk]

RegiStar has a calibrate function where all other images can be matched to the starting image as long as there is enough overlap to get a meaningful corellation.

Bert

[bojan]

Thanks, Bert :-)
I will try this again on first opportunity..

[avandonk]

Bojan Calibrate is under Operations or F7. You may need to trim as any gradients caused by vignetting tends to play havoc with the colour matching.

Bert

[erick]

Nice resolution on the main stars, Bert - doubles visible. So often they are blown out in such images. That highres image is stunning for a widefield. :thumbsup:

[gregbradley]

Yes you've got that HDR technique down pat now.
Quite a breakthrough and a very lovely image.

Greg.

[multiweb]

Nice one Bert! :thumbsup: I had problem getting the high-res version. Is is still there?

[DavidU]

A real beauty Bert. This HDR is comming along well.

[avandonk]

Try again as there has been some sort of major internet pipe breakdown today. I could not access it myself. Yet it was there????


Nert

[avandonk]

I stlll think I have further to go Greg. But the latest results are very promising. I have to gather my ideas on what is really happening.

EasyHDR uses human visual logarithmic sensitivity as the basis for their tone mapping algorithm. The tricky bit is getting the settings correct for the available data. The adjustment controls are not independant and there is no 'correct' set of settings. These settings depend on the starting images in both their content and collection parameters.

All I can do is keep practising!

Bert

[DavidU]

Yep, Hi res comming up as 10026px × 7890px

[Hagar]

Very nice Bert. Starting to get some real depth in the mosaic. Previous images often left me with a washed out feeling but this is sharp and clean.
Well done.

[avandonk]

Thanks Hagar. We are all chasing perfection. We will never get there. It should not stop us from trying though.

I was overdoing the dynamic range compression. I have repented my sins.

I still feel like I am groping in the dark with one hand tied behind my back.

Without the interaction here I would still be down a blind alley!

Regards bert

[Matt Wastell]

Stunning image Bert!
I found it to be quite captivating - it really held my eye for a long time - a real success in my book!

[avandonk]

On another note Peter Ward was correct all this time. He said my stars were square (at 300%) and he was correct. Like most people I thought a bit of slack autoguiding is adequate for a 300mm optic. By guiding as if I was recording data with a 1200mm FL optic with the 300mm optic and dithering I have been able to enhance the resolution merely by upsizing and median stacking.

The evidence is in the image.

Bert