HDR Composition of M42

[jdonald]

I took 20x10s for the core and stars, 10x120 for main nebula and 13x300 for the fainter nebula, when I Follow Harry's video tutorial he mentioned to use shorter exposures to fix the cored stars.  I thought 10s exposures would work but no matter what setting I use the stars are cored.

Is there something I can do?

[jdonald]

I have tried all kinds of things, debayered, different settings and nothing I do is working on the stars.  I have included my data here if anyone could help me or point me in the right direction I would appreciate it.

http://dl.dropbox.com/u/27516702/M42/M42_10x120s_20m_Iso1600_cls1_RGB_VNG-Debayer.tif
http://dl.dropbox.com/u/27516702/M42/M42_13x300s_1hr5m_Iso1600_cls1_RGB_VNG-Debayer.tif
http://dl.dropbox.com/u/27516702/M42/M42_20x10s_4m_Iso1600_cls1_RGB_VNG-DeBayer.tif

Thanks a lot,

Jeff

[vicent_peris]

Hi Jeff,

I've just downloaded your 10 sec exp image. The problem is that this image has the brighter stars completely saturated. HDRComposition cannot recover data that there isn't in your image. The best solution would be to shoot some short exposures with lower ISO (maybe 200). At 1,600 ISO you have a very little dynamic range.


Best regards,
Vicent.

[jdonald]

So should I spread the range over different ISOs? what do you suggest?

[georg.viehoever]

Add images with either adjusted ISO or shorter exposure times. Should not make much of a difference (at least with my Canon EOS40D). Important is that star cores are not saturated in the "darkest" image.
Georg

[vicent_peris]

For targets like M42, I would shoot the shorter images at lower ISO... By shooting the 10-sec exp at 200 ISO you'll have the star cores as a 1,2-sec exp at 1600 ISO.

Regards,
V.

[jdonald]

for difficult shots like Alnitak and the Horsehead, can this method be used as well?

[georg.viehoever]

Yes, of course. What HDRComposition cannot do is to remove reflections of bright stars and other artefacts caused by the optics. But as far as contrast, burnout, ... are concerned, this tool is your friend.
Georg

[jdonald]

Yes, reflections, You saw them did you?  They bother me, not sure how to remove them other than the clone tool in Lightroom.

[jdonald]

I just wanted to give everyone an update to my issue here as I am sure others will come searching.  I managed to get SOME success:


M42_HDR_ISO1600_15x300s-1hr15m_Core-40x5s_3m by AstroJeff, on Flickr

What I discovered was that my star subs were offset from the main subs, I had to run a staralignment to fix the two subs, once they were registered the HDR Process worked like a charm, however, you can see in my image the trapezium is still blown out, I can't figure out how to fix it.  My friend who used Images Plus came up with this:


JeffsM42 by AstroJeff, on Flickr

His version is more like what I intended it to be.  The good news is he used MY data, so I am running out of things to try.  By all means you can play with my data (In linear form and not registered/not Star Aligned):

https://skydrive.live.com/redir?resid=655A3B876F8D2EE6!302 (300s subs)
https://skydrive.live.com/redir?resid=655A3B876F8D2EE6!303 (5s subs)

Is it possible to get the image to look similar to my friends in PI 1.7?  What he said he did was that he stretched each image separately then combined them using two layers and a blurred luminance mask.

Thanks for everyone's support, I am learning new techniques in PI everyday!

Jeff

[georg.viehoever]

Did you try HDRMultiscaleTransform or LocalHistogramEqualization? Here is a quick attempt using StarAlignment, HDRComposition (with 0.17 as threshold), DynamicCrop, HistogramTransform, HDRMultiScaleTransform and HistogramTransform. I am sure you come closer to your friends result by tweaking the parameters for  HDRMultiScaleTransform or LocalHistogramEqualization, plus some ColorSaturation.

Georg

[jdonald]

It works a lot better I haven't heard of HEQ before, never used it, nor have I used colorsaturation, I just used Curves boosting saturation using a lightnessmask.

When I process like you have I can see hard edges on the regions that overlap.  I tried changing the smoothness which didn't help, also how did you figure out the threshold?

Thanks again

Jeff

[georg.viehoever]

Jeff,

In HDRComposition,  I used the defaults except for threshold. I dont see any hard edges. I determined the threshold by exploring the values in saturated regions (hover mouse over such a region, look at the values displayed at the lower border of the PI window). They were in the range of 0.2..0.25 (dont know why). Subtract something from this (to account for non-linearity as the sensor approaches satuaration), and you have your 0.17.

ColorSaturation is an intensity transform, just like Curves or Histogram, only specialized at adjusting color. You can get more intense red color by moving up the red end of the curve (or try shifting the color scale so red is in the center, then move up the central portion of the curve). Also try adjusting the scale parameter: You can set it up to 10 to get more dramatic (and sometimes rather artificial) color effects. The preview functionality is rather helpful.

Georg

[jdonald]

Thanks for the help again, this is what I came up with:


M42_HDR_ISO1600_Faint-15x300s-1hr15m_Core-40x5s_3m-Outer_10x120s-20m by AstroJeff, on Flickr

What do you think?

Subs Faint: 15x300s(1hr15m)
Subs Core: 40x5s(3min)
Subs:Outer: 20x120s(20m)

I added 20 120s subs for more DR.

[georg.viehoever]

Not bad. I can't see the trapecium stars however. I would prefer something where there are still visible.
Georg