HDR Workflow for SMC

[Ryderscope]

I would appreciate some advice from this forum on how to approach the workflow and processing of a data set that I have captured of the Small Magellanic Cloud (SMC). The data set has been captured using my Canon F2.8, 200mm prime lens and QSI 683 camera and includes the following subs:
- L, R, G, B = 20 x 30s
- Ha  = 20 x 90s
- L, R, G, B = 20 x 300s
- Ha = 20 x 900s
The field of view of the image includes the SMC and the globular clusters NGC 104 and NGC 362. Additional short exposure subs were captured to support a HDR combination to fix saturation of the core of NGC 104.

The workflow process that I am considering is as follows. Any suggestions or alternative ideas would be appreciated.

1. Run the raw subs through the batch processing script (with darks, flats bias) to produce linear masters thus:
30 second master for each of L, R, G, B
90 second master for Ha
300 second master for each of L, R, G, B
900 second master for Ha

2. Colour combine the individual R, G and B masters to give:
30 second master for RGB
300 second master for RGB
We now have linear masters thus:
L 30s / L 300s / RGB 30s / RGB 300s/ Ha 90s  / Ha 900s

3. Execute the HDR combo function thus:
L 30s + L 300s ->  Lum HDR
RGB 30s + RGB 300s ->  RGB HDR
Ha 90s + Ha 900s -> Ha HDR

4. Process these masters now with following steps:
DBE/Colour Calibration/Noise Reduction/Histogram Stretch/Curves/Sharpen/Obsessive fiddling until we are happy
We now have non linear HDR masters for:
Lum
RGB
Ha

5.  Add Ha to RGB using NBRGB script
We now have masters for:
HaRGB
Lum

6. Channel combine using LRGB tool

7. More obsessive fiddling to produce final image

Feedback on this workflow would be appreciated to make sure that I am on the right track.

[RickS]

Seems like a reasonable approach, Rodney.  Not sure you really need HDR on this target but maybe the stars saturate more quickly in the DSLR, especially since your image scale is large?  I was working on this target last weekend and got good stars without HDR (AP140EDF scope and Apogee U16M camera, image scale of 2.4 arcsec/pixel.)

One simple trick you might want to try for the Ha combine is to clip it a little to make the background black.  Apply the modified Ha as a mask on your RGB or LRGB image and use curves to boost the red.

Cheers,
Rick.

[Ryderscope]

Thanks Rick. I'm using the QSI 683 CCD which has limited well depth so bright objects such as the core of NGC 104 (47 Tuc) tend to max out rather quickly. Your idea of using the Ha image as a mask is a good one that I will give a go. Will be interesting to see how it comes out.

Rodney W.

Seems like a reasonable approach, Rodney.  Not sure you really need HDR on this target but maybe the stars saturate more quickly in the DSLR, especially since your image scale is large?  I was working on this target last weekend and got good stars without HDR (AP140EDF scope and Apogee U16M camera, image scale of 2.4 arcsec/pixel.)

One simple trick you might want to try for the Ha combine is to clip it a little to make the background black.  Apply the modified Ha as a mask on your RGB or LRGB image and use curves to boost the red.

Cheers,
Rick.

[RickS]

Sorry, Rodney.  I can't have read your post too carefully.  I must have just seen "Canon" and got DSLR stuck in my head    I'd be interested to hear how you go.

Cheers,
Rick.

[Ryderscope]

And to add a more specific question to my post ... I note from the video tutorials from Harry Page that the HDR combo process is executed whilst still in the linear stage. I have also seen posts that indicate that it is ok to do the HDR combo in the non linear stage as well. My process above has the HDR combo happening before I have stretched the data.

What is the general consensus then? Should I be doing the HDR combo at the linear stage or after I have stretched the data? Does it make a difference either way?

Thanks in advance.
Rodney W.

[RickS]

What is the general consensus then? Should I be doing the HDR combo at the linear stage or after I have stretched the data? Does it make a difference either way?

Given the sort of analysis and processing that is done by HDRComposition I'm pretty sure it is intended only for use on linear data.

Cheers,
Rick.