#hobby update
Battling today’s challenges, whether is #greenland or #eu or #us or … list goes on. One needs a break.
Yeah; so, I like challenging stuff. It makes me tick. Some (not many 😅) have asked “how”, okay, so … nerd stuff (short; see comments for the full story and hi-res);
Today I had an intensive processing day working on my Orion mosaic with M42 and the Horsehead (B33). The main goal was to maintain maximum control over the dynamic range of M42, so that the Trapezium core would remain well preserved while still revealing the much fainter surrounding nebular structures.
For M42 I therefore deliberately went fully HDR. For Ha, OIII and Tri-RGB I used multiple exposure times and first combined them per channel, panel-locally, into HDR masters (HDRComposition). This resulted in three solid base images for M42: HDR Ha, HDR OIII and HDR RGB (64-bit), each with a large dynamic range and no clipping in the core.
The Horsehead panel was less critical in terms of dynamic range, but was built with the same care in Ha, OIII and RGB so that both panels would match well in content and appearance.
I then processed each panel completely separately in PixInsight. Both panels went individually through the full workflow:
In PixInsight, for all Ha/OIII/RGB data: BlurX, STF, StarX, Curves; and for the RGB set specifically: SPFC, MGC, SPCC, HDRMT and LHE, Curves, etc. (Yah; easy man)
After that, PixelMath was used (controlled blending of Ha and OIII into RGB), resulting in two fully finished RGBHOO panels, processed independently. After this, the stars were added back, the images were solved again, and finally GradientMergeMosaic was applied, followed by another ABE and the corrections that were needed.
With respect to PixelMath, I approached it as follows: processed, starless masters for Ha, OIII and RGB. From the RGB image, a Luminance image was extracted. The factor (1 − L) suppresses Ha/OIII in the bright core; this was also the reason for not using a simple MAX operation or similar when recombining the stars.
\(Red: $T[0] + M42_Ha * 0.6 * (1 - M42_B33_Mosaic_L) Green: $T[1] + M42_OIII * 0.22 * (1 - M42_B33_Mosaic_L) Blue: $T[2] + M42_OIII * 0.12 * (1 - M42_B33_Mosaic_L)
Once each panel was finished individually, the stars were added back and the panels were solved again.
By merging the two fully processed panels with GradientMergeMosaic, the HDR structure was much better preserved and the transition between M42 and B33 became clean and natural. Ha was suppressed again to some extent, but this was easily corrected with ABE (and can always be refined further…), followed by some S-curves mainly on red, which fortunately brought everything back in line with the individual panel result. The core of M42 is slightly less pronounced in the mosaic than in the standalone panel, but it looks natural and well balanced with the rest of the image.
Enjoy.
#astrophotography #space
