The weather has been pretty good for the last month so in between roof painting, I've had a chance to take some shots of the area of the sky around Crux and Carina. First is a wide one, with the LMC just showing up at the top of the frame and Orion on the left: Subscribe to see EXIF info for this image (if available) This was shot at 7.5mm FE, f/3.5, 10s, ISO1600, with a 13 stack for the top half, 12 stack for the bottom half, aligned and reprojected in Hugin, then manually blended and processed in Photoshop. I decided to try and do manual dark frames this time around, whereas previously I used to stack light frames with built-in dark frame subtraction enabled. The initial stacks from this sequence were a complete disaster! I discovered that for super wide field shots with the fisheye at least, DeepSkyStacker just craps out if fed with the RAW light frames and dark frames for some reason. The result was noisy as hell and had trails of hot pixels everywhere. In the end I did some hefty initial processing, noise reduction, and manual hot pixel removal with ACR on just the light frames and stacked the exported DNGs for this salvaged result. Next is a close-up with the Carina Nebula and Coalsack prominently visible: Subscribe to see EXIF info for this image (if available) This one was shot at 25mm, f/1.4, 4s, ISO1600, with an 18 stack of light frames and dark frames. For this narrow FoV, DeepSkyStacker did well when fed with all the RAWs. I really should have stopped down a bit more, the coma at the corners is a bit too noticeable... Finally, a trial run on a different night with my new PL15: Subscribe to see EXIF info for this image (if available) This one was shot at 15mm, f/1.7, 8s, ISO400, with a 6 stack of light frames (automatic dark frame subtraction enabled). It's not quite wide enough to be this close to the foreground (a pretty poor foreground at that), but I think it will be workable in other scenarios. After this experience, I don't think I'll bother with shooting manual dark frames - I feel that the camera actually does a better job at this as each dark frame is taken with the same conditions as the corresponding light frame and perfectly aligned. Moreover, if anything screws up, I'll have more salvageable light frames with dark frame subtraction enabled than without.