Stacking vs Single Frame
I have seen many lunar images where the author lists the processing as being stacked from a number (often hundreds) of images. This would be achieved using a high frame rate camera (or video camera) and stacking software like Registax. Perhaps I tend to review these on social media posts where the quality of the images is reduced, but I have to admit I am frequently underwhelmed by stacked images. They take on a distinctive blurriness or over-processed appearance that I just don’t like. However, recently I tried the technique myself and was impressed with my own effort (if I do say so myself). So going forward, should I stack or not?
A barlowed single frame ‘lucky shot’ (click to expand)
Until this point, for lunar photography I have been DSLR all the way. My technique would involve getting the camera in focus, either as prime photography or using a 2x barlow, and then taking maybe 50 RAW shots with the hope of finding 5 good ones. As the atmosphere swims around the images, craters appear to bulge in or out in successive frames, but if you throw enough mud at the wall you get a couple of still, crisp frames. You might get one ‘lucky shot’ where the atmosphere behaves for the 1/200th of a second you depress the shutter. This is harder with the barlow but worth it if you can pick out more detail. You have to react to your conditions. Like with visual astronomy, if you push the magnification too far the results get progressively worse.
2 pane composite of sharp top and bottom
I have also done a couple of cut n’ shut images, as the above technique rarely gives a crisp view across the whole frame. Of your 5 best, you might get a good top and and a good bottom half, so stitching them together gives the sharp overall image. The difficulty with this is achieving an invisible join due to the distortions made by the atmosphere. This takes a while and can be a fun-sapper.
Combining frames may be a necessity it your focal length is such that you cannot frame the whole moon in one shot. From here you can extrapolate to full mosaic images where you might take multiple images of 10-15 regions, selecting the best for each region then matching all regions together. This is a real labour of love when done manually.
30 x 0.01sc exposures stacked in Registax
I tried the stacking technique by accident. I was doing a test on my newly collimated RC scope and CMOS camera. Using the shortest exposure of 0.01sec I turned it onto the moon and ran off 100 images BIN 2. I then ran off the same preview on BIN 1, which even from the preview I could tell was better much crisper. I scrapped the BIN2s and ran off 30 at BIN 1.
I knew about Registax but had never used it before. I use a Mac running boot camp so I downloaded a copy and clicked my way through the intuitive process of alignment and stacking using all the default settings. Despite my inexperience, within half an hour I had an aligned and stacked .tif file that I took into Adobe.
Processing was mainly done in Lightroom. I opening it first in Photoshop and saved as a .PSD. I couldn’t immediately find the adjustment I was looking for so jumped into Lightroom and played with the clarity slider, the sharpening and noise reduction filters. I have never bothered to learn what these sliders do, I just play with them and review the preview; better/worse, without making them appear artificial or overly adjusted.
Conclusions…so far
I think that stacking could be the way to go, although this apparent improvement may be down to the quality of the new camera, shorter exposures (meaning less exposure to the wobble) and stacking using software rather than manual alignment. I think I next need to try detail using the 2x barlow or a full mosaic. This will tell me if the whole process is as seamless and where the sharpness can be achieved with higher magnification. There is greater uniformly and sharpness across the whole images, no fringing or apparent aberrations that the DSLR used to give me. The stacking process doesn’t beat the magic of getting one awesome shot, but should ensure a greater consistency and enable me to try other projects like images with multiple moon phases or animations. I hope to return to this within the next few weeks with an update and more images.
May 2019 update….
I’ve tried some more lunar images using this stacking technique but have struggled a little. I’ve taken set of images over successive nights in FITS format on the ASIair but have not had great success with Registax, I think because of how it debayers the files. The final stacks come out very poor quality or in random colours. Indeed Registax only seems to recognise my (colour) FITS as B&W and although it has debayer prefilters, none give a preview that looks like my image. Its a little frustrating as I want to find a successful way to image the moon during the summer while other objects are too faint in the twilight. Given how easy my first image was, everything since has been hard work - I must be doing it wrong! I have even tried converting FITS to TIFF (also not easy - I found a way through APP) but Registax only recognises 8bit not 16bit TIFF…another lesson learned. Registax also had a habit of crashing during the stacking process, so after a few hours I had to give up for my sanity.
March 2020 update….
Since I’ve returned to lunar imaging on my new OMC-140 MakCass, I’ve given Registax6 another go and taken some effort to learn how the programme works. This involved watching YouTube videos and no small amount of trial & error. With this new knowledge, stacking does seem to be working. I’ve had it pointed out to me that stacking can increase the sharpness of the images, but mainly reduces the noise (like with deep sky imaging). As the Moon has its own texture, noise is less intrusive and a sharp, single frame can therefore still look great.
Stacked and then processed in Photoshop
Registax Rules:
I’ve learned now to do the following:
if I’m imaging a crater, run off 100-200 frames
stack in Registax using c.300 align points (slider to the right for ‘strongest’)
limit the stack to maybe only the best 25 frames (use the stack chart to show how many reach the desired quality)
use Wavelets to sharpen up detail
Next stages:
further experiments needed with colour .FITS files and how Registax debayers the images.
further experimenting with the ‘right hand tools’ in Wavelets.
get in practice before the planets re-ermerge later in 2020. I’d love to capture some surface detail on Mars & Jupiter