Time sure flies...

8/19/2025. Time sure flies when you only take one frame every 30 seconds. I've been working on solar time-lapses. Expected and unexpected challenges kept them interesting. What you see at the link down below is actually my second try. The first was a wreck, but it suggested a workflow. I made some mistakes that I now know to avoid (for one, just leave PIPP out of it). Then came clouds for three weeks. When the weather allowed a second try, the sketched-in workflow sufficed to guide me to a decent video after hours of trial and error. A clearer-ish overview follows. Next time should be quicker and better. But first, internalize these two words to the wise: take a good flat, and tweak polar alignment. The former saves lots of time in post, and the second means you can walk away from the 'scope for at least a couple of minutes without sacrificing hard-won frames or harder-won field of view.

Here are a couple of stills. See? Things were working today:

 

With those out of the way, I configured Firecapture's "autorun" (not time-lapse) and let it run for about 40 minutes. That's 20,000 frames, more or less. Collecting data was routine; turning out a video remains adventurous.

Hardware: TMB92SS, Quark, 178MM, uncompressed, SWSA.
Software: Firecapture 2.7.x., 7ms, gain 200, gamma .65, 8-bit, binned 2x2, high speed mode. Autorun set for 200 frames every 30 seconds, repeat 180x (I stopped after 100 captures).
Post: AutoStakkart!4, ImPPG, PixInsight, Photoshop, Premiere Elements in that order.

AS!4: from each 200-frame clip generate a 40-frame TIFF image, cropped.
ImPPG: deconvolve and histogram stretch in batch mode, output as FITS.
PixInsight: convert FITS to TIFF
Photoshop: Load TIFFs into stack (align); crop; make frame by frame corrections for dust, gradients, etc. as needed; open a timeline, load frames from layers, render.
Premiere Elements: adjust speed, gamma, color, scale...

https://vimeo.com/1111181519 It's the same 3-4s clip slowed down a bit and repeated about four times; thanks Elements. After all is said and done, that's shown at 600x the speed of life, give or take about 100x.

 

There's a certain amount of handwaving in the conversion to video steps in Photoshop. I tried to keep track of what worked, but I lost the thread a couple of times and just stumbled around until I got a decent result. Next time for sure.

 

8/28/2025. I suffered an epiphany about surface post-processing. ImPPG defaults to 50 iterations of L-R deconvolution. I usually use 70, but that's just nibbling. Today, I thought to try 200 iterations and less aggressive parms. I've got a fast processor; use it. This appears to be the ticket! Try more adjustments along these lines, do a little prior file prep with this in mind (nr, mult...). The path from AS!4 to this was very short (just raw, ImPPG, raw tweaks, nr, usm). Clean this workflow up and only then go nuts.

Best 200 of 4,000
8-bit, Binned 2x2, gain 190, gamma 27, 11ms, native FL

 

8/29/2025. One day later...

Less foreshortening, a wider field of view, still 200 out of 4,000 images
350 passes of L-R deconvolution..

 

8/31/2025. Photoshop's shake reduction tool takes care of some of the remaining blur. Maybe it comes from keying the AS! files to unstable anchors. Whatever its origin, the "sharpen/shake removal" often makes a big difference (so does BlurXterminator, though BX is not as simple to invoke and not as easy to preview and fine tune). Examples to come when I get this worked out a bit better. Every parm matters, and a certain amount of fading of the full effect is often important.

:: top ::