Ancestors

Written by Mallory's Musings & Mischief on 2025-01-26 at 17:05

For my fellow #astrophotography #astronomy #FOSS nerds:

There's a project to make a touch interface/app for NINA! It looks pretty functional right now, but there's a lot of progress yet to be made. I cannot wait until the real-time stacking plugin gets integrated with this.

I love ASIair, and it works really well, but I'm not a fan of closed proprietary platforms and would love to have an alternative.

https://youtu.be/NMCeTYhAyuw

=> More informations about this toot | More toots from malcircuit@thingy.social

Written by Mallory's Musings & Mischief on 2025-01-26 at 17:10

Speaking of NINA, I find it incredibly obnoxious that it's Windows only software. Not only does Windows increasingly suck from AI enshitification, Linux has much lower system requirements. I shouldn't need 16 GB of RAM to operate my fucking telescope.

=> More informations about this toot | More toots from malcircuit@thingy.social

Toot

Written by Mallory's Musings & Mischief on 2025-01-26 at 17:34

Another project I've been thinking about is an FPGA-based system for capturing images, guiding, plate solving, and stacking. Like, FPGA talking USB directly to the camera, no PC at all.

To be honest, it's probably not a great idea, but I think it would be neat. It would probably make more sense to use a small embedded platform with a GPU and use that for image processing acceleration.

I often wonder if the ASIair leverages any GPU acceleration. Live stacking has always felt too slow.

=> More informations about this toot | More toots from malcircuit@thingy.social

Descendants

Written by Mallory's Musings & Mischief on 2025-01-26 at 17:51

A big reason I think about using an FPGA is to get the maximum capture framerate possible. For DSO imaging, this is basically irrelevant, but for planetary imaging capture framerate is very important. The more frames you can feed into a lucky imaging stacker, the better the end result is. There's always a tradeoff between resolution (pixel or bit depth) and framerate, limited by the maximum throughput the system can handle.

=> More informations about this toot | More toots from malcircuit@thingy.social

Written by Mallory's Musings & Mischief on 2025-01-26 at 18:00

A modern Intel CPU has crazy throughput, but only on moderately powerful systems. So basically that means a mid to high-end mini PC, desktop, or laptop.

But throughput is what FPGAs excel at. I would expect it would be possible to have an FPGA capture system that could perform as well as (or better than) a PC, but with lower power requirements and possibly cost.

That's the idea, anyway. I wish I had the motivation, energy, and focus to actually try it.

=> More informations about this toot | More toots from malcircuit@thingy.social

Written by Mallory's Musings & Mischief on 2025-01-26 at 18:22

Plate solving, in particular, is of interest to me. I wonder if GPU or FPGA acceleration could be applied to the problem. If there's a way to parallelize plate solving it'd be dumb and slow to do it with a single CPU thread. I don't know enough about algorithms to know if it's possible.

=> More informations about this toot | More toots from malcircuit@thingy.social

Written by Medea Vanamonde on 2025-01-26 at 18:04

@malcircuit talk astro foto tech to me…

=> More informations about this toot | More toots from MedeaVanamonde@chaosfem.tw