@linmob Is there any progress on sound in videos?

@ati1 Sorry, what do you mean?

Video recording is still experimental on mainline in general. I recorded a short clip in the dark (on a libcamera stack and with a person in it, won't share) and colors are off and it's upside down, but it has sound.

@linmob I meant Librem 5. Good to hear things are moving in the right direction.

@ati1 Thanks for the answer! I just tested with my Librem 5 (postmarketOS v25.06, frankensteined with Millipixels 0.23.0 (0.22.0 ships with the distro), and it records sound and it includes it in the video - it does not sound great though (some hissing and other artefacts), but I could make out what I had said.

@linmob @ati1 millipixels main limitation will be video length. Im working for solution with unlimited length, see tui/ccam.
@linmob @ati1 Correct. c18db48824e5 should be good version. Basics are right (debayer + processing on GPU), but there's plenty of work to do.

@pavel Thank you! Since you're also using a patched kernel, did you try it with megapixels2 yet?

@linmob
@martijnbraam Megapixels video support is newer version of Millipixels video support. Same limits there. Once 2.0 is released, I may want to update to recent scripts versions, but don't expect miracles there. It would be hard to integrate debayer + ISP onto GPU with Megapixels. You could put GPU ISP into separate process; that's bwtest in debayer-gpu repository. That should be workable, but... Clicks Machine should be superior solution.
@linmob @martijnbraam (And yes, you could / should modify Clicks Machine so that it can do debayer + ISP on files on demand.) (And yes, eventually this should be all integrated into libcamera).

@pavel Thanks, I was just thinking about photos right now. I'll play with the config and see if I can fix the colors :-) (Assuming I can figure out where Flatpak puts the config files ;-) ) @martijnbraam

@linmob @martijnbraam Yep, sensor calibration can be a lot of fun, too. There's enough work to be done all over the place. (And first step -- linearize data from sensor -- is something I'll need to investigate for Clicks Machine).

@pavel @martijnbraam @linmob I was looking at these results some time ago and they turned out to be nothing but clipped data and lack of data pedestal handling.

@dos @martijnbraam @linmob Do you have some data to back up that linearity claim? I believe there's consensus that it will not be linear near saturation. OTOH I don't believe there's consensus of how important that effect is :-(. https://www.mdpi.com/1424-8220/24/6/1841

@pavel @martijnbraam @linmob After fixing Megapixels' tool, the result is linear up until the image starts clipping (at around 0.6 in this case). Since that tool averages the values across the whole frame, what you're seeing above that is lens shading.

@pavel @martijnbraam @linmob When you take only a few pixels from the center (with G channel clipped at the very top):

@dos @martijnbraam @linmob So... you are right only center pixels should be considered. And it surely looks like it is linear from 0 to 95% or so. I assume there would be similar effect at the 5% near saturation on red and blue ... and that surely is non-linearity, but may not be important enough to be worth correcting.

@pavel @martijnbraam @linmob No, that's clipping. My light source was 5% too bright, so you can see that it stays linear up until it clips.

@pavel @martijnbraam @linmob Same thing, but brighter.

While there will be some tiny non-linearity if you look really close (there are no perfect sensors after all), it's absolutely nowhere even close to 5% and not really something you have to concern yourself with for photography :P

@dos @martijnbraam @linmob Aha, thanks for this. Agreed, there's nonlinearity at the top, but we should not need to correct for it as it is really small.
Follow

@pavel @martijnbraam @linmob You would need a more rigorous setup to actually see that non-linearity, what's here is still more of an artifact from averaging bunch of pixels. When you take just a single one, it's below the noise floor.

@dos @martijnbraam @linmob Yes, agreed, effect may be from differences between the pixels. If it was bigger, it may be worth it to try to work with it, but it is likely too small to exploit.
@dos @martijnbraam @linmob Martijn, if you agree with the analysis, would it be worth it to somehow "retract" that blog post / publish updated one?
Sign in to participate in the conversation
Librem Social

Librem Social is an opt-in public network. Messages are shared under Creative Commons BY-SA 4.0 license terms. Policy.

Stay safe. Please abide by our code of conduct.

(Source code)

image/svg+xml Librem Chat image/svg+xml