Next semester is starting up soon, and one of the most useful things I've taken to doing is recording lectures. I doubt I'll ever be able to get the quality anywhere that would be considered 'good', but I have no idea what I'm doing outside of some trial and error, so I imagine I can probably get it quite a bit better with some optimizations.
Set up... recording with a laptop, and a Ugreen 15728 to capture both video and audio. Everything's done on OBS Studio. I sit at the front of the class with the webcam mounted to a little tripod and aimed at the projector screen with the end goal of basically having narrated powerpoint presentations.
The field of view of the webcam is way higher than I need it to be, so I zoom it in by fitting just the presentation to frame in OBS... which works, but probably 75% of its resolution potential is lost to just cropping it out. I don't think anything can be done about that, since the webcam doesn't have a physical zoom and I can't position it anywhere closer to the screen than I have previously. ...and the resolution as-is comes out... okay. They're powerpoints, and with the exception of the odd bit of unusually fine font, the final product is readable, in a 1995 jpeg kinda way.
I think the real room for improvement is managing the lighting settings - the issue is that the camera is still recording the bits that aren't actually being saved, so if the room lights are on and the ppt has a dark background, it gets 'corrected' to where the light font on the page also get kinda blacked out, I think because it's compensating for the bright surfaces beyond the edges of the presentation. Room lights off with a light background ppt, it all gets white-washed. Tweaking the gamma and... backlight? whitelight? in the device-specific settings I can usually get it to where the text is readable, but again, it kinda looks like shit.
Audio... I used to have some of those interview mics for the prof to wear, which were great quality for the... like two or three? times they actually decided to wear it. They were for some reason opposed to doing that, so they'd just set it on the podium and let it record from there... and then walk all over the front of the classroom, making audio quality all over the place. Tried using the mic on my gaming headset, with the entire headset perched next to the webcam, but it captured me breathing and typing more clearly than the prof; admitted defeat on audio quality and have since used the ones built into the webcam, and they do okay, but with that same kind of barely passable quality as the video.
I've seen some optimization guides, but pretty much all of them are geared toward streaming or web conferences, not lecture halls, and their tips only yielded slight improvements.
I don't really understand what any of the settings actually do, so I just fuck with the sliders or check the toggles and if it makes it better I keep it, if it doesn't I put it back... but I'm probably doing something in a stupid order or overlooking a whole category of settings. Literally just figuring this all out as I go - no prior experience with this or similar software.
...I know my recording conditions and equipment are all far from ideal, but I do share these recordings with my classmates, so if any of know any optimizations I can make with the tools on hand, please let me know!
Thanks all!
Check out adding video filters too; compressor, equalizer and gain may help. There’s a bunch more on github, unfortunately I can’t think of anything that adapts to all the situations you described, but you could have more than one instance of your capture device set up, one for room lighting on and one for off.