Assumptions that programmers make that are wrong: Videos have a constant single resolution over their duration.
@karolherbst that's a new one to me

@kate I think it's a new one to everybody except the one person who thought filing bugs for this is a fun thing to do :)

https://mastodon.social/@nekohayo/116767161957388450

@karolherbst oh heh i thought that was some kind of secret knowledge of people dealing with media stuff
@karolherbst @kate Nah, this has legitimate uses in broadcast TV. It can occur in both the bitstream and containers.
@kanongil i can't imagine any broadcaster taking the chance at changing a stream and crashing a bunch of crappy TV firmware.
@karolherbst @kate i vaguely remember there have been some security bugs in video encoding/decoding caused by this feature
@karolherbst assumptions that video format standards committees make that are wrong: changing the resolution mid-stream will be a desirable, useful, and well-implemented feature.
@rotopenguin you won't believe it, but there are good uses cases for it 🙃
@karolherbst @rotopenguin same as variable bitrates for sound I would assume?

@borup @rotopenguin mixed content for streaming, but also.. what if the network is unstable and you want to adjust it to send lower/higher quality midstream?

What if the device is capable of changing orientation, e.g. a phone going from portrait to landscape mid stream?