alt Hacker NewsH.264 isn't efficient though. AV1 (and even H.265) are anywhere from 20-50% more space efficient for the same visual quality.
AAC vs Opus is a little bit more debated, but on average should outperform AAC.
Replies
So? File size matters for delivery formats. But not for intermediate formats straight out of camera. For editing, bitrate and decode performance are often way more important.
Hardware decode support for AV1 and H265 is still pretty bad. If you use those formats, a lot of computers will stutter when you move the play head around while editing. On my mirrorless cameras, I much prefer filming in medium to high bitrate h264. I fill up my storage cards faster, but my camera batteries last longer recording like that and the footage is much easier to work with. I have a 4090, which supports h265 decode. But it only works for 4:2:0 footage. If I film in h265 422, my computer falls back to a software decoder. It's easier to just use h264, then I don't have to think about it.
If you want to ship high quality deliverables, you don't want resolve to do the final encode anyway. Resolve tries to use NVENC for encode, which is a much worse encoder than x264 and friends in ffmpeg. From resolve, output prores or something. Then use ffmpeg to do the final encode. I think there might also be ffmpeg encoding plugins.
tldr; Codec matters, but there's a lot more nuance to getting a good image than just picking the newest video codecs.
The problem with AV1 is it requires a massive amount of compute resources. This affects both recording (encoding) and playback (decoding) and certainly affects editing. My AMD RX 480 (8 GB) card doesn't even support AV1 encoding too.
Being able to edit without lag is really important, also keeping your CPU load reasonably low when recording is important to avoid dropped frames, or your recording streams becoming out of sync (ie. a webcam video with audio recorded from another input device while you record your desktop).
I noticed on Linux, if I do anything CPU or disk intensive while recording (even with an SSD), I often see these things get out of sync at the point of load. It's an easy fix usually by just cutting the video where it starts to get out of sync and shifting the audio track over ~200-300ms, but it's something that could happen many times so it becomes an issue in the end. I recorded thousands of videos on Windows and never had this problem but it happens every few videos on Linux just doing casual screenshare videos.
AAC vs Opus is similar, except audio related.
In general when recording videos I usually use x264 as mkv which is CPU bound because it produces a much better looking video vs using h264. That's using a bit rate and other settings that tries to balance quality with low disk space.