... it's worse than your datacenter.
Latest update: 2021-07-11
I recently found out that mpv was stuttering when i used it on my old Intel Pineview Atom box because the normalisation audio filter munched all the CPU time. I wondered if i could load a profile according to the CPU used, and finally managed to do it.
I thought that it would be integrated with the
but it did not work at all.
So at first,
i defined options for the default profile,
then the Atom specific profile,
with youtube-dl options as a goodie:
# Better youtube quality on higher specced machines ytdl-format=best[height<=?1080] # Normalize the audio output af=loudnorm=I=-15 [atom] # Reduce youtube quality on that slow Atom box ytdl-format=best[height<=?360] # And clear the normalization audio filter (it uses a lot of cpu) af=
mpv has lua scripting capabilities that will be required to make the profile autoloading work. As expected, if the CPU model is an Atom, it will apply the "atom" profile.
$ cd ~/.config/mpv $ mkdir scripts $ cat > scripts/autoprofile.lua h = io.popen("sysctl hw.model") v = h:read("*a") h:close() if string.find(v, "Atom", 1, true) ~= nil then mp.commandv("apply-profile", "atom") end ^D
You can check out if everything is correct with
mpv -v [...]. Enjoy :)