>>106914135
This was on #mpv-devel
The proof is concrete
<llyyr> that one is not right
<llyyr> this means that when we delinearize, we need to use the inverse of the display or output EOTF, it has no relation to the input file
<kasper93> delinearize before sending to display is clear to use display characteristics
<kasper93> the issue is how to linearize, using also display to both have round trip and display appearance in linear space and follow the spec and you piece wise to use "optimized encoding"
<llyyr> yeah I'm just clarifying what I said to you on discord yesterday, I was explaining stuff under a misunderstanding