Uhm, where did you get this info from?
It is able to contain the data of fps both lowest, highest if it’s VFR, and constant if its constant - it’s not added manually with -metadata. It simply IS metadata which the container is able to show without any prompts. You can set -fps_mode vfr, cfr or passthrough. Setting vfr is basically passthrough - or not setting fps_mode at all in ffmpeg. Either way. frame rate will be reported to the file.
Hell, it can show you exactly how many frames are in the video (although, overhead can give a false exact framerate by a couple or promille from what I’ve heard about other’s findings).
If that’s the case, who/what is right? Is the info provided by MediaInfo accurate? Is the VFR issue solved? Does TVAI add something to the exports that shouldn’t that tricks/confuses MediaInfo?
I am not a master of this field as some of you are, but I am just reporting what some of the available “master” tools report. By pure luck I discovered this VFR issue and reported it first, from what I know. After that, the hell’s gates broke open. So in a way I feel like I have to contribute on behalf of those who don’t have this deep knowledge like some of you do.
I am willing to do test beta but I wouldn’t use the “release” version for my main work until it’s confirmed that the issue is solved.
As you can see in a post I made higher up running mpdecimate on different models and containers. Which shows similarly to your findings.
Interpolation had frames that were identical down to the exact pixel. However, not many enough to cause any kind of noticeable stuttering, unless perhaps you’re running ultra slowmo at fps not high enough.
I too, will continue using 5.3.6 as I personally dislike the new GUI. It requires a ton more clicks to get the same thing done, and I don’t use hyperion or rhea xl anyway.
