Closed smoogipoo closed 2 years ago
for whatever it's worth i tried to look into this one yesterday, and while part of the issue is related to the slider velocity bindable range being too small, there seems to be something else in the conversion algorithms that seems to yield incorrect results.
Do you know if mania was at least fixed? Taiko is especially flimsy and must rely on bringing about FP error under magnitudes of 1E-6 to work correctly. If mania was still broken then that would be a pretty good target to focus on rather than taiko, because I'm fairly certain the taiko calcs are as accurate as they can get... https://github.com/ppy/osu/blob/b90a5864b1726cd854c6b709e07b8d9a8ba4b432/osu.Game.Rulesets.Taiko/Beatmaps/TaikoBeatmapConverter.cs#L146-L190
unfortunately no, sorry. i got hard stuck on the taiko case and put the mania case off for later.
Do you know if mania was at least fixed?
Version 2022.306.0
@zachmanthethird I was asking for @bdach 's own changes, not something already present either on master or in a release build.
I think I might have finally found our answer; though I'm not sure if it was the original problem. Converts do not seem to be correctly referencing per-hitobject Velocity values, and thus the entire map has a constant scroll speed. This means that the "slider" in question was not complaining about a 0.01 value; instead, it's receiving a value of 1 instead of 0.1. Therefore, the range of this value may not have been the problem at all. I specifically tested with https://osu.ppy.sh/b/2169357 and https://osu.ppy.sh/b/2002882 but after coming to this conclusion, I have yet to find a converted beatmap that disproves this theory.
Discussed in https://github.com/ppy/osu/discussions/17120