Derekku wrote:
ykcarrot wrote:
I added 3rd offset. I feel very wrong at after second half of the song.
Lybydose wrote:
Put a red section here at 105,815 offset.
I disagree with both of these. I tested a few more times, and feel that just two offsets work best, although I tweaked my originals by a few ms each. 2,990 and 25,818 should work nicely, and I didn't get more than 5ms early or late while testing. Remember to resnap on both diffs and adjust your inherited sections if need be.
Basing timing entirely off the "Mistimed Hits" in test mode isn't the most accurate way to fix timing. It fluctuates too much from human error and personal whims: if I'm utterly convinced that -10ms is correct, I subconsciously play 10ms earlier and validate my claim by looking at the mistimed hits. I can then play again and, without even changing the offset, hit the same thing 10ms late. In both situations I've fooled myself into thinking I'm hitting the notes with 100% accuracy. I could play for hours and get different results each time.
I time things by ear listening to the metronome at 25% speed. It's much more accurate because I get the same results every time I do it. However, even that is subject to human error at times.
In this case, an even better way to do it would be to use a program like FL Studio to analyze the waveform of the mp3 itself, because they I
definitely get the exact same results every time.
First though, I need to use osu! itself to find the timing of one note so that it's possible to properly sync up one note between osu! and the external program. Everything else follows from that guide note (in other words, if my guide note is 5ms off, everything else I find will be the same 5ms off, etc). The timing section at 25,818 sounds pretty good to me, so I'll use that. Since this "guide note" is based off Derekku's timing anyway, this means that everything else I find in the song will ultimately be based off his timing.
I then
further confirmed those results by re-timing by ear in the editor and seeing if they matched up.
Timing Results:Unfortunately, there are no drums at the very beginning, so I can't really get much out of that. I think the 2,980 it's set to right now is early though. It should probably be around 2,990-2,995
2nd offset: 25,818
3rd offset: 74,380
4th offset: 194,374
All 84.000 BPM
