Timing window?

[Finalraver]

does anyone know the timing windows for perfect?

[Satoru]

You generally hvea to hit it on the beat. If you're having issues getting perfect you might want to re-do the audio and button calibrations

Also the end screen does tell you on a general level if you are hitting the notes too early or too late. This might give you some insight as well

[Cheggf]

Originally posted by Satoru:

You generally hvea to hit it on the beat. If you're having issues getting perfect you might want to re-do the audio and button calibrations

Not his question.

Originally posted by Finalraver:

does anyone know the timing windows for perfect?

Super perfects are ±7.5ms & normal perfects are ±37.5ms.

[Arras]

There's super perfects?

[Cheggf]

Originally posted by Arras:

There's super perfects?

Well the game never calls them super perfects so you can call 'em whatever you like, but if you turn on early/late hit indicators you can see E PERFECT & L PERFECT. Perfects are worth 555 points, and that gets multiplied by the multiplier. You get an extra bonus point if you're within 7.5ms so there's no early/late indicator, and another extra bonus point if you're frame perfect. But those extra 1 or 2 points don't get multiplied by the multiplier, they're just there for tiebreakers. So really there's three perfects.

So a perfect is worth 555 at 1x, and 2,220 at 4x, while the other perfects are 556 & 557 at 1x and 2,221 & 2,222 at 4x.

[Cheggf]

I would like to correct previous information. I had seen it repeated several times so I assumed it was accurate, but I've learned what seems to be the correct information. The window is not a static thing but depends on BPM, and each timing is a fraction of a beat.

Rank	Time	Score
MISS	>0.5	0
OK	0.5	111
GOOD	0.35	222
GREAT	0.25	333
E/L PERFECT	0.1	555
PERFECT	0.05	555+1
TRUE PERFECT	0.01	555+2

[Cheggf]

Hey guys. It's me again. The person who posted those previous numbers seemed to have done thorough testing so I posted them assuming they were correct, but they've done even further testing and think that they were wrong. It might have worked differently in the demo which caused the incorrect information.

I'll just post what they said verbatim:

Originally posted by Katie:

I have redecided that my numbers are wrong and it is ms(ish) , it just turns out that accuracy is rounded so aggressively that for like 110bpm to 150bpm the buckets they end up in are identical to the actual maths +/- frame discrepencies in almost all cases. (the offset gets swapped around between beats and seconds several times and stuff gets casted to ints for no reason in this bit of the code which is why it was hard to be sure)
It's also really hard to play while intentionally hit super early and super late.
I'm pretty sure the windows are supposed to be

MS > 175.0ms (175*1.00)
OK > 122.5ms (175*0.70)
GD > 087.5ms (175*0.05)
GR > 035.0ms (175*0.20)
EL > 017.5ms (175*0.10)
PF > 003.5ms (175*0.02)
TP > 000.0ms (175*0.00)

but in testing True Perfects happen at +/-4ms and Perfects happen at 18.5ms, but thats probably just a >= vs > and rounding

They also investigated the minigames for me when I asked.

minigames are

MS > 250.0ms
OK > 237.5ms
GD > 200.0ms
GR > 150.0ms
EL > 062.5ms
PF > 025.0ms

although I dont think they use the OK or PF windows

I also asked about boss fights and they said "240ms probably, but they might still be bpm based".

[Nauskills]

That's some beautiful data, thank you Cheggf