forum

[Archived] "Only allow changes to local offset in first 10 seconds"

posted
Total Posts
53
show more
Wishy
No wiping scores because of bad modding please.
statementreply

lolcubes wrote:

Fixing offsets is easy and if the offset is bad we usually fix it already (if reported).
But yeah as I stated in my previous post, there are multibpm maps with different offsets and maps with wrong BPM which I guess need some waiting now for peppy to implement the possibility to fix BPMs as well.

I wonder how would this look like, cause if it's something like "online offset at" such things could effectively fix a lot of problems without a need to unrank & wipe scores.
Starting from xxxxxx ms, offset timing by +/- yy ms and skew timing by +/- zzz ppm
Topic Starter
plaird

merchat7 wrote:

@plaird
Timing is an objective thing. It's either right or wrong. I guess it's fine if people prefer to adjust it themselves through, but this is not up for me to decide.
Timing is objective but perception of it is subjective. The point is not that a person should be able to use their subjective perceived timing because they do not agree with an objectively correct timing, it is that the "objectively correct" timing may not be objectively correct because it is merely what another person subjectively perceived to be correct, hence why discrepancies exist at all.

This game is not like other rhythm games where all players use the same hardware and play songs timed by the same people, so I felt that the ability for players to do what felt correct to them was a good thing. But if human error were to be removed from the process then I won't really care.
TheVileOne
This thread seems like a Feature Request to me, so my response will be feature oriented.

Perhaps out of sync replays might not be the best thing actually. I certainly wouldn't want my old replay completely ruined because the BPM was changed. I also wouldn't want the scores wiped. Perhaps there would be a way that wouldn't affect existing saved replays. I think it might be worth including such a feature sometime in the distant future.

I could be wrong, but aren't the replays using the actual beatmap's object data? If you save a local replay and then change the where the objects are in the song, will that not affect how the objects appear in the replay? I haven't tested this, and plan to when I am in a position to test. If this is the case, wouldn't this make replays potentially unstable, and could be changed regardless if the user saved the replay. If this is not the case, then that means object data is saved on a per replay level. I don't see it being too difficult to add code that decides whether to overwrite or keep existing replay object data.

My Idea

A map that has BPM updated creates a replay file that goes in the beatmap folder or whereever it needs to go. Upon map update, it will inform the users that BPM has changed, and that old replays will need to use replay file to stay in sync. Use whatever method works best to distribute said replay file. Old replay formats will need to use the replay file, while newer recorded replays will not require the replay file. This could probably be done with the design of a shiny new replay format entirely or just a modified one that allows this.

The replay file will in no way affect score submission and is only used for older already processed replays.

A lot of text trying to say: Changes to replay data should be handled in a stable way, if we are going to prevent score wiping.
MillhioreF

TheVileOne wrote:

If you save a local replay and then change the where the objects are in the song, will that not affect how the objects appear in the replay?
Replays contain an md5 hash of the .osu file they were played on. If the .osu is modified in any way, then the replay and score won't show up on the modified map.
lolcubes
Online offsets don't affect the replays themselves directly I believe, so if such things can be implemented online (the bpm fix, or the "offset at" idea I described before), then you wouldn't need a score wipe, nor will the replays go out of sync.
peppy
If "20% of maps are mistimed", then please list these 20%. Or don't make up false statistics.
statementreply
20 most recently ranked beatmaps:
http://osu.ppy.sh/s/72324 too early
http://osu.ppy.sh/s/73769 either bpm off or needs multiple sections
http://osu.ppy.sh/s/58787 a bit late
http://osu.ppy.sh/s/46862 too early
http://osu.ppy.sh/s/68717 a bit early
http://osu.ppy.sh/s/63233 bpm a bit fast
http://osu.ppy.sh/s/50453 too early (already fixed online)
http://osu.ppy.sh/s/59830 bpm a bit slow, offset too late on main timing section
http://osu.ppy.sh/s/70744 a bit late
http://osu.ppy.sh/s/72805 a bit early
http://osu.ppy.sh/s/60113 a bit early
lolcubes
Small offset variations under 10ms are rankable.
As I said before, maps with bad offsets don't need to be reported here. Only if they have bad multi BPMs or just wrong BPMs. We can fix offsets already and the local offset already fixes the issue if the player wants to use it.
Sure

statementreply wrote:

20 most recently ranked beatmaps:
http://osu.ppy.sh/s/73769 either bpm off or needs multiple sections
http://osu.ppy.sh/s/63233 bpm a bit fast
http://osu.ppy.sh/s/59830 bpm a bit slow, offset too late on main timing section
Nah.. Are you sure about that BPM is wrong? I don't know how accurate you are, but I don't see any problem on those.
statementreply

ykcarrot wrote:

Nah.. Are you sure about that BPM is wrong? I don't know how accurate you are, but I don't see any problem on those.

statementreply wrote:

http://osu.ppy.sh/s/73769 either bpm off or needs multiple sections
This one is off, but I'm not quite sure about the nature of off-timing

statementreply wrote:

http://osu.ppy.sh/s/63233 bpm a bit fast
~131.996 i.e. original bpm gradually getting ~6ms early throughout the song

statementreply wrote:

http://osu.ppy.sh/s/59830 bpm a bit slow, offset too late on main timing section
~120.003 i.e. original bpm gradually getting ~6ms late throughout the song
statementreply

lolcubes wrote:

Small offset variations under 10ms are rankable.
As I said before, maps with bad offsets don't need to be reported here. Only if they have bad multi BPMs or just wrong BPMs. We can fix offsets already and the local offset already fixes the issue if the player wants to use it.
http://osu.ppy.sh/s/10435 wrong bpm and multiple offsets, LOTS of ranked plays
http://osu.ppy.sh/s/156 wrong bpm
http://osu.ppy.sh/s/3661 wrong bpm
... and more
Oinari-sama
I face a different problem at the moment.

Previously I have as much time as needed to "tune" the offset as I play (eg +10ms, feel it, then trickle according to the feedback). But now I have to guess a value right at the beginning of the song and multiple restarts are required before I can get the value right. This is quite annoying for maps with no Skip button. Sorry I'm not so pro like some others who can sense a 1ms difference, I actually need to "tune" the offset as I play to guesstimate the right value.

Can we please have the old setup back?
Xivaxi
Is there any harm in letting us change it during breaks? I miss being able to tweak the offset without having to restart every time.
peppy
I initially removed this as some people were using it in dodgy ways (while macroing to make accuracy look more random), but I will allow changes during breaks in the next build.
Oinari-sama

peppy wrote:

I initially removed this as some people were using it in dodgy ways (while macroing to make accuracy look more random), but I will allow changes during breaks in the next build.
Yes allow changes during breaks will be helpful. Is there any chance to extend this "grace period" to say the first 20-30 seconds? Assuming the non-skipable intro isn't too long, 30s should give ample time for those who genuinely need to "tune" the offset.
peppy
intro is not included
Winshley

MillhioreF wrote:

http://osu.ppy.sh/s/43
Highlighting this because I know how offtimed that map is.

I managed to SS that map because I used Local Offset wizard on the fly during gameplay: +30ms at start of map, +20ms while spinning the first spinner, +10ms on slider before 2nd spinner, and 0ms on slight pause before the 3rd (last) spinner.

I'm curious how this is going to happen later when the change takes place, though I would bet that score wipe is likely going to happen instead.


ykcarrot wrote:

I don't care about miss timed beatmaps. Just I can't adjust offset during testing in the middle of the map, it's really annoying.
Restore it please peppy
Perhaps enable local offset change all the time only in editor mode?
theowest
I've never changed the local offset more than once while playing a beatmap. Mostly because I don't have the correct offsets written down, just don't bother to or I don't play beatmaps with such problems. Even if I do, I don't notice them, I just retry until I get decent accuracy.

Maybe we could have offset changing exceptions for these maps?
AndyPertamax
I usually set offset while Autoplay, especially in osumania mode
Please sign in to reply.

New reply