1. osu! forums
  2. Beatmaps
  3. Ranked/Approved Beatmaps

show more
posted
im a cat meow
posted

ScubDomino wrote:

im a cat meow
HI im a cat meow, I'm dad
posted
what

it doesnt matter if a buzz is buffered or not, the chance of getting a sliderbreak from a reverse arrow cutting out the hitwindow is the same

buffering only affects when you can leave a buzzslider, if anything it makes it harder for people to click early. the only way to keep that hitwindow-clipping from happening is by not using buzzsliders at all

people dont buffer buzzsliders to make clicking the next object easier, they buffer them so they dont have to hold their key down the entire time until the next object
posted

UndeadCapulet wrote:

what

it doesnt matter if a buzz is buffered or not, the chance of getting a sliderbreak from a reverse arrow cutting out the hitwindow is the same

buffering only affects when you can leave a buzzslider, if anything it makes it harder for people to click early. the only way to keep that hitwindow-clipping from happening is by not using buzzsliders at all

people dont buffer buzzsliders to make clicking the next object easier, they buffer them so they dont have to hold their key down the entire time until the next object
Sent from my iPhone
posted
broken game please fix
posted
ive got some issues in this normal
nothing objectively unrankable but i really find a lot of this diff questionable at best..

00:21:216 (1) - the gap here seems too short considering this map usually uses 4/1 rhythms... start spinner a measure earlier and have it go for 2 measures to fix
the use of .9 spacing is jarring and gives a very weird feeling, plus i don't really see the point of it
00:45:622 (1,2) - these two sliders seem to have no correlation with how or why anchor points are used, leaving this feeling of random movement which doesn't really play well in this diff. also, the 1/1 gap is incredibly weird considering the majority of the map uses 4/1 sliders.. i'd increase it to a 2/1 gap.
00:52:402 (2,3) - not gonna lie this overlap is ugly :/
00:53:758 (3,1) - another 1/1 gap but this one fits a bit better so its ok i guess
00:56:470 (1,1,1) - i don't really understand your logic with these slider angles (this applies to the whole section). maybe i'm missing something but they seem kinda randomized and nonsensical
01:18:165 (1) - i have no idea why this section is so lightly mapped.. its still kiai but it gets much easier???
01:49:351 (1,2) - i'd suggest just removing this.. this might throw players off too easily and doesn't feel right with the slider ending on the downbeat.
01:58:843 (1) - push this back 2 beats to allow for a wider gap between the spinner and slider and also to represent the beginning of the static (or at least what i hear as the beginning?)
02:23:249 (1,2) - same as 00:45:622 (1,2) -
02:31:385 (3) - this slider is confusing because the most "major" noise in it is the downbeat but the anchor is at the beat before which i feel really doesnt represent that large noise on the downbeat at all
02:55:792 (1) - same as before, i'd really up this rhythm for at least some form of consistency..
posted
man that sucks
posted
Sounds Good, Doesn't Work
posted
Regarding the buffered slider things, it would be more reasonable if all sliders consistently had their tails closest to the next object. It would make it more playable and still doesn't use buffering.

Here are some examples of non-buffered slider distance snaps:
00:57:317 (2,3) - 01:08:165 (2) - 2x
01:00:029 (2) - 01:02:741 (2) - 1x
01:00:538 (1) - 01:03:631 (2) - 0x

Changing sliders with 1/8 gaps that have 2x distance (closer to the head) to 1x distance (closer to tail) would be more fair to the player since they'd be able to keep their cursor on the previous slider circle for slightly longer.
posted
its not a matter of human skill its a matter of hardware being shite

forgot to say but nathan's problem is completely different from buffered sliders, but the main reason such a thing is a problem is the same reason as unbuffered sliders are a problem, and they are closely related
hardware input, output, and transfer has LATENCY, buffering sliders and not using sliders with 0.000000000000000000000000000000000000000000000000000001 ms repeats is supposed to combat this latency error

the problem isnt that humans are not good enough to play them
the problem is that hardware is not good enough to handle it

and BECAUSE the hardware is not good enough to handle such technicalities, the game ends up not becoming a rhythm game because u arent even playing to the rhythm at that point anymore

theres a reaason UR sub 50 is inhuman/luck and its mostly because of hardware

in this regard, the chance of sdliderbreak due to a reverse arrow is NOT the same because hardware issues cause players to stay on the note longer than intended and hit the next note late, OR hit the 1/16 not early BECAUSE the latency between hardware does not allow players to hit notes timely

i mean we cOULD just disregard all bugs and just start mapping auto-fail maps while we'r at it Tho,m or maybe ur from Wakanda and got that usb 6.3 0 ms LATENCY
posted
posted

Mishima Yurara wrote:

its not a matter of human skill its a matter of hardware being shite

forgot to say but nathan's problem is completely different from buffered sliders, but the main reason such a thing is a problem is the same reason as unbuffered sliders are a problem, and they are closely related
hardware input, output, and transfer has LATENCY, buffering sliders and not using sliders with 0.000000000000000000000000000000000000000000000000000001 ms repeats is supposed to combat this latency error

the problem isnt that humans are not good enough to play them
the problem is that hardware is not good enough to handle it

and BECAUSE the hardware is not good enough to handle such technicalities, the game ends up not becoming a rhythm game because u arent even playing to the rhythm at that point anymore

theres a reaason UR sub 50 is inhuman/luck and its mostly because of hardware

in this regard, the chance of sdliderbreak due to a reverse arrow is NOT the same because hardware issues cause players to stay on the note longer than intended and hit the next note late, OR hit the 1/16 not early BECAUSE the latency between hardware does not allow players to hit notes timely

i mean we cOULD just disregard all bugs and just start mapping auto-fail maps while we'r at it Tho,m or maybe ur from Wakanda and got that usb 6.3 0 ms LATENCY
Any evidence to back this up?
posted
common sense suggests that hardware has latency, combine the latencies with your mouse/tablet with ur keyboard, audio driver, audio peripheral, video driver, and monitor, and u got a pretty significant size

idr most of the threads since they were since like 2014 and osu!'s search is acutally garboo, but here are specifics

t/208094
https://www.reddit.com/r/osugame/commen ... d_and_how/
t/153515
https://superuser.com/questions/317217/ ... b2-0-drive
https://superuser.com/questions/16893/d ... ond-faster
https://docs.microsoft.com/en-us/window ... ency-audio
https://www.rtings.com/monitor/tests/in ... rison_1436

how would 0 latency hardware even work,,,,, what the hell bros?
posted

Mishima Yurara wrote:

common sense suggests that hardware has latency, combine the latencies with your mouse/tablet with ur keyboard, audio driver, audio peripheral, video driver, and monitor, and u got a pretty significant size

idr most of the threads since they were since like 2014 and osu!'s search is acutally garboo, but here are specifics

t/208094
https://www.reddit.com/r/osugame/commen ... d_and_how/
t/153515
https://superuser.com/questions/317217/ ... b2-0-drive
https://superuser.com/questions/16893/d ... ond-faster
https://docs.microsoft.com/en-us/window ... ency-audio
https://www.rtings.com/monitor/tests/in ... rison_1436

how would 0 latency hardware even work,,,,, what the hell bros?
Yes, but can you point to specifics in the map and stuff? And can you say that whatever latenc etc.... the map demands is too hard / unreasonable for people to hit? I'm not saying ur stuff is wrong at all, but ur just listing a bunch of resources for why latency exist but not for why it negatively affects this map in particular.
posted
when osu meet maths
posted
oh in the map lol i thought u meant proof that latency exists in hardware and i was bewildered

all the unbuffered 1/8 sliders
all the 1/16 sliders

1/16 hitwindow is 21 ms
1/8 hitwindow is 42 ms

the 2nd question makes it sound like u dont know what latency Actually is tho; Latency is not a skill,, it doesnt make sense to say that a map demands hard latency, latency is a time interval of delay that is prevalent in all hardware

the latency required for these sliders is very minimal and for most players it will quite literally be RNG because of such hardware limits, tho there are some players who use high quality graphics cards, keyboards, video cards, e.t.c but Yall ForReal? Spending 2k CAD on osu btw

also i think it speaks for itself why latency would negatively affect a rhythm game............... delayed responses due to hardware in a game that requires timing is Lol, this negatively affects all maps due to timing window constraints it causes

i dont acutally care about the map btw i just wanted to make sure that people realize that this is an actual thing and this is the original/main reason mappers buffered sliders my first post before this was even ranked wasnt Srs Lol but uc never responded to my How Tho so gl osu dead game anyway peaked in 2001

sent from my telegram
posted
has VINXIS been secretly smart all this time!?!?!?!?
posted

Mishima Yurara wrote:

oh in the map lol i thought u meant proof that latency exists in hardware and i was bewildered

all the unbuffered 1/8 sliders
all the 1/16 sliders

1/16 hitwindow is 21 ms
1/8 hitwindow is 42 ms

the 2nd question makes it sound like u dont know what latency Actually is tho; Latency is not a skill,, it doesnt make sense to say that a map demands hard latency, latency is a time interval of delay that is prevalent in all hardware

the latency required for these sliders is very minimal and for most players it will quite literally be RNG because of such hardware limits, tho there are some players who use high quality graphics cards, keyboards, video cards, e.t.c but Yall ForReal? Spending 2k CAD on osu btw

also i think it speaks for itself why latency would negatively affect a rhythm game............... delayed responses due to hardware in a game that requires timing is Lol, this negatively affects all maps due to timing window constraints it causes

i dont acutally care about the map btw i just wanted to make sure that people realize that this is an actual thing and this is the original/main reason mappers buffered sliders my first post before this was even ranked wasnt Srs Lol but uc never responded to my How Tho so gl osu dead game anyway peaked in 2001

sent from my telegram
??? Universal offset exists. If you're consistently getting 15ms delay due to latency you account for that through adjusting offset. So I guess what you really mean is that 21ms is somehow unfair as a timing window? But at OD 10 you only have 1 +/- 19.5ms hitwindow anyways?

So what is your solution? Removing even more repeatsor something? I'm actually still confused since this isn't the same issue Nathan brought up (which isn't going to be fixed by removing repeats anyways).
posted
what
posted
show more
Please sign in to reply.