forum

[Mapping] Technical problems with new hypers

posted
Total Posts
25
Topic Starter
Drafura
So I tried to adapt some maps from the old build to the new one and I've found some problems wich force me to breaks patterns shape into ugly and/or non intuituve ones.

After a discution with Millhi, I have finally uploaded a test map to clearly illustrate each problems brought by this system : https://osu.ppy.sh/s/95487

The map is actually divided in 3 sections separated by breaks (I can extend it if someone wants me to show another issue I've missed), each combo starts with two dead beats to make it more confortable to play. I'll not use time to explain so if I say "2.3" I'm talking about the second combo in the third section, I think this is more user friendly (If it is asked I'll edit to use time).


First section :
I use the maximum spacing for a non dash jump in each pattern (1.98x, 2x is hyper'd). This section is designed to show how hyperdashes impact the difficulty of the same jump. In a first look it could be a good idea but integrate into another hard patterns the result is that the player is permanently changing the way he have to play wich is wrong cause this is the same exact jump everytime and should be played in the exact same way.
1.1 Just the 1.98x jump.
1.2 A 2x hyperdash followed by a 1.98x jump. This one feels similar the other pattern, just a little bit harder nothing really annoying.
1.3 Again the 1.98x jump, placed here to feel the differences.
1.4 A 2.84x hyperdash followed by a 1.98x jump. This one is way more tricky to catch, and it appears that with even higher spacings (by using 1/2 1/4 hypers with 3.0, 4.0, 5.0x ... spacing) the jump can become near impossible to catch.

This section demonstrate pretty well how the new hypers affect the hit window of a post hyper jump. Basically it feels like in standard mode the OD was changing constantly and sometimes you have to be hyper accurate and sometimes you're not even sure there's a rythm involved in the game. So basically the player tends to try to maintain a mega timing all map long wich sometimes makes him start too soon or too late and the fun is gone fast.
About the mapping two solution are possible :
- Reduce the hyper, seems the best solution, but the hyper will feel slow and will maybe not fit with the song.
- Reduce the jump, pretty bad solution imo cause if you are actually using this jump in another pattern you certainly wants to conserve consistency and use the same spacing.

Best solution :
- This needs a fix, 2 jumps with the same spacing should be played the same way, hyperdash or not.


Second section :
Shows a behavior wich already existed in the old build. But this behavior makes me actually thinks about : Are all jumps over the hyper trigger really needed ?
2.1 This a 2.04x jump (1/1). osu! detect it as hyperdash.
2.2 This another 2.04x jump (1/2). osu! does not detect it as hyperdash.
2.3 This a 2.31x jump (1/2). This is the minimum spacing for osu! to detect it as hyperdash.
2.4 This another 2.04x jump (1/4). osu! does not detect it as hyperdash. (As you can see it is almost half circle difference, is this really due to frames ?
2.5 and 2.6 Same behavior with more difference for 1/4.

This behavior just makes me wonder if the 1.1 hyper (2.04x at 1/1) is really needed or not.


Third section :
This section is about the "slipping" effect those hypers generate.

3.1 2 consecutive 1/4 hyperdashes with the minimum trigger spacing.
3.2 2 consecutive 1/4 hyperdashes with a huge spacing. The feeling is totally different here and without a good timing the ryuuta will "slip" away and you'll probably miss the next 1/4 hyper.
3.3 The only solution I've found so far to fix this, you can appreciate how ugly the pattern is only to get the same feeling as in 1.1.


All of these things together are removing possibilities to mappers in term of patterns shapes and diffculty wanted. What was wrong with old hypers ? I never heard anyone complaining about them, now you can see in the "what's your biggest weakness" thread a lot of people having hyperdashes as their weakness. Does old hyper required more rythm ? I don't think so, it should be the mapper to decide when he want to add rythm or not, hypers were already something wich was hard enough cause they can makes you read from part to part of the screen.

Some asked tldr so :
Please, peppy... Give us good hyper behavior and I can assure that we mappers will add the rythm CtB needs.

Do not post screenshoot of SSs. I have SS'd it so I know it is possible, this thread is here to discuss about those issues and maybe make the things change before a lot of CtB maps gets ranked and make those needed changes impossible.

Thanks for reading, discuss...

Edit :
- 28/05/13 : Fixed section number.
eldnl
I really can't understand what is that wrong ... I play your demo just fine
Topic Starter
Drafura

eldnl wrote:

I really can't understand what is that wrong ... I play your demo just fine

Drafura wrote:

So basically the player tends to try to maintain a mega timing all map long wich sometimes makes him start too soon or too late and the fun is gone fast.
You're one of the few players in the game wich can maintain a very good timing.

As I said, this map is just made to show the different cases, if you use al lot this kind of patterns in a long map you'll maybe see how this can impact your ability to play it fine. Even if it not impacts you due to your skill not everyone can reach these abilities in a long map, cause these are technical behavior wich are intented from what I understood. They are intended to give to hyperdashes more timing but I strongly thinks this is wrong : It exists others ways to make rythm via the mapping itself and some hypers seems to be unneeded. Why forcing the mapper to use an hyperdash when he want to add rythm ? If a mapper wants for his jump the same difficulty as the 1.4 jump in my map he's forced to use 3 object and an hyperdash. Why complicate so much the mapping when it exist another way wich is more free, easier and obviously not bugged (I never heard about an uncatachable old hyperdash).

I'll use VelperK's maps as example here : He use a lot of hyperdashes in his maps cause he want to make them challenging in term of timing and reading, maybe if he had another option to give timing he'll use more non hyper jumps or even if he won't do it during mapping, maybe during the modding process when you say "this hyper is uneeded" and he replies "I want rythm here" you could have an option to propose using a timed non hyper jump.

I understand peppy already spent a lot of his time to fix the old bugged jumps, but I think this system still needs refinement including the old hyperdash come back and more tweaks on 2 fruits non hyper jumps (Pretty much all old jumps where catchable, and some frame differences should not impact normal jumps that much imo). I tried to help when this bug was being fixed, one of the peppy tests was prefect resolving all bug jumps with the old hyper system it only had a problem for droplets in sliders, I think this is the best direction to get to fix those issues.
eldnl
Oh I see, so you can't put hard left-to-right jumps?
I mean, like this map.
If that's the point, then this should be fixed.
ZHSteven
Professional prove. you get my support.

1st section is the prove of the inconstant speed of "kid"(sorry I forget his name) in an extreme situation.
So far as I know, this situation only appears at Hyper jump + normal jump(old system only appears at extreme normal jump) or Hyper jump + a left-to-right jump(o-o-o)
Actually, when the new system come, a lot of people complain about the new system "already lose the feeling of the jumps","shitty hypers","completely playing a new game", I guess it has been solved.

***And also, some people complained about the inconstant speed of the same reverse hyper jump, could this get further proved or rejected?


2nd section is the prove of how a bug jump is been generate.
I thought this have been solved, but seems I was wrong....
This have been discuss a long time. I remembered I have proved it long time ago, but whatever..
if you want to see more prove, click(huge image size!! warning!!!
【2.2】I add a note makes a left-to-right jump, also keep the distance to x2.04.


firstly, only 3 note with 1/2 time.



then, 3 note with 1/4 time.



then, 5 note with 1/4 time.


last, 10 note with 1/8 time.(yes i am trolling)

(I cannot set this in a proper distance. maybe is my mouse problem?)


See the difference? Yes, this is definitely not the problem about "frame" at all. This is just a bug. Peppy fixed the bug jump in some way we dont know, but the bug still there.
Extra?




with extreme SV(3.6)

I think we don't need 3rd section, because this is personal biased. Some people really like this kind of pattern as I know, and they think this not affect their performance at all.
(whispered: I really hate them)
Topic Starter
Drafura

ZHSteven wrote:

I remembered I have proved it long time ago, but whatever..
Sorry if I pointed out something well known, I'm not that old into CtB mapping and didn't followed discutions about technical problems in ctb since the hyper update (I just encountered all these issues when I tried to map with this build). But I know this issue was happening in older builds, and gives serious questions about the fix of the hyper trigger distance. I understand frames can bring some irregularities but not that much (only from 1/1 to 1/2 it's almost half a circle ! A frame issue should only affect a bunch of pixels, am I wrong ?).

ZHSteven wrote:

I think we don't need 3rd section, because this is personal biased. Some people really like this kind of pattern as I know, and they think this not affect their performance at all.
(whispered: I really hate them)
No personnal opinion here. This section show how the ryuuta slides away when you press a bit too long the key. I used 1/4 hyper cause the effect is greatly increased with fast hyperdashes. On 1/8, 1/16 it's even harder to get the fruit and stand in the same position. I used 3.3 to show a way I've found to give more timing to the player but it breaks the pattern and makes it ugly.

For the personnal preference, the question isn't "Like it or not ?" but "1/4 hypers should be allowed in CtB mapping ?" I think yes, why not ? Well used they add new patterns and creativity. But I agree that well spammed they're a pain in the *ss. If you want to allow them, they need a fix cause they're pretty ugly to map if you wants to keep a decent difficulty.

eldnl wrote:

Oh I see, so you can't put hard left-to-right jumps?
I mean, like this map.
If that's the point, then this should be fixed.
Yes, the second section is a clue of why I think they doesn't need hyperdash cause the frames shouldn't affect that much the ryuuta. First section is a way do to them but you'll need 3 hits starting with an hyperdash. Well, forcing the mapper to use an hyperdash to let him put a hard jump doesn't really make sense.
ZHSteven

Drafura wrote:

Sorry if I pointed out something well known, I'm not that old into CtB mapping and didn't followed discutions about technical problems in ctb since the hyper update (I just encountered all these issues when I tried to map with this build). But I know this issue was happening in older builds, and gives serious questions about the fix of the hyper trigger distance. I understand frames can bring some irregularities but not that much (only from 1/1 to 1/2 it's almost half a circle ! A frame issue should only affect a bunch of pixels, am I wrong ?)..
That is I talked into one of the Chinese Ctb thread if I am not wrong, but that seems old and out of date since the new hyper systems here. I just want to share something I already known. Apologize about it.

Drafura wrote:

No personnal opinion here. This section show how the ryuuta slides away when you press a bit too long the key. I used 1/4 hyper cause the effect is greatly increased with fast hyperdashes. On 1/8, 1/16 it's even harder to get the fruit and stand in the same position. I used 3.3 to show a way I've found to give more timing to the player but it breaks the pattern and makes it ugly.
I think this is speed based. hyperdash ends only when the rhythm is there(kid touch the hole note, I guess), not when you touch the note. If the speed of the jump is too fast, high speed hyperdash will "slip" you away just like your map stats, that is because some little "dash" haven't being used up. When you end the jump at a later timing, I mean just a little late, the extremely high speed will definitely leads you to the wrong place.
I personally think that is not a big thing that needs to be fixed. There is some people willing to play this kind of trolling pattern, and they can catch it well. Then why we need to forbid it? Or take a step back, if this must be fixed, then how should this be fixed? increase the allowance for it? would it affect other jumps?

Drafura wrote:

For the personnal preference, the question isn't "Like it or not ?" but "1/4 hypers should be allowed in CtB mapping ?"
This is not the only hyper-trolling pattern. At the beginning, I was highly against them, think this was broke the flow of the whole map, and should be nuked. But now, since almost everybody like them, I have no ideas already.. After all this should be played by players, not mappers. So I think if this not flood, this will be ok to me.

Drafura wrote:

If you want to allow them, they need a fix cause they're pretty ugly to map if you wants to keep a decent difficulty.
I don't see it is too "ugly" in fact.. It is good isnt it? XD
ok back to topic, it is impossible to make this pattern smooth. if you really needs to add this pattern, seriously I think your changes is good, or maybe reduce the hyperdash speed will work.
eldnl
I'm not sure but this is probably an example for a 1/4 jumpy map ... for me, plays very well.
VelperK
Keep dreaming if you really think peppy is going to care about CTB again. Needless to say about this very thread.

eldnl: Verdisphena's diffs are shit now with this behaviour, so are TAG4. But since this kind of mapping is minority, none cares.
He did a good job in reducing mapping possibilities for CtB I have to admit.

peppy: if you just passed by to the last post of this thread, just fix this kind of patterns and we're ok (I know you won't but just saying <3)



Also, eldnl, your map doesn't have the kind of patterns I just showed above, so it doesn't count. As CtB mappers, we now have to prevent that kind of patterns (and oh boy are we reducing ourselves with that limitation) since they're 100% luck depending, just like the old pixels he wanted to fix.
eldnl

VelperK wrote:

Keep dreaming if you really think peppy is going to care about CTB again. Needless to say about this very thread.

eldnl: Verdisphena's diffs are shit now with this behaviour, so are TAG4. But since this kind of mapping is minority, none cares.
He did a good job in reducing mapping possibilities for CtB I have to admit.

peppy: if you just passed by to the last post of this thread, just fix this kind of patterns and we're ok (I know you won't but just saying <3)



Also, eldnl, your map doesn't have the kind of patterns I just showed above, so it doesn't count. As CtB mappers, we now have to prevent that kind of patterns (and oh boy are we reducing ourselves with that limitation) since they're 100% luck depending, just like the old pixels he wanted to fix.
I put my map for 1/4 jumps so... anyways, you're right, Rainbow Dash map is the perfect example, is not lucky based, nothing is lucky based, but you need to catch the perfect timing to do them.
VelperK

eldnl wrote:

VelperK wrote:

Keep dreaming if you really think peppy is going to care about CTB again. Needless to say about this very thread.

eldnl: Verdisphena's diffs are shit now with this behaviour, so are TAG4. But since this kind of mapping is minority, none cares.
He did a good job in reducing mapping possibilities for CtB I have to admit.

peppy: if you just passed by to the last post of this thread, just fix this kind of patterns and we're ok (I know you won't but just saying <3)



Also, eldnl, your map doesn't have the kind of patterns I just showed above, so it doesn't count. As CtB mappers, we now have to prevent that kind of patterns (and oh boy are we reducing ourselves with that limitation) since they're 100% luck depending, just like the old pixels he wanted to fix.
I put my map for 1/4 jumps so... anyways, you're right, Rainbow Dash map is the perfect example, is not lucky based, nothing is lucky based, but you need to catch the perfect timing to do them.
So basically old pixels were not luck based since you needed to have a perfect timing in order to catch them.

OK.
eldnl

VelperK wrote:

So basically old pixels were not luck based since you needed to have a perfect timing in order to catch them.

OK.
You're misunderstanding, pixel jumps were technically impossible to catch, but this kind of jump are technically possible, but hard as hell.
VelperK

eldnl wrote:

VelperK wrote:

So basically old pixels were not luck based since you needed to have a perfect timing in order to catch them.

OK.
You're misunderstanding, pixel jumps were technically impossible to catch, but this kind of jump are technically possible, but hard as hell.
They were possible to catch, but only if you were lucky enough. Withouth any tricks whatsoever I was able to catch the 40 combo of Cirno's Perfect Math Class TAG4 MANY times, and I felt that the timing in those ocassions was splendid, not too soon neither too late, but since it was too hard for someone to catch it in a skill based way, a hyperdash was added (and I totally support that). There are hundreds of other examples in which hyperdashes were added in unnecesary places (and i'm not saying I dislike the extra hyperdashes, just stating that they were unnecesary because you didn't need them in order to catch the pixels) because it was "too lucky based". You can't deny this as it's a fact, luck DOES exist here.

Also the "pixel" term refers to a jump on which by a margin of some pixels IS actually possible, but just too hard for a human to do correctly in a skill based manner.
eldnl
:facepalm:

you're misunderstanding again, that's why I said technically ... well I guess we need to stop now.
VelperK
We can agree on that both of these situations were/are possible but too luck-based and the later I just stated needs to be fixed ASAP, ok?
ursa
oh you two :o

VelperK wrote:

We can agree on that both of these situations were/are possible but too luck-based and the later I just stated needs to be fixed ASAP, ok?
Agree :D

anyway :

https://osu.ppy.sh/s/51611 (Diff end's wing) the beggining pattern

Seems this kind of pattern makes me annoyed , cus i've maked simmiliar pattern like that in the past & now it's completely make my old map almost UNcatchable cuz of the hyper changes >:(
Topic Starter
Drafura

urs4chan wrote:

https://osu.ppy.sh/s/51611 (Diff end's wing) the beggining pattern

Seems this kind of pattern makes me annoyed , cus i've maked simmiliar pattern like that in the past & now it's completely make my old map almost catchable cuz of the hyper changes >:(
This is related to section 3 of the demo map (In boku boku case those are 1/2 due to bpm but it equals 1/4 of most maps with decent bpms) and yes this is a good example of something wich will be used a lot in CtB mapping and is completly ruined by the new hypers. I think these kind of patterns are not meant to be hard in CtB mapping imo they're meant to play with your reading abilities, that's why old hypers fits so well in those situations. They allow you to use patterns wich calls the reading skill of the player without increasing the timing difficulty, it gives coherence and allows you to balance your map as you wish. I want to add that one of my map has been affected by this (Oriens) and I think it is actually the reason why there's not so many FCs/SSs on it (Those 1/4 sliders are too random if you want to keep a good accuracy, so most people get pissed of it and ignore it. Randomly missing droplets or combo isn't fun for anyone and I really didn't designed to be this way when I mapped it).

ZHSteven wrote:

I don't see it is too "ugly" in fact.. It is good isnt it? XD
ok back to topic, it is impossible to make this pattern smooth. if you really needs to add this pattern, seriously I think your changes is good, or maybe reduce the hyperdash speed will work.
My recent mapping experiences with this build shows me that if you want to keep them smooth to play you have to affect the pattern even at a small hyperdash distance, cause when they're integrated in a whole map it's again very different than just showing the case in a demo map.
Here's one of these said experience : https://osu.ppy.sh/b/247418 @ 01:14:602 (7,1,2,3) -
The original pattern was like this I had to modify it to this for the only reason the hyperdashes was dizzy, so I lost in map design and readability to gain timing on hyperdashes, and I have no other solutions at the moment.

eldnl wrote:

I'm not sure but this is probably an example for a 1/4 jumpy map ... for me, plays very well.
Yes, for a 140 bpm song they're not so dizzy, do it with +200 bpm songs this will leads the player to countless stupid jumps wich should only call the reading skill of the player not a perfect timing skill.

About post hyperdash jumps you're both right. Velperk is right cause yes, luck is involved since it require perfect timing (when you miss it feels impossible and when you catch it feels too easy, I don't know if i'm the only one feeling this when I play them) and eldnl is right too cause a frame issue could or not be involved in an already timed jump. Difference is that you have more chances to catch the new ones than the old ones, cause there isn't any frame issues in the new build (In the old build even if irl you had the perfect timing the frame issue in the game could makes you miss, so it's an addition).

VelperK wrote:

As CtB mappers, we now have to prevent that kind of patterns (and oh boy are we reducing ourselves with that limitation) since they're 100% luck depending, just like the old pixels he wanted to fix.
I think the same about those jumps they should be banned from CtB mapping they pretty much makes no sense in term of mapping. Basically you make some hypers followed by a jump and the result for this pattern is that the hardest jump is the non hyper'd one ? And it's actually the way to make the hardest jumps of the game ? This makes no sense, the hardest jump of the game should be doable out of all hyperdash'd pattern in order to permit mappers to balance their maps in term of difficulty. So yes, this will certainly go for an unrankable pattern if the jump is too big and again remove something we could use to make maps.

VelperK wrote:

Keep dreaming if you really think peppy is going to care about CTB again. Needless to say about this very thread.
I sadly know how this is going to lead to nothing, but this discution have to exist at least to inform mappers about those issues, discuss about them and try to find ways to fix by modifying patterns. Plus in a long term if many mappers are having problems with those issues maybe something will change one day.
You know, in a first time I just ignored the build and kept mapping under the old build and then I got censored when I linked the old build in the forum, so for me the only way to share my maps the way they are supposed to be played is just "not allowed" with no other words wich could stand for a reason. So yes I'm fiighting back for this, cause I love mapping and CtB.
peppy
No pattern should be "banned". I don't mind making post-hyperdash more lenient, but this may cause complains about other maps becoming too easy as a result (which was the case during testing).

Also avoid the shit statements about me not caring about stuff. Oh wait, this is VelperK. Nevermind.
Last Remnant_old
Well, I think that you could make the catcher reach the destination a bit faster, because now I think we only have like 5ms or even less for edge-to-edge hyperjumps (the time between arriving at destination fruit before it actually hits the plate). Actually, maybe it is time to use the OD difficulty setting of map to determine this leeway time, what do you think? So lower OD would mean more time before you arrive under destination fruit (therefore, faster hyperdash speed).

Another important thing that should be discussed is hyperdash after hyperdash in changed direction.
Example:
.(3)
         .(2)
.(1)

Imagine catching fruit 1 and holding shift+right heading to fruit 2. As I have said, you have around 5ms time to change to left direction after reaching under fruit 2. Most often, you will still be holding right arrow key after catching fruit 2, and it would cause you to hyperdash in wrong direction. Since hyperdash speed is too high, this will mean that you will miss fruit 3.
I would propose that the game shouldn't allow hyperdashing in wrong direction OR as I have previously said, increase the hyperdash speed to allow more reaction time.
eldnl

peppy wrote:

Also avoid the shit statements about me not caring about stuff. Oh wait, this is VelperK. Nevermind.
xDD

so there is nothing that you can do?
peppy
I thought I already stopped hyperdashing in the wrong direction. If not, this is a definite change that can be made.
Topic Starter
Drafura

peppy wrote:

No pattern should be "banned". I don't mind making post-hyperdash more lenient, but this may cause complains about other maps becoming too easy as a result (which was the case during testing).
And what about the second section ? Don't we have a chance here to get more difficult non hypered jumps ? This could counter balance the difficulty (Making hyperdashes easier only focusing in reading skill of the player and normal jumps harder focusing on the timing skill of the player, so mappers can choose on what they want to play on their maps like it was the case.)

About the hyperdash going to the wrong direction (at least it seems like it still in hyperdash state since the timing is too small), it's basically the "slipping effect" if you press a bit too long the ryuuta slips away wich actually cause post hyper jumps being extremly hard (compared to the other jumps) depending on timeline/bpm used for the pattern.
VelperK

peppy wrote:

No pattern should be "banned". I don't mind making post-hyperdash more lenient, but this may cause complains about other maps becoming too easy as a result (which was the case during testing).

Also avoid the shit statements about me not caring about stuff. Oh wait, this is VelperK. Nevermind.
I would happily test the results of such experimentation in the osu!test build, if you decide to let users see if the changes fit them or not.

I'm sure I won't be the only one tho.
Kurokami

VelperK wrote:

peppy wrote:

No pattern should be "banned". I don't mind making post-hyperdash more lenient, but this may cause complains about other maps becoming too easy as a result (which was the case during testing).

Also avoid the shit statements about me not caring about stuff. Oh wait, this is VelperK. Nevermind.
I would happily test the results of such experimentation in the osu!test build, if you decide to let users see if the changes fit them or not.

I'm sure I won't be the only one tho.
I will keep my supporter 'till then. :P
ConanCloud

WasteK wrote:

peppy wrote:

No pattern should be "banned". I don't mind making post-hyperdash more lenient, but this may cause complains about other maps becoming too easy as a result (which was the case during testing).

Also avoid the shit statements about me not caring about stuff. Oh wait, this is VelperK. Nevermind.
I would happily test the results of such experimentation in the osu!test build, if you decide to let users see if the changes fit them or not.

I'm sure I won't be the only one tho.
/sign

I want to test it, too. We need a fix :)
Please sign in to reply.

New reply