DUST Search - [Feedback] Stick Deflection vs. Rotation Speed (Graph included)

Sete Clifton wrote:

I did a quick analysis of how stick deflection is related to rotation speed. Here are the results I found:

https://docs.google.com/spreadsheet/ccc?key=0AqMshc-A6HVHdFlKbklqY1hQWnE4WVJIc1pCVHE4Smc#gid=0

Quick note: The blue line ("current") is how the system works now, and the red line ("proposed") is what I believe it should be.

Anyway, as we can see (after accounting for some small error), rotation speed appears to increase linearly with the distance from center that the stick is moved. This is a large part of the reason why I think aiming with the DS3 still feels imprecise and "slippery". The current system makes it way too easy to accidentally build up unwanted turning speed.

Most modern FPS it seems use a system where the effectiveness of stick deflection starts off low and increases as you get closer to the edge. This allows you to have a wide range at which you can make smaller and smoother movements, but also still gives you the ability to turn quickly by pushing the stick further.

The second line on that graph I provided does this, and is what I propose it should be (in terms of the general design). Reaching 50% rotation speed should happen around 80-85% stick deflection. Basically, the idea is that the deflection scale needs to have more detail/range at the low end, and less at the high end. Which makes sense (at least to me), because when you want to turn really fast, you're already giving up precision, but when you want to make slow fine movements, you want as much detail as possible.

Any thoughts on this?

Pages: 1 2 :: [one page]

Author	Thread Statistics \| Show CCP posts - 1 post(s)
Sete Clifton PSU GHOST SYNDICATE DARKSTAR ARMY 187	Posted - 2013.05.27 02:28:00 - [1] - Quote I did a quick analysis of how stick deflection is related to rotation speed. Here are the results I found: https://docs.google.com/spreadsheet/ccc?key=0AqMshc-A6HVHdFlKbklqY1hQWnE4WVJIc1pCVHE4Smc#gid=0 Quick note: The blue line ("current") is how the system works now, and the red line ("proposed") is what I believe it should be. Anyway, as we can see (after accounting for some small error), rotation speed appears to increase linearly with the distance from center that the stick is moved. This is a large part of the reason why I think aiming with the DS3 still feels imprecise and "slippery". The current system makes it way too easy to accidentally build up unwanted turning speed. Most modern FPS it seems use a system where the effectiveness of stick deflection starts off low and increases as you get closer to the edge. This allows you to have a wide range at which you can make smaller and smoother movements, but also still gives you the ability to turn quickly by pushing the stick further. The second line on that graph I provided does this, and is what I propose it should be (in terms of the general design). Reaching 50% rotation speed should happen around 80-85% stick deflection. Basically, the idea is that the deflection scale needs to have more detail/range at the low end, and less at the high end. Which makes sense (at least to me), because when you want to turn really fast, you're already giving up precision, but when you want to make slow fine movements, you want as much detail as possible. Any thoughts on this?
KAGEHOSHI Horned Wolf Seraphim Initiative. CRONOS. 4013	Posted - 2013.05.27 03:16:00 - [2] - Quote Very good work, CCP needs to see this
KA24DERT Not Guilty EoN. 54	Posted - 2013.05.27 03:26:00 - [3] - Quote My thoughts on this is that CCP needs to stop meddling with our controls. Let them make sane defaults and some presets, and give us all the sliders and tick boxes we need to tailor our controls to our playstyles.
Moejoe Omnipotent Imperfects Negative-Feedback 459	Posted - 2013.05.27 03:38:00 - [4] - Quote The sloppy controls is almost entirely due to input lag.
Zeylon Rho Subdreddit Test Alliance Please Ignore 385	Posted - 2013.05.27 04:30:00 - [5] - Quote I like the analysis. It's basically adding acceleration on the far end of the stick I guess. I find the current system mostly workable, but I can see how your way makes sense.
Sete Clifton PSU GHOST SYNDICATE DARKSTAR ARMY 187	Posted - 2013.05.27 05:27:00 - [6] - Quote Zeylon Rho wrote: I like the analysis. It's basically adding acceleration on the far end of the stick I guess. I find the current system mostly workable, but I can see how your way makes sense. Yeah, sort of. Though just to be clear I would like to mention that this has nothing to do with aim acceleration. This is purely about the maximum speed at a given distance the stick is pushed. Anyway, the idea is that right now from 0% to 100%, rotation speed is spread out evenly across the entire stick deflection range. The problem with this is that we don't need 50% of the space being taken up by speeds that are impossible to use for fine/general aiming anyway. With what I'm proposing (and what most FPS do now days) is that you essentially expand the first 50% and contract the second 50%. This way you get more room to make smaller and smoother adjustments. If you want to turn around really quickly, then you push the stick all the way to the edge. This frees up about 85% of the deflection range to be used to for regular aiming (as opposed to the roughly 50% that we get now).
Ludvig Enraga KILL-EM-QUICK RISE of LEGION 203	Posted - 2013.05.27 05:47:00 - [7] - Quote I like it a lot. Like most in depth and sound thoughts it will not be implemented I am afraid.
Protocake JR Internal Error. Negative-Feedback 61	Posted - 2013.05.27 06:49:00 - [8] - Quote I would really like a Dev's input on this subject. Anyone know the people responsible for tuning our controls?
Sete Clifton PSU GHOST SYNDICATE DARKSTAR ARMY 189	Posted - 2013.05.27 07:02:00 - [9] - Quote Protocake JR wrote: I would really like a Dev's input on this subject. Anyone know the people responsible for tuning our controls? Not positive, but Wolfman seems to be involved in aiming, as he's generally the main DEV that comments/makes threads about it. Anyway yeah, I'd also really like a DEV to at least see this. I mean, if it's not right for this game or them, then I have to respect their creative direction. But I think this would drastically help aiming on the DS3.
Laurent Cazaderon What The French CRONOS. 1420	Posted - 2013.05.27 07:51:00 - [10] - Quote Wolfmann is still working on controls for KBM and DS3. He tries to see what's up with the input lag and tweaks the acceleration curve and such using new test builds pretty much every day. No idea when he will feel like it's done but stay put and wait for some news. By the way, it seems they loved your thread dude ^^

XxGhazbaranxX The Unholy Legion of Darkstar DARKSTAR ARMY 46	Posted - 2013.05.27 07:52:00 - [11] - Quote Hello, I really respect this idea, it's well thought. This would help a lot on the kbm vs ds3 battle. CCP should not only take weapon and suit balancing as a priority but also go into the basics. I like any idea that tries to give them ds3 the same precision as a mouse. I personally use keyboard and mouse because how sluggish and inaccurate the ds3 feels.
XxGhazbaranxX The Unholy Legion of Darkstar DARKSTAR ARMY 46	Posted - 2013.05.27 07:54:00 - [12] - Quote Laurent Cazaderon wrote: Wolfmann is still working on controls for KBM and DS3. He tries to see what's up with the input lag and tweaks the acceleration curve and such using new test builds pretty much every day. No idea when he will feel like it's done but stay put and wait for some news. By the way, it seems they loved your thread dude ^^ Good to see that good ideas are being looked at.
Arramakaian Eka Tronhadar Free Guard Minmatar Republic 607	Posted - 2013.05.27 09:17:00 - [13] - Quote Great analysis, agree on your suggestion.
EXASTRA INVICTAS Crux Special Tasks Group Gallente Federation 36	Posted - 2013.05.27 11:51:00 - [14] - Quote I agree with OP, and it's nice to at least hear from CPM on the issue.

CCP Wolfman C C P C C P Alliance 403	Posted - 2013.05.28 09:13:00 - [15] - Quote Hi guys, Awesome graph Sete! You are correct the input/speed is linear atm. We are currently testing different curves. At the end of last week we were using a curve very similar to the one you outlined (we have in the past as well). Unfortunately weGÇÖve not been very happy with the results. We saw a limited increase in low input sensitivity, as you would expect, but at higher input levels it became harder to judge how much input would deliver the speed you wanted when tracking a target. The benefit we have found with the linear GÇ£curveGÇ¥ is that its predictability when ramping up/down makes tracking speed changes easier to judge. Anyway, weGÇÖll let you know how it goes as we continue the work. I agree with Sete that part of the GÇ£slipperyGÇ¥ issue may be building up unwanted turning speed. However right now I donGÇÖt think this is caused by the input curve. I believe it may partially be a result of a faster time to max rotation speed when the edge acceleration kicks in (which happens just before max input). This is something we are testing at the moment. We are also testing the different parts of the aim assist system to check for any hidden issues there. This does take a while which is why weGÇÖve not made an update yet. Issues with performance related input lag and hit detection are also being investigated. These can have a really big impact on how the aiming feels so weGÇÖre looking at this from all angles. CCP Wolfman

KAGEHOSHI Horned Wolf Seraphim Initiative. CRONOS. 4040	Posted - 2013.05.28 09:16:00 - [16] - Quote CCP Wolfman wrote: Hi guys, Awesome graph Sete! You are correct the input/speed is linear atm. We are currently testing different curves. At the end of last week we were using a curve very similar to the one you outlined (we have in the past as well). Unfortunately weGÇÖve not been very happy with the results. We saw a limited increase in low input sensitivity, as you would expect, but at higher input levels it became harder to judge how much input would deliver the speed you wanted when tracking a target. The benefit we have found with the linear GÇ£curveGÇ¥ is that its predictability when ramping up/down makes tracking speed changes easier to judge. Anyway, weGÇÖll let you know how it goes as we continue the work. I agree with Sete that part of the GÇ£slipperyGÇ¥ issue may be building up unwanted turning speed. However right now I donGÇÖt think this is caused by the input curve. I believe it may partially be a result of a faster time to max rotation speed when the edge acceleration kicks in (which happens just before max input). This is something we are testing at the moment. We are also testing the different parts of the aim assist system to check for any hidden issues there. This does take a while which is why weGÇÖve not made an update yet. Issues with performance related input lag and hit detection are also being investigated. These can have a really big impact on how the aiming feels so weGÇÖre looking at this from all angles. CCP Wolfman Thank you very much for working on it
Sete Clifton PSU GHOST SYNDICATE DARKSTAR ARMY 195	Posted - 2013.05.28 10:12:00 - [17] - Quote Thanks for the reply and detailed response Wolfman, it's very much appreciated. That's a shame it didnt really work out in your tests, but in any case it's just nice to know you're at least actively looking into all this kind of stuff. I'm sure you guys will figure it out though. Keep up the good work!
Laurent Cazaderon What The French CRONOS. 1429	Posted - 2013.05.28 10:15:00 - [18] - Quote told you he liked it
Aeon Amadi Mannar Focused Warfare Gallente Federation 1400	Posted - 2013.05.28 11:35:00 - [19] - Quote CCP Wolfman wrote: Hi guys, Awesome graph Sete! You are correct the input/speed is linear atm. We are currently testing different curves. At the end of last week we were using a curve very similar to the one you outlined (we have in the past as well). Unfortunately weGÇÖve not been very happy with the results. We saw a limited increase in low input sensitivity, as you would expect, but at higher input levels it became harder to judge how much input would deliver the speed you wanted when tracking a target. The benefit we have found with the linear GÇ£curveGÇ¥ is that its predictability when ramping up/down makes tracking speed changes easier to judge. Anyway, weGÇÖll let you know how it goes as we continue the work. I agree with Sete that part of the GÇ£slipperyGÇ¥ issue may be building up unwanted turning speed. However right now I donGÇÖt think this is caused by the input curve. I believe it may partially be a result of a faster time to max rotation speed when the edge acceleration kicks in (which happens just before max input). This is something we are testing at the moment. We are also testing the different parts of the aim assist system to check for any hidden issues there. This does take a while which is why weGÇÖve not made an update yet. Issues with performance related input lag and hit detection are also being investigated. These can have a really big impact on how the aiming feels so weGÇÖre looking at this from all angles. CCP Wolfman Now, are these tests you're performing conducted on a neutral platform separate of current in-game Dust mechanics or are you using the mechanics that we are using? Allow me to elaborate. Are you using the engine itself as your control? Certain suits turn slower than others and application of modules that increase mass (I.E Armor Plating) will affect the rotation speed considerably. A Heavy with a single complex armor plate turns slower than one that does not. The reason I ask is because what may be a good system for say, a Caldari/Gallente Assault suit, may over-throw the usability of the Minmatar Scout suit by side-effect. It's a bit jarring. If I play the Gallente Assault suit for an entire week than go and try the Minmatar Scout (which I have done!) I suddenly find myself in a position where I can't aim at all because I'm over-compensating for the very fast aiming associated with lower suit mass. For that matter, is there a system in place that dynamically alters the aiming based on mass or are their pre-set determinations on this?
Vrain Matari ZionTCD Unclaimed. 545	Posted - 2013.05.28 12:22:00 - [20] - Quote Sete Clifton wrote: I did a quick analysis of how stick deflection is related to rotation speed. Here are the results I found: https://docs.google.com/spreadsheet/ccc?key=0AqMshc-A6HVHdFlKbklqY1hQWnE4WVJIc1pCVHE4Smc#gid=0 Quick note: The blue line ("current") is how the system works now, and the red line ("proposed") is what I believe it should be. Anyway, as we can see (after accounting for some small error), rotation speed appears to increase linearly with the distance from center that the stick is moved. This is a large part of the reason why I think aiming with the DS3 still feels imprecise and "slippery". The current system makes it way too easy to accidentally build up unwanted turning speed. Most modern FPS it seems use a system where the effectiveness of stick deflection starts off low and increases as you get closer to the edge. This allows you to have a wide range at which you can make smaller and smoother movements, but also still gives you the ability to turn quickly by pushing the stick further. The second line on that graph I provided does this, and is what I propose it should be (in terms of the general design). Reaching 50% rotation speed should happen around 80-85% stick deflection. Basically, the idea is that the deflection scale needs to have more detail/range at the low end, and less at the high end. Which makes sense (at least to me), because when you want to turn really fast, you're already giving up precision, but when you want to make slow fine movements, you want as much detail as possible. Any thoughts on this? Hi Sete. Great work on this, it will help the community and the game greatly, i'm hoping. From my own testing, it looks like the blue curve is missing a 'knee'. I'll qualify that by saying I haven't measured this in the last few days(i'm at work right now), but there is an abrupt and pronounced transition in rotation rate somewhere around mid-deflection. By abrupt i mean a few degrees of deflection and by pronounced i mean maybe a 20% plus increase/decrease in rotation rate Furthurmore, this increase/decrease is instantaneous and is not affected by suit inertia at all. I'll measure it carfully tonight and post results here.

Sete Clifton PSU GHOST SYNDICATE DARKSTAR ARMY 200	Posted - 2013.05.28 13:11:00 - [21] - Quote Vrain Matari wrote: Hi Sete. Great work on this, it will help the community and the game greatly, i'm hoping. From my own testing, it looks like the blue curve is missing a 'knee'. I'll qualify that by saying I haven't measured this in the last few days(i'm at work right now), but there is an abrupt and pronounced transition in rotation rate somewhere around mid-deflection. By abrupt i mean a few degrees of deflection and by pronounced i mean maybe a 20% plus increase/decrease in rotation rate Furthurmore, this increase/decrease is instantaneous and is not affected by suit inertia at all. I'll measure it carefully tonight and post results here. Thanks. As for the "knee" I haven't noticed anything like that, and as Wolfman said above it should be completely linear. But if you're able to find that in testing then that would be very interesting. Definitely post it if you find something.
hgghyujh Expert Intervention Caldari State 6	Posted - 2013.05.29 02:05:00 - [22] - Quote bump! you sir are a god among men, tho really what I would like to see is a graph like that in game that you can tweak your self.
Celus Ivara DUST University Ivy League 13	Posted - 2013.05.29 02:43:00 - [23] - Quote A little more detail on the very low end would be nice. When trying to counter-snipe enemy snipers, I'm trying to line my reticle up with a target that's three pixels wide; doing so with a reticle that likes to take eight pixel jumps is tricky. ...That said though, the difficulty of making this shot gives attentive snipers a chance to notice someone trying to snipe them, prior to the bullet flying through their head; fixing this may make counter-sniping too easy. I think my only real lament here is that presently the difficulty feels artificial. Perhaps, adding more detail on the low end, while also slightly upping the effects of scope sway (especially after sprinting/jumping) may be the best solution.
Artificer Ghost Bojo's School of the Trades 68	Posted - 2013.05.29 04:20:00 - [24] - Quote Moejoe Omnipotent wrote: The sloppy controls is almost entirely due to input lag. Nevermind this, I'm off my game, and forgot what input lag was. That's what was wrong with the grenade throw before, right? Well, sorry, I pretty much said lag is because of PS3 age and CPU space, as I have a brand new PS3 and work fine. :P
Ten-Sidhe Osmon Surveillance Caldari State 492	Posted - 2013.05.30 19:55:00 - [25] - Quote why not add a linearity slider? 0 would be current liner, 100 could be a near logarithmic. Default set where ccp thinks it is best since newberries may not find setting or know what it does. An option to set where 100% is would also be nice, some people using third party controllers may have a different limit on stick travel. I know many use xbox controllers with Cronus chip since they feel dual-shock is to small and girly. I use a dual-shock with billet aluminum aftermarket sticks that allow more travel, so i hit full turn at 80% or so of my stick travel. Some of the people with off brand controller may have less then stock travel and need to adjust the 100% point lower to match their stick travel. I checked a usb pc joystick in ps3, it works. People that want to use a keyboard and joystick would probably love more options for adjustment.
Heligg LA 501 eme Citizens of Nowhere 12	Posted - 2013.05.30 20:26:00 - [26] - Quote How about atan (x)^3?
xharbiex Elements Of Death Elite 0	Posted - 2013.05.31 16:37:00 - [27] - Quote First off- that proposed curve is horrid. I prefer a straight f'n line. I expect my turn speed to be exactly the same at all times throughout the stick movement. But what I prefer shouldn't affect everyone. The real solution to the problem is to give us MAG style control over our DS3. I would to see the following: Setting for dead-zone sensitivity Higher top end on X & Y Setting to adjust top-end acceleration on the horizontal axis. I would like to turn off acceleration in my case for example. Bottom line: give players their freedom to choice the feel they desire and set a default that CCP feels is a good template.
Draco Cerberus Purgatorium of the Damned League of Infamy 100	Posted - 2013.06.06 15:41:00 - [28] - Quote hgghyujh wrote: I would like to see is a graph like that in game that you can tweak your self. This is a great idea, could lead to an "advanced" customization menu in our options menu
Sontie Ill Omens EoN. 622	Posted - 2013.06.07 12:54:00 - [29] - Quote Oh please god ccp, do something like this. Aim controls are so out of date with this game :(
ChromeBreaker SVER True Blood Unclaimed. 555	Posted - 2013.06.07 13:02:00 - [30] - Quote Any update on this?

JonnyAugust Internal Error. Negative-Feedback 230	Posted - 2013.06.07 15:10:00 - [31] - Quote Great analysis. This articulated how I feel too. +100
Scottie MaCallan Krusual Covert Operators Minmatar Republic 25	Posted - 2013.07.10 21:20:00 - [32] - Quote xharbiex wrote: First off- that proposed curve is horrid. I prefer a straight f'n line. I expect my turn speed to be exactly the same at all times throughout the stick movement. But what I prefer shouldn't affect everyone. The real solution to the problem is to give us MAG style control over our DS3. I would to see the following: Setting for dead-zone sensitivity Higher top end on X & Y Setting to adjust top-end acceleration on the horizontal axis. I would like to turn off acceleration in my case for example. Bottom line: give players their freedom to choice the feel they desire and set a default that CCP feels is a good template. just wanted to echo this, like, I actually adjust my sensitivities in other games based on what gun/playstyle I'm using at the moment. I'd also like to suggest one more setting: have a separate adjustable sensitivity for ADS, so that snipers and other long range players can switch more seamlessly between scanning/zoomed out observations & precision firing basically, while I'm all for CCP Wolfman trying to optimize vanilla aiming as much as possible, I'd also really appreciate a full set of options for players with specific preferences or playstyles
Rogatien Merc Red Star. EoN. 916	Posted - 2013.08.31 06:18:00 - [33] - Quote in the spirit of New Eden complexity, CCP could just say "F- it!" and let every player manually adjust every aspect of DS3 sensitivity including that curve. Of course keep working on the default... but hey, why not? Crap, I don't even remember how I got to this thread. Guess I'll necro it anyway.
Jathniel G I A N T EoN. 810	Posted - 2013.08.31 08:30:00 - [34] - Quote This thread is probably the most beautiful and detailed exchange between Dev and Player I have ever seen on these forums... ...brought a tear to me eye...


Pages: 1 2 :: [one page]