Aller au contenu

Photo

How frame rate affects game difficulty [UPDATE 13/03/2013]


  • Veuillez vous connecter pour répondre
412 réponses à ce sujet

#101
Menthus

Menthus
  • Members
  • 75 messages

Caratinoid wrote...

No you can set any frame rate you want as client, if your pc can handle it, but there a system in place called frame rate smoothing (active by default) which will alter you frame rate depending on the processing power of your machine. So you can start a game at 30 FPS but then you'll get a map that needs more resources to render or look at a lot of enemies at once which can lock your frame rate at a lower value (22 FPS is the lowest it can set) and will stay there even if your computer can pump out more frames.


Thanks again Cara.

So the Client frame rate has nothing to do with the host? Whenever I get visual lags, the frame rate seems to get stuck at ~22FPS and never improves, even when I move to a smaller area with less enemies. Such visual lag also goes away when I host my own games, on the same map and with the same enemy. Thus I suspect the lag comes from the host. I don't really have solid proofs though.

How about the enemy's shooting accuracy? Is it dependent on the frame rate on the host, or the client?

#102
OneMore1968

OneMore1968
  • Members
  • 792 messages
This is a quote taken from the developer's of the Unreal Engine 3.0 on which ME3 is based:

We’re 30 frames per second yet again, locked across the board. You can really feel the difference in the controls.” by Jesse Houston

If it is played at 30+ fps then maybe we should expect strange behaviour from time to time.

#103
Caratinoid

Caratinoid
  • Members
  • 982 messages

Menthus wrote...

Thanks again Cara.

Cara? Hm ... have never been called that yet, but then again my tag doesn't really mean anyting, didn't knew it would be seen by other peole, I think it was something I came up with while studing biochemistry.

Menthus wrote...

So the Client frame rate has nothing to do with the host?

Correct.

Menthus wrote...

How about the enemy's shooting accuracy? Is it dependent on the frame rate on the host, or the client?

The host.

Modifié par Caratinoid, 28 janvier 2013 - 12:34 .


#104
Daihannya

Daihannya
  • Members
  • 763 messages

Caratinoid wrote...

.458 wrote...

This game has a forced vsync, so no matter how fast your video is, it will lower if your monitor rate drops. On most flat screen LCD types, this means 60 fps. I've noticed in the load screens it locks to 30 fps.

About host calculations...not all of this is done in host. Most of what you mention is indeed host only. Rules as to whether you were capable of hitting or getting hit depend on the host. Rules to where you can be are on the host...this is why there is rubber-banding in lag, your computer is getting set back to what the host thinks you are.

There is a small amount of information that IS dependent on the off-host though...fine aiming comes to mind. When you set up that head shot, there is no way it would work if there were a quarter second or half second delay from the round trip to the host. Frame rate will have a profound effect on how easy it is to fine-aim, and thus a small part of the game (but an important part) does indeed get modified by a fast frame rate.

Also, if you use warp or throw, and it tracks to the target but does no damage, this proves the animation is off-host while actual impact rules are host-only. Animations are always local to the off-host. FYI, this is why a network can lose an event and animations like throwing a grenade appear to work but it doesn't actually do anything (consider that as long as you have a grenade in inventory, it will always work...host will never disagree without a major network failure).

There is another case I speculate about, and not sure of. That is the case of whether hits to a target with some weapons can be a "glancing" blow, thus causing less damage. It does not seem possible that a bullet type weapon could have a partial damage or reduced damage without this rule...and with network lag, damage seems to sometimes be reduced even if the target is hit. So it seems that possibly a glancing blow is considered, and that this is host-dependent, while fine aim like a head shot is off-host dependent.


Hit testing on enemies is performed off host.

Hit testing on players is performed on host.


Are you sure this is 100%?    

The wierd occurrance I am trying to understand is when  the host has a partially laggy client.    It seems to effect even folks with good connections.   Case in point I play with a group of friends, all within California and Nevada.   One of us would typically leave, leaving a space for one pug.  Given the hours we play we get a kiwi, ausy, or someone from asia as a PUG.   When that person joins, our regular host becomes a slightly better killing machine.  I am downed a lilttle more than when playing with my friends purely on the west coast. I also get the I was in freaking cover, when he shot me. Yet host didn't change, just a different client.  

I also was playing with stardusk once.  Only played with him once, we have a common friend.  But I really hope I never play with him again.  Host was also in california, but when he joined the match for me it went completely south. Everything became supper laggy.  My connection to our common friend has always been a little laggy, but when he was playing I never got hit, but wasn't hitting squat.  Given it was a reaper match, I ended up sitting out a number of waves.  I couldn't get hit, but the invisible banshees were out for my blood,  and sync killed me every round.  

And to answer your questions.  I play at 60FPS consistently its vsync'd to my monitor so it caps it there.   There is also an older laptop which is only capable of around 30-40FPS.   Surviablility is definately easier when using the laptop.  But because the screen is smaller, its harder for me to actually hit things when shooting. 

Modifié par Daihannya, 28 janvier 2013 - 12:14 .


#105
Caratinoid

Caratinoid
  • Members
  • 982 messages

Carlina wrote...

This is a quote taken from the developer's of the Unreal Engine 3.0 on which ME3 is based:

We’re 30 frames per second yet again, locked across the board. You can really feel the difference in the controls.” by Jesse Houston

If it is played at 30+ fps then maybe we should expect strange behaviour from time to time.

Yeah, I'm pretty sure It should be played at 30 FPS to get the intended difficulty. Which to me visually looks like s***.<_<

Daihannya wrote...

Caratinoid wrote...

Hit testing on enemies is performed off host.

Hit testing on players is performed on host.

Are you sure this is 100%?

It would be impossible to hit anything if it wasn't so.

Modifié par Caratinoid, 28 janvier 2013 - 12:22 .


#106
P3G4SU5

P3G4SU5
  • Members
  • 346 messages
Playing at 120FPS:

Just played a couple games; the first at the default locked frame rate of 62FPS and the second modified to 120FPS. My system is higher end and frame rate fluctuated by only ~1-4 frames (barely noticeable) during the 120FPS game. The difference between 62FPS and 120FPS is pretty noticeable.

AI performance aside, the ease of aiming and camera movement improves gameplay. There is a significant increase in overall accuracy to both movement and shooting - shots were easier to line up and animations of enemies and players seemed much smoother. Altogether the game just feels more fluid. I don't think anyone that tries the game at 120FPS would be able to go back from it. I know I certainly won't.

Tl;dr    Try the game at 120FPS if your system can handle it, you won't regret it.:o

#107
Caratinoid

Caratinoid
  • Members
  • 982 messages

P3G4SU5 wrote...

Playing at 120FPS:

Just played a couple games; the first at the default locked frame rate of 62FPS and the second modified to 120FPS. My system is higher end and frame rate fluctuated by only ~1-4 frames (barely noticeable) during the 120FPS game. The difference between 62FPS and 120FPS is pretty noticeable.

AI performance aside, the ease of aiming and camera movement improves gameplay. There is a significant increase in overall accuracy to both movement and shooting - shots were easier to line up and animations of enemies and players seemed much smoother. Altogether the game just feels more fluid. I don't think anyone that tries the game at 120FPS would be able to go back from it. I know I certainly won't.

Tl;dr    Try the game at 120FPS if your system can handle it, you won't regret it.:o


Agree. So, you tried hosting platinum yet?:D

Usually not a problem for the host because you can react quicker to everything but clients will have it tough. That's my main problem here, just want to be able to host without making it an elite only lobby.

#108
Daihannya

Daihannya
  • Members
  • 763 messages

Caratinoid wrote...

Carlina wrote...

This is a quote taken from the developer's of the Unreal Engine 3.0 on which ME3 is based:

We’re 30 frames per second yet again, locked across the board. You can really feel the difference in the controls.” by Jesse Houston

If it is played at 30+ fps then maybe we should expect strange behaviour from time to time.

Yeah, I'm pretty sure It should be played at 30 FPS to get the intended difficulty. Which to me visually looks like s***.<_<

Daihannya wrote...

Caratinoid wrote...

Hit testing on enemies is performed off host.

Hit testing on players is performed on host.

Are you sure this is 100%?

It would be impossible to hit anything if it wasn't so.


I can see the weapons working this way, but the hit testing on players is what I am questioning.   It would explain why diving into cover doesn't help on a laggy connection.  But why would a change in a client to the host, but not the host itself effect my hit rate?

#109
GallowsPole

GallowsPole
  • Members
  • 4 216 messages
Why indeed?

#110
Caratinoid

Caratinoid
  • Members
  • 982 messages

Daihannya wrote...

Caratinoid wrote...

Carlina wrote...

This is a quote taken from the developer's of the Unreal Engine 3.0 on which ME3 is based:

We’re 30 frames per second yet again, locked across the board. You can really feel the difference in the controls.” by Jesse Houston

If it is played at 30+ fps then maybe we should expect strange behaviour from time to time.

Yeah, I'm pretty sure It should be played at 30 FPS to get the intended difficulty. Which to me visually looks like s***.<_<

Daihannya wrote...

Caratinoid wrote...

Hit testing on enemies is performed off host.

Hit testing on players is performed on host.

Are you sure this is 100%?

It would be impossible to hit anything if it wasn't so.


I can see the weapons working this way, but the hit testing on players is what I am questioning.   It would explain why diving into cover doesn't help on a laggy connection.  But why would a change in a client to the host, but not the host itself effect my hit rate?


Yes diving into cover doesn't help because you are still standing out of it on the hosting players machine, which makes you look like a noob. Same with shooting something 3 times in the face just to die yourself after a short delay.

If that player is lagging very bad then it will make him less of a priority I think so the game will seem more aggressive to other players. I think this is more a question  for an aggro mechanics thread. Sometimes you can see some weird things happen like an infiltrator in clock being targeted more than other 3 players. In some games enemies don't shoot at me at all for some reason. Also it is possible that the host was recording at that time at a lower frame rate or something.

#111
Caratinoid

Caratinoid
  • Members
  • 982 messages
Anyway, gtg now, and I better not find this thread locked when I come back!:police:

#112
.458

.458
  • Members
  • 2 113 messages

Caratinoid wrote...

.458 wrote...

This game has a forced vsync, so no matter how fast your video is, it will lower if your monitor rate drops. On most flat screen LCD types, this means 60 fps. I've noticed in the load screens it locks to 30 fps.

About host calculations...not all of this is done in host. Most of what you mention is indeed host only. Rules as to whether you were capable of hitting or getting hit depend on the host. Rules to where you can be are on the host...this is why there is rubber-banding in lag, your computer is getting set back to what the host thinks you are.

There is a small amount of information that IS dependent on the off-host though...fine aiming comes to mind. When you set up that head shot, there is no way it would work if there were a quarter second or half second delay from the round trip to the host. Frame rate will have a profound effect on how easy it is to fine-aim, and thus a small part of the game (but an important part) does indeed get modified by a fast frame rate.

Also, if you use warp or throw, and it tracks to the target but does no damage, this proves the animation is off-host while actual impact rules are host-only. Animations are always local to the off-host. FYI, this is why a network can lose an event and animations like throwing a grenade appear to work but it doesn't actually do anything (consider that as long as you have a grenade in inventory, it will always work...host will never disagree without a major network failure).

There is another case I speculate about, and not sure of. That is the case of whether hits to a target with some weapons can be a "glancing" blow, thus causing less damage. It does not seem possible that a bullet type weapon could have a partial damage or reduced damage without this rule...and with network lag, damage seems to sometimes be reduced even if the target is hit. So it seems that possibly a glancing blow is considered, and that this is host-dependent, while fine aim like a head shot is off-host dependent.


Hit testing on enemies is performed off host.

Hit testing on players is performed on host.


Which makes sense...enemies don't seem to do head shots.

PS: I'll run when I see an enemy with Dog of War banner :lol:

#113
Shampoohorn

Shampoohorn
  • Members
  • 5 861 messages
OP, thanks for the insights. A shame about the pissing match in the middle of the thread though.

#114
oOmpie

oOmpie
  • Members
  • 342 messages

Lord_Sirian wrote...

GallowsPole wrote...

Man, I love pushing buttons to the easily pushed.

 

Who writes these lines...? How can you be so illiterate? Is English even your first language?


Damn, dude... you just went full retard...

#115
Deerber

Deerber
  • Members
  • 16 836 messages

oOmpie wrote...

Lord_Sirian wrote...

GallowsPole wrote...

Man, I love pushing buttons to the easily pushed.

 

Who writes these lines...? How can you be so illiterate? Is English even your first language?


Damn, dude... you just went full retard...


Hey! The children left, better leave it as is, ok? Thanks.


Caratinoid wrote...

Menthus wrote...

Thanks again Cara.

Cara? Hm ... have never been called that yet, but then again my tag doesn't really mean anyting, didn't knew it would be seen by other peole, I think it was something I came up with while studing biochemistry.


It feels strange to me, since cara in italian means... Darling. A female, darling :?


Caratinoid wrote...

UnknownMercenary wrote...

HolyAvenger wrote...

I still haven't played on your host Caratinoid. Sling me an invite next time you have space, I want to see this for myself.


Bring plenty of  lube  oops packs.


Hey don't act like you have seen it all, that was only 120, we haven't tried 300 yet!:D

EDIT: and geth:devil:


Ok, I'm officially proposing this. We do a Platinum match with Caratinoid hosting at 200+ FPS, on Condor against Geth. As Drell Adepts.

... If we make it to wave 2, I'd say we're awesome players B)

Modifié par Deerber, 28 janvier 2013 - 08:30 .


#116
Guest_Lord_Sirian_*

Guest_Lord_Sirian_*
  • Guests

Deerber wrote...

Hey! The children left, better leave it as is, ok? Thanks.

 

You realise your holier-than-thou BS makes you utterly unlikeable, don't you?

#117
Feneckus

Feneckus
  • Members
  • 3 076 messages
Playing at 120 FPS does make a difference. Obviously, the difference isn't as big as between 30 FPS (:sick:) and 60, but it's definitely noticeable.

#118
P3G4SU5

P3G4SU5
  • Members
  • 346 messages

Deerber wrote...

oOmpie wrote...

Lord_Sirian wrote...

GallowsPole wrote...

Man, I love pushing buttons to the easily pushed.

 

Who writes these lines...? How can you be so illiterate? Is English even your first language?


Damn, dude... you just went full retard...


Hey! The children left, better leave it as is, ok? Thanks.


Caratinoid wrote...

Menthus wrote...

Thanks again Cara.

Cara? Hm ... have never been called that yet, but then again my tag doesn't really mean anyting, didn't knew it would be seen by other peole, I think it was something I came up with while studing biochemistry.


It feels strange to me, since cara in italian means... Darling. A female, darling :?


Caratinoid wrote...

UnknownMercenary wrote...

HolyAvenger wrote...

I still haven't played on your host Caratinoid. Sling me an invite next time you have space, I want to see this for myself.


Bring plenty of  lube  oops packs.


Hey don't act like you have seen it all, that was only 120, we haven't tried 300 yet!:D

EDIT: and geth:devil:


Ok, I'm officially proposing this. We do a Platinum match with Caratinoid hosting at 200+ FPS, on Condor against Geth. As Drell Adepts.

... If we make it to wave 2, I'd say we're awesome players B)



Count me in :D.

Note that when there is no frame rate cap though, the host will experience very strange mouse behaviour. Movement becomes inconsistent so the same mouse movement will turn the camera by different degrees, depending on the situation.

Modifié par P3G4SU5, 28 janvier 2013 - 08:48 .


#119
Brohak

Brohak
  • Members
  • 263 messages
 So basically, my **** laptop that gives slightly under 30 FPS since release makes me better because enemies can't hit me ? I should go back to Bronze. :crying:

#120
Caratinoid

Caratinoid
  • Members
  • 982 messages
@P3G4SU5

I can set any fixed framerate I want in the game config file so my mouse aiming will be fine. I can go up to 300 fixed because my minimum frame rate when in overclock is around 300, it goes much higher (up to 500) in some places but playing with a jumpy frame rate like that is horrible like you said.

#121
Caratinoid

Caratinoid
  • Members
  • 982 messages
@Deerber

Drell have very low shields and will hit the no regen bug sooner than other char, the only way to win would be to not get hit at all since one burst is enough to kill drell even with full shields, which will be pretty difficult on platinum with increased accuracy.

#122
udinbak4

udinbak4
  • Members
  • 587 messages
I remember this thread.

Biowur, pls tell us what is the intended framerate.

#123
Caratinoid

Caratinoid
  • Members
  • 982 messages
@udinbak4

Yeah threads like that usually speak about very low frame rates, which I don't really care about, I just want to be able to host at a higher frame rate without making my lobby exlusive to elite players (PUGs will just leave me soloing last waves).

#124
spudspot

spudspot
  • Members
  • 2 447 messages

Caratinoid wrote...

@udinbak4

Yeah threads like that usually speak about very low frame rates, which I don't really care about, I just want to be able to host at a higher frame rate without making my lobby exlusive to elite players (PUGs will just leave me soloing last waves).


Is it really that much of a difference? My guess is that my standard host plays at solid 60fps and I can definitely tell the difference in terms of enemy accuracy between that and hosting with my crappy 20fps tops notebook. 

#125
Caratinoid

Caratinoid
  • Members
  • 982 messages
Like I said for skilled players it doesn't really matter unless they have a high network latency because enemies can not shoot you if they keep dieing after 1 second as they spawn. It's the moments when things get tense that count, like wave 9 on platinum collectors with all those phantoms, you can't even get away after using a medi-gel (which usually works at a lower frame rate because you will be running). Basically when the 's**^ hits the fan you can almost always just run away which doesn't always work at high frame rates.