3 ms·

a lot of people are using 360 fps now, which may actually be unnecessary. playing on 240 fps feels extremely different. it’s about how the occasional rapid mov

a lot of people are using 360 fps now, which may actually be unnecessary. playing on 240 fps feels extremely different. it’s about how the occasional rapid movement feels. spinning around, or making a series of edits in a fight. in a typical interactive movie type game with a controller, the maximum delta-v of camera or character is very low. 60 fps feels smooth the same way movies are fine at 30, a lack of low latency high delta interaction. in fortnite with a 24x48 inch mousepad, delta-v peaks pretty high, and when frame time goes above 4ms, you notice. low latency interaction makes the world feel more real in a way graphics can’t. graphics are important, but secondary.

paulddraper 3y ago
I can believe that 60 to 120 is a noticable difference to the careful observer. But 120 to 240? Really?
- nathants 3y ago
  definitely! it’s easy to mess around with since you can limit framerate to a lower level. the other part is frame dips. no matter the game, frames drop sometimes. a drop from 240 to 180 will be harder to notice than a drop from 120 to 80. 4ms consistent frame time is a challenging engineering target. whatever game dethrones fortnite will improve client side performance and reduce the effect of server ping on client side interactions.
  - paulddraper 3y ago
    You're assuming it will drop from 240 to 180. I don't see why it wouldn't drop from 240 to 80.
    - nathants 3y ago
      it definitely could. unreal engine doesn’t seem to prioritize high performance, nor does the fortnite dev team. in fortnite competitive play, as soon as too many players are in the same part of the world, client and server performance goes south fast. in typical casual play, this never happens. in competitive play, it always happens. i’m exploring gamedev out of curiosity for this exact discussion. the metric i’m trying to optimize for is minimum frame rate[1]. frame rate is top left, min on left, average on right. video is 120 fps. i’m not sure yet what’s possible in varying scenarios of complexity, but there is definitely room for improvement! most interesting so far is how fast physx is. graphics are the bottleneck, not physics. 1. https://r2.nathants.workers.dev/world_building.mp4 https://r2.nathants.workers.dev/world_building.mp4

paulddraper 3y ago

I can believe that 60 to 120 is a noticable difference to the careful observer. But 120 to 240? Really?

nathants 3y ago
definitely! it’s easy to mess around with since you can limit framerate to a lower level. the other part is frame dips. no matter the game, frames drop sometimes. a drop from 240 to 180 will be harder to notice than a drop from 120 to 80. 4ms consistent frame time is a challenging engineering target. whatever game dethrones fortnite will improve client side performance and reduce the effect of server ping on client side interactions.
- paulddraper 3y ago
  You're assuming it will drop from 240 to 180. I don't see why it wouldn't drop from 240 to 80.
  - nathants 3y ago
    it definitely could. unreal engine doesn’t seem to prioritize high performance, nor does the fortnite dev team. in fortnite competitive play, as soon as too many players are in the same part of the world, client and server performance goes south fast. in typical casual play, this never happens. in competitive play, it always happens. i’m exploring gamedev out of curiosity for this exact discussion. the metric i’m trying to optimize for is minimum frame rate[1]. frame rate is top left, min on left, average on right. video is 120 fps. i’m not sure yet what’s possible in varying scenarios of complexity, but there is definitely room for improvement! most interesting so far is how fast physx is. graphics are the bottleneck, not physics. 1. https://r2.nathants.workers.dev/world_building.mp4 https://r2.nathants.workers.dev/world_building.mp4