2K and Hangar 13 Mess up Mafia III release with dumb 30 FPS cap

Discussion in 'Frontpage news' started by Hilbert Hagedoorn, Oct 7, 2016.

  1. sammarbella

    sammarbella Guest

    Messages:
    3,929
    Likes Received:
    178
    GPU:
    290X Lightning CFX (H2O)
    "General people" (PC gamers) dumbness and patience have limits, Batman AK was a milestone for both.

    Refund policies on PC and bad image from real life performance vs hype can overcome millions in marketing.
     
  2. Neo Cyrus

    Neo Cyrus Ancient Guru

    Messages:
    10,776
    Likes Received:
    1,388
    GPU:
    黃仁勳 stole my 4090
    Not only do the eyes and mind not work like that, not even all parts function at the same speed (or resolution). Any article claiming any sort of human limitation along those lines is rubbish; they always lack any real science behind them.

    I don't need to remind you that reaction time and how many "frames" a person can see differs wildly between people. A fighter jet pilot, a pro gamer, an average Joe, and a 50 year old woman who enjoys her Wii U are not going to perceive and process the same amount of data per second.

    I play League of Legends locked at 300 fps (through Afterburner) and I can't go back to a vsync'd 60 fps because it makes the game look like a slideshow. The smoothness difference is a massive enough gap that I overlook screen tearing and go with the higher frame rate. I've tried lower caps just to see what it looks like, and 200 or lower frame rates are choppy enough compared to 300 that it makes the screen tearing unbearable. Whoever wrote an article about humans not being able to perceive high frame rates must actually be retarded with severely degraded senses.
     
    Last edited: Oct 10, 2016
  3. SA SYS

    SA SYS Guest

    Messages:
    81
    Likes Received:
    0
    GPU:
    GEFORCE GTX 1080 FE
    Guys why the hell guru3d didnt post Gears 4 benchmark???
    cuz it's UE4 and nvidia wins ? :| ( anyway nvidia allways wins in frame time ;) )

    :cheers::thumbup:
     
  4. yasamoka

    yasamoka Ancient Guru

    Messages:
    4,875
    Likes Received:
    259
    GPU:
    Zotac RTX 3090
    What you're seeing is less latency from rendering at 300FPS (5 frames per refresh) and display the last (5th frame) instead of the first (had you been rendering at 60FPS). That way you get 16.67 / 5 ms = 3.33ms of frame latency instead of 16.67ms.

    Your monitor is refreshing at 60Hz. You're still seeing 60FPS, no more. Just less latency. Wrong example, right idea.

    If you're not getting a perfectly smooth VSynced 60FPS (it has to be smoother than any other FPS given it is perfectly synchronized with the display refresh rate) then you have an issue elsewhere.

    Anyways, yes, the human eye and brain do not see in frames.
     

  5. Neo Cyrus

    Neo Cyrus Ancient Guru

    Messages:
    10,776
    Likes Received:
    1,388
    GPU:
    黃仁勳 stole my 4090
    What? If the latency is 3.33 ms and every 5th frame is output why would that cause perceived smoothness to increase if the spacing of each frame is still 16.67 ms?

    All these years I never stopped to think about any of this.
    I'm not. In the example I gave running it at 300 fps for example allows me to clearly see when a character is turning. At 60 fps it's a slideshow in comparison. No nothing like 30 fps or even 50 fps would be, but I can't see as many frames of the character turning. At 300 fps I can super clearly see it, which is why I didn't stop to think and thought it was actually displaying more fps.
     
    Last edited: Oct 10, 2016
  6. yasamoka

    yasamoka Ancient Guru

    Messages:
    4,875
    Likes Received:
    259
    GPU:
    Zotac RTX 3090
    I am referring to a very particular latency, the part of the whole latency chain that is due to frame rendering. That is, when a frame starts rendering, it takes input at that instant. Let's call that instant t0 = 0s. Since you're rendering at 300FPS, that frame latency is 3.33ms. Your monitor refreshes and displays that frame, showing the result of your input 3.33ms later. Of course, the whole latency chain is a lot longer, but you basically reduce your frame latency from 16.67ms to 3.33ms.

    Let's assume, as a first example, perfectly paced frame delivery, at 60FPS then at 300FPS. | indicates a monitor refresh.

    60FPS: 01--> | 02--> | 03--> | 04--> |
    300FPS: 01 02 03 04 05 | 06 07 08 09 10 | 11 12 13 14 15 | 16 17 18 19 20 |

    Your monitor is refreshing every 16.67ms. Your frames are being rendered every 3.33ms --> 4 out of 5 frames are being dropped (with old information) --> you're not seeing any additional frames. You're seeing more recent frames. That's all.

    Remember, this is assuming you have no stutter from other external factors at 60FPS. If you do, then rendering at 300FPS will give more even FPS delivery (which manifests into you seeing 'additional' frames - they're simply more even paced frames).

    To test if your 60FPS VSync is perfectly smooth (no stutter), pan around the camera using the keyboard. You should see no jumping in the camera.

    Let's now assume there is frame variance. 5ms variance at 60FPS would become 1ms at 300FPS (percentage of frametime) --> each frame displayed has less variance and you have a more even output.

    It's very likely then that you do not have an even frame delivery at 60FPS VSync. Try the strafe test I mentioned and see if that works out for you.
     
    Last edited: Oct 10, 2016
  7. Neo Cyrus

    Neo Cyrus Ancient Guru

    Messages:
    10,776
    Likes Received:
    1,388
    GPU:
    黃仁勳 stole my 4090
    I took a moment to think about it and edit my post. Quite a while ago, yours took a while. If you could please update yours to answer my updated one that'd be appreciated.
     
  8. yasamoka

    yasamoka Ancient Guru

    Messages:
    4,875
    Likes Received:
    259
    GPU:
    Zotac RTX 3090
    Oh sure! Will do.
     

Share This Page