rev-vr-podcast-feature-image

It is the week of July 14th, as the first shipments of the much anticipated Oculus Rift DK2 are expected to head to developers, and the VR community has one thing on their mind: when are they going to get an email, or a statement, or something to indicate where their DK2 is? Jesse and I have decided to be patient, and instead talk about a few other subjects to keep our minds off of it.

Jesse is part of the team that is developing VR Chat. They have been doing several regular meetups, including Developer chats on Wednesdays, and a Sunday evening chat after Cymatic Bruce’s stream.

Congratulations to developer, RedofPaw, for  his successful Kickstarter campaign!

Oculus Connect is happening in September, and we are still waiting for more information about it. It seems like everyone in the VR community is scrambling to make plans to go. How much will it cost? What will they show? A mobile SDK? CV1? Controllers? We discuss the possibilities and what it might mean to other hardware developers already in the industry.

One big piece of news released by Oculus is that you need to have V-Sync enabled for the Oculus Rift DK2—some folks barked at that. We discuss what this means and how your experience will be changed by this.

SEE ALSO
Watch Oculus Connect 6 Day Two Keynote Livestream Here @10 AM PT

Is your computer DK2 ready? Are you going to need to buy a new video card? Only time will tell.

I want to thank Jesse for chatting with me in this episode. If you would like a chance to chat with Jesse and join in on the VR Chat fun, check out vrchat.net and watch Reddit’s Oculus section for chat announcements.

Subscribe to Get Episodes as They Air

Subscribe to the Rev VR Podcast on iTunes or Stitcher

Direct download of episode 72: MP3 Download

This article may contain affiliate links. If you click an affiliate link and buy a product we may receive a small commission which helps support the publication. See here for more information.


  • Simon

    Regarding VSYNC; the biggest issue with it is that it ‘hard locks’ you into having the frame ready to display by the time the monitor wants it, if you miss this (even by a micro-second) you have to wait another frame interval to have it displayed (13ms @ 75Hz). With the whole concept of ‘motion to photons’ this could make latency appear much higher.

    What Oculus is banking on is the prediction of head positioning, and tricks like time warping (where they can perform most of the ‘grunt’ work rendering a frame, but still tweaking it slightly before final parts of rendering) to ensure that the displayed frame is as correctly positioned as possible at time of display.

    A good explanation of time warping is here:
    https://www.youtube.com/watch?v=WvtEXMlQQtI&list=UUTlgIRnl3viQb1zCJqFJwqw

  • sp82

    Hi Reverend,

    Why asking for 75fps seems like asking too much?

    It’s 2014 and play triple A games on PC, most of the times, sucks because you got low fps and 70-80ms of input latency, I mean for tiple A games. Even worst with vsync on.

    Play on console sucks too, for me, because it’s impossible plug-in mouse and keyboard on most of the games. WHY? Olso most of the times TVs have 50-100ms of input latency.

    My PC have teraflops of possibilities but I must deal with latency, tearing, stutter and motion blur.

    This sucks.

    I’m software developer and system integator, I spend enogh time optimizing system for speed and correctness to know that there is a plenty of space to optimize software and hardware.
    I like go deep in understanding graphics stuff to, obviously I follow the R&D from Valve and Oculus about VR it’s so exciting moment in the history of VR and this are my thoughts:

    Monitor producers and video standards associations slept for years.
    Today, If you look at the display quality and resolution in laptops and monitors compared to smartphone and tablet displays the situation is embarrassing.

    You must choice: vsync-on or tearing, good image quality or more fps (and lower latency).

    Video transport standards sucks too.
    Vesa specifications, hdmi, displayport and all this stuff works like the monitor is a CRT with fixed refresh rat, but LCDs are not CRTs, come on guys!!!! wake up!!!.

    NVIDIA with GSync maybe pushed VESA to work on that, DisplayPort1.2 and HDMI2 should handle variable refresh rates.
    But we all know that variable rafresh rate and low persistence do not work well with variable refresh rates!!! Damn it!!! Much more R&D required to have the perfect display system.
    Maybe is possible that with very high refresh rate (200+Hz) is possible to handle some little variance in rafresh rate moduling the luminance frame by frame, I don’t know.

    Monitors with motion blur techniques are very rare and more expensive without any reason, it’s only a strobing backlight, and input latency is not very good in the consumer monitors.

    I don’t know any monitor with the perfect combination of features because lack of R&D in the recent years. I’m really disappointed.

    The perfect specs for monitors and HMD should be:
    – high refresh rate (120+Hz), OLED pixels can update thausand times per seconds
    – low persistence, fast switching time, color accuracy and pitch black accuracy, may OLED displays are the definitive display.
    – global display
    – retina resolution
    – smart video transport protocol to copy the front framebuffer from the GPU to a buffer embedded on the display as fast as possible. In this way the buffer on the display is the front buffer and the buffer on the GPU is the rendering buffer, no more double and triple buffer to handle!
    – I want add one important thing to the list: if you eliminate the black gap betwin pixels and blend the color components you can eliminate the screen door effect! Avegant Glyph proof this. You don’t need 10Kx10K displays to eliminate the screen door effect.
    – Time Warping on the HMD itself!! with very low latency from imu input to pixel update, and you can increase the fps as higher as possible.

    With timewarping on the HMD you can reach very high framerate (over thousand fps), but this require hundreds of GB/S from the hardware distortion engine to the pixels,
    you need to put the hardware engine (a dedicated chip) on the back of the display itself. I talk about display created for the VR (8-10 years from now?).
    I ask to Carmack about this. He reply that there is a bandwidth problem from the GPU to the HMD: https://twitter.com/mad_max82/status/464113653466533888
    I think it’s a relative problem, maybe you can double the bandwidth from the GPU to the HMD to transport the z-buffer or you canuse a simplified distorsion algorithm just to timewarp with rotation only correction (eVRydayVR explain this better then me, serch on youtube).

    It’s all about high refresh rate, low latency and low persistence.

    Personally I played FPS for years but now I just play racing games, I just eaging I think, and I’m obsessed with latency problems.
    My setup is very simple:
    – very old 60Hz low input latency monitor with not so high resolution.
    – vsync off
    – I setup the racing simulator to do almost 200fps, lowering the settings.
    – After that set a limit on the fps to 120fps, it’s a option on the game settings, it’s not a vsync or something like that, it’s just a sofware limit on the number of fps.
    – disable triple buffer on the graphics driver settings.
    – I use downsampling from the resolution form 1920×1080(rendering resolution) to 1680×1050(native monitor resolution) for better details, better than antialiasing.

    with this setup I get very low fps.

    Sorry, this is just my open letter to the display manufactors, game studios and video standard associations to innovate and to not sleep.

    • sp82

      edit
      ….. with this setup I get very low fps…..
      in
      …. with this setup I get very low latency….