I wrote previously a post about how crappy most of the current VR games available at Steam look like. And no, I didn't deliberately pick&choose the worst examples I could find. I took directly the top rated VR games on Steam, without any kind of bias.
When I comment about this online, many people jump to the defense, stating that VR is so resource-heavy that you just can't expect even a VR-capable PC to render modern games at the required resolution and refresh rate. After all, it has to render the view of the game twice, 90 times per second minimum. That's like almost triple compared to your regular games, right?
Well, no. For some reason these people seem to think that the game needs to render the game at full resolution twice (thus requiring twice the rendering speed), and at a significantly increased framerate at that (compared to the standard 60 fps.) In actuality the resolution for each eye is 1080x1200 pixels. That sounds like a rather high resolution, but it's actually just 62.5% of your standard 1920x1080 resolution.
The total amount of pixels to render for (current) VR is 2160x1200. This is 25% more than the standard 1920x1080. It's more resource intensive yes, but not by all that much.
So, I went ahead and did a small test to see what VR could realistically look like, taking into account how resource intensive it is. I performed the test as follows.
I do not have a monitor capable of VR resolution, so I had to rely on using the 1920x1080 resolution. However, since what matters is how many pixels per second the hardware is capable of rendering, 2160x1200@90Hz is roughly equivalent to 1920x1080@112.5Hz. (Well, mathematically they are exactly equal in terms of pixels per second, but in terms of rendering speed there are other variables in play as well, thus they are "roughly" equivalent.)
I think that it's pretty safe to say that if a min-spec PC is capable of rendering a game at 1920x1080@120Hz, then it will be capable of rendering it in VR. (Any additional CPU processing required for VR is more than accounted for in that extra framerate. After all, the higher the framerate, the more taxing it becomes for the CPU.)
My PC just happens to be perfect for this test because it barely meets the minimum specs recommended for VR: It's an i5-2500K with a GTX970 graphics card. (The minimum recommended spec for VR has a CPU that's just a tad bit faster than mine, so this is actually perfect. If my PC is capable of rendering a game at 1920x1080@120Hz, then an actual min-spec PC will also most certainly.)
So I ran some modern games with vsync turned off, and adjusting graphics settings if needed, with the goal of seeing if they could run at a minimum of 120Hz.
So, here are some screenshots of what VR games could look like (click on the images to get full resolution versions):
The Talos Principle:
Battlefield 4:
Alien: Isolation. (For some reason Steam's screenshot feature did not grab the MSI Afterburner overlay in this game, so you'll just have to trust me, but the framerates were well over 130Hz even at maximum graphical settings.)
When I comment about this online, many people jump to the defense, stating that VR is so resource-heavy that you just can't expect even a VR-capable PC to render modern games at the required resolution and refresh rate. After all, it has to render the view of the game twice, 90 times per second minimum. That's like almost triple compared to your regular games, right?
Well, no. For some reason these people seem to think that the game needs to render the game at full resolution twice (thus requiring twice the rendering speed), and at a significantly increased framerate at that (compared to the standard 60 fps.) In actuality the resolution for each eye is 1080x1200 pixels. That sounds like a rather high resolution, but it's actually just 62.5% of your standard 1920x1080 resolution.
The total amount of pixels to render for (current) VR is 2160x1200. This is 25% more than the standard 1920x1080. It's more resource intensive yes, but not by all that much.
So, I went ahead and did a small test to see what VR could realistically look like, taking into account how resource intensive it is. I performed the test as follows.
I do not have a monitor capable of VR resolution, so I had to rely on using the 1920x1080 resolution. However, since what matters is how many pixels per second the hardware is capable of rendering, 2160x1200@90Hz is roughly equivalent to 1920x1080@112.5Hz. (Well, mathematically they are exactly equal in terms of pixels per second, but in terms of rendering speed there are other variables in play as well, thus they are "roughly" equivalent.)
I think that it's pretty safe to say that if a min-spec PC is capable of rendering a game at 1920x1080@120Hz, then it will be capable of rendering it in VR. (Any additional CPU processing required for VR is more than accounted for in that extra framerate. After all, the higher the framerate, the more taxing it becomes for the CPU.)
My PC just happens to be perfect for this test because it barely meets the minimum specs recommended for VR: It's an i5-2500K with a GTX970 graphics card. (The minimum recommended spec for VR has a CPU that's just a tad bit faster than mine, so this is actually perfect. If my PC is capable of rendering a game at 1920x1080@120Hz, then an actual min-spec PC will also most certainly.)
So I ran some modern games with vsync turned off, and adjusting graphics settings if needed, with the goal of seeing if they could run at a minimum of 120Hz.
So, here are some screenshots of what VR games could look like (click on the images to get full resolution versions):
The Talos Principle:
Battlefield 4:
Alien: Isolation. (For some reason Steam's screenshot feature did not grab the MSI Afterburner overlay in this game, so you'll just have to trust me, but the framerates were well over 130Hz even at maximum graphical settings.)
Comments
Post a Comment