WEBVTT

00:00:00.160 --> 00:00:08.080
virtual reality or vr has been hailed as the next paradigm for gaming and

00:00:04.960 --> 00:00:10.719
immersion ever since it rolled out then

00:00:08.080 --> 00:00:16.320
rolled over back in the late 90s with early implementations like sony's glass

00:00:12.960 --> 00:00:20.160
tron being described by john carmack as

00:00:16.320 --> 00:00:23.199
like looking through toilet paper tubes

00:00:20.160 --> 00:00:25.840
fast forward to 2016 though and the tech

00:00:23.199 --> 00:00:30.800
was awesome but it still had a big problem the price

00:00:28.080 --> 00:00:35.760
because on top of the headset gamers in particular needed to shell out

00:00:33.280 --> 00:00:39.520
for much more powerful hardware than they would need to play the same game in

00:00:38.160 --> 00:00:44.480
flat mode but there's some good news

00:00:41.840 --> 00:00:48.320
since that time oculus who sponsored this video has dropped the price of the

00:00:46.800 --> 00:00:52.960
rift and continued development on some pretty

00:00:50.480 --> 00:00:58.160
neat invisible mechanisms that work in the background to make vr run on even

00:00:55.520 --> 00:01:03.359
relatively modest machines and we spent the last week testing to see

00:01:00.399 --> 00:01:14.799
just how modest a machine you can use and i think you'll be surprised

00:01:14.799 --> 00:01:21.280
let's begin with some of the efficiency improvements that are handled by your

00:01:18.320 --> 00:01:25.280
graphics card because in theory every frame of an image in vr needs to

00:01:23.600 --> 00:01:29.920
be rendered twice once for each eye

00:01:27.520 --> 00:01:35.280
so technologies like NVIDIA's single pass stereo make it so your pc only

00:01:32.720 --> 00:01:41.360
needs to set up the scenes geometry once and their lens matched shading actually

00:01:38.320 --> 00:01:43.600
cuts off any pixels which would not be

00:01:41.360 --> 00:01:46.479
visible anyway due to how the lenses in the headset work

00:01:45.040 --> 00:01:52.159
pretty neat then we've got the headset specific tech

00:01:49.280 --> 00:01:57.119
like oculus's asynchronous time warp and space warp which kick in any time your

00:01:54.799 --> 00:02:01.360
machine can't maintain the steady 90 frames per second that's required to

00:01:58.960 --> 00:02:07.360
prevent motion sickness by generating synthetic preemptive frames that act

00:02:04.399 --> 00:02:12.959
like a sort of suspension mechanism to smooth out the low FPS bumps

00:02:10.800 --> 00:02:17.760
getting this tech working right was quite a bit of work because

00:02:14.879 --> 00:02:21.920
gpus are designed for high throughput but not

00:02:18.800 --> 00:02:24.400
preemptive frame guessing so oculus

00:02:21.920 --> 00:02:30.879
worked with both GPU manufacturers as part of NVIDIA's vr works and AMD's

00:02:27.760 --> 00:02:32.400
liquid vr initiatives so it's more than

00:02:30.879 --> 00:02:37.360
a catchy slogan now let's introduce our bare minimum

00:02:34.879 --> 00:02:45.599
system so we paired uh where is it here it is a core i3 8100 quad core processor

00:02:41.519 --> 00:02:48.480
and a gtx 1050 ti and ran future marks

00:02:45.599 --> 00:02:53.599
vr mark which has a neat feature both the rift's minimum and vr ready

00:02:51.599 --> 00:02:59.360
performance requirements are baked into the results graph and what we found was

00:02:56.400 --> 00:03:05.519
that even our bottom of the barrel actually managed to meet the minimum

00:03:01.840 --> 00:03:07.840
spec bar so there you go video done

00:03:05.519 --> 00:03:13.200
peace out right no no wrong so upon further analysis we

00:03:10.319 --> 00:03:18.720
actually found that while our rig was outputting 90 frames per second in the

00:03:15.599 --> 00:03:20.080
basic orange benchmark which matches up

00:03:18.720 --> 00:03:25.000
with the refresh rate of the head mounted displays in the rift

00:03:22.560 --> 00:03:30.799
the actual target is 109 frames per second this is because

00:03:28.400 --> 00:03:35.599
aside from the graphics going on on the screen in vr the system also needs to

00:03:33.760 --> 00:03:40.159
read sensors and do a bunch of stuff in the background so

00:03:37.120 --> 00:03:43.519
with the overhead of vr the calculation

00:03:40.159 --> 00:03:47.040
is that 109 frames per second in flat

00:03:43.519 --> 00:03:50.319
mode should correspond to 90 frames per

00:03:47.040 --> 00:03:51.680
second in vr and superposition basically

00:03:50.319 --> 00:03:57.680
said the same thing so that setup was only good enough for

00:03:54.560 --> 00:03:58.480
like the very barest minimum spec

00:03:57.680 --> 00:04:02.319
so we do need to go a little higher but do

00:04:00.959 --> 00:04:06.799
we go for CPU or GPU

00:04:04.959 --> 00:04:11.840
well we were watching our CPU load pattern during the runs and our

00:04:08.720 --> 00:04:14.720
processor even though it's a measly core

00:04:11.840 --> 00:04:20.079
i3 well it is a quad core and it didn't look that busy so we bumped our graphics

00:04:17.280 --> 00:04:23.600
up to a gtx 1060 to see where that would get us

00:04:21.280 --> 00:04:29.759
usually people would pair such a card with an i5 class chip but

00:04:26.639 --> 00:04:32.479
we're focusing on vr here and even with

00:04:29.759 --> 00:04:40.080
all the trickery it is still mostly GPU bound and check this out so it turns out

00:04:36.000 --> 00:04:43.040
that i3 plus 1060 combo here

00:04:40.080 --> 00:04:49.199
actually looks good to go with a super rating in vr mark orange room and a full

00:04:46.400 --> 00:04:54.320
bar in super positions optimum preset so that puts our total system cost at just

00:04:52.160 --> 00:04:58.720
over 900 us dollars and bear in mind that that is including the Windows 10

00:04:56.720 --> 00:05:03.840
home operating system which microsoft recently jacked up to 140 bucks so if

00:05:01.520 --> 00:05:07.039
you already have Windows or whatever then

00:05:04.720 --> 00:05:12.080
well you can subtract that of course though those are synthetics so

00:05:09.520 --> 00:05:17.520
the next step is to go lab rat mode and put this to the test now

00:05:14.080 --> 00:05:19.199
normally i game in vr at home on a gtx

00:05:17.520 --> 00:05:23.520
980 ti so i will be able to tell the difference if

00:05:21.280 --> 00:05:27.680
this doesn't manage to stay as smooth so let's fire up the oculus performance

00:05:25.360 --> 00:05:32.080
profiler analysis tool and head up display which should help us pinpoint

00:05:29.759 --> 00:05:35.520
the cause of any frame rate drops that we experience

00:05:33.680 --> 00:05:41.360
oh cool okay so this is oculus's redesigned home

00:05:39.680 --> 00:05:45.520
interface thing now you've got this control panel here that you can move to

00:05:43.600 --> 00:05:49.280
wherever it is that you want it to go and then you can uh

00:05:47.520 --> 00:05:53.520
go ahead and decide well you know which menu do you want to see oh

00:05:51.120 --> 00:05:58.560
close them so there's my library there's my explore tab here then you can kind of

00:05:56.320 --> 00:06:02.160
adjust this one over here oh neat you can interact with the

00:05:59.759 --> 00:06:05.600
desktop as well oh that's trippy let's go ahead and look at the other

00:06:03.759 --> 00:06:09.520
monitor the one that's not capturing my experience right now ah yes

00:06:07.520 --> 00:06:12.639
hey you can mess with ivan hey ivan i am opening your start menu

00:06:11.280 --> 00:06:16.560
what are you gonna do about it i can touch you inappropriately

00:06:14.880 --> 00:06:22.000
oh wait i probably shouldn't open fraps oh that just glitched it out

00:06:20.560 --> 00:06:25.280
i don't know i don't know what this thing is yeah thank you

00:06:24.160 --> 00:06:30.080
let's go are your monitors different sizes yep

00:06:27.520 --> 00:06:34.560
one is 4k one is 1080. dang it ivan game now right

00:06:31.600 --> 00:06:37.280
uh walk around the place somewhat because see they've made it social so

00:06:36.000 --> 00:06:44.319
you can invite people to come over to your virtual home friends currently zero friends online

00:06:42.880 --> 00:06:48.800
thanks that's why you wanted me to open that isn't it all right what's

00:06:46.840 --> 00:06:53.039
next i play some robot recall you've played this game before haven't you i've

00:06:50.560 --> 00:06:56.560
played vr shooters dang it

00:06:54.560 --> 00:06:59.919
yeah it's really fun all right well i think that's good i

00:06:58.160 --> 00:07:04.400
think that's all the evidence i need this is working awesome

00:07:01.680 --> 00:07:08.400
so for our last trick we're trying out the climb

00:07:07.120 --> 00:07:13.360
kind of the the crisis of vr if you

00:07:11.680 --> 00:07:18.000
think about it that way like the graphics in this game are

00:07:15.680 --> 00:07:23.039
especially when it was first released are pretty incredible compared to

00:07:21.280 --> 00:07:25.919
what else there is now what i just

00:07:24.639 --> 00:07:30.400
all right oh that's a long way to fall

00:07:28.880 --> 00:07:34.960
huh this is maximum settings too

00:07:33.120 --> 00:07:41.280
yeah go go go go look at me i'm like superman

00:07:38.960 --> 00:07:46.400
superman oh wait oh oh no superman falls sometimes

00:07:44.639 --> 00:07:50.479
okay so actually that went surprisingly well but

00:07:48.880 --> 00:07:54.800
still the point remains that in a massive surprise to no one just because

00:07:52.479 --> 00:08:00.560
a benchmark gives you a good score and you can run many vr games well on a

00:07:57.840 --> 00:08:05.039
relatively inexpensive budget does not automatically mean that you can run any

00:08:02.800 --> 00:08:08.160
game maxed out in vr it's i mean same idea as in

00:08:06.479 --> 00:08:12.720
flat mode right so just like flat mode if you do have

00:08:11.039 --> 00:08:16.400
some more budget and you're buying a vr gaming rig it wouldn't be a terrible

00:08:14.879 --> 00:08:21.919
idea to bump your specs a little if you want to play uh vr versions of flat

00:08:18.879 --> 00:08:23.919
games like fallout 4 and skyrim since vr

00:08:21.919 --> 00:08:28.479
is more demanding but frankly i've actually found that the

00:08:25.520 --> 00:08:35.519
most fun that i've had is in games that were designed for vr and what's cool is

00:08:32.159 --> 00:08:37.919
that the climb in my opinion is more of

00:08:35.519 --> 00:08:41.839
a tech demo than an actual game and that one still ran decently

00:08:39.839 --> 00:08:46.880
and most of the vr games out there that i've tried some of which are good and

00:08:44.159 --> 00:08:50.880
free like robo recall are not nearly as demanding because it's in the developers

00:08:48.640 --> 00:08:55.920
best interest to reach as many customers as possible with their games so our pc

00:08:54.080 --> 00:08:59.040
right here and of course we're not factoring in some of the RGB nonsense

00:08:57.760 --> 00:09:05.680
and stuff like that in our in our cosport but our pc here can enable a ton

00:09:02.720 --> 00:09:09.040
of awesome vr experiences exactly the way it is

00:09:07.600 --> 00:09:12.399
so thanks to you guys for watching thanks to oculus for sponsoring this

00:09:10.880 --> 00:09:16.080
video if you guys dislike this video you can hit that button but if you liked it hit like get subscribed or maybe

00:09:14.959 --> 00:09:19.920
consider checking out where to buy the stuff we featured at the link in the

00:09:18.080 --> 00:09:24.320
video description also down there is our merch store which has cool shirts like

00:09:21.600 --> 00:09:30.240
uh is anyone wearing an lt t-shirt not no one where's your team spirit um

00:09:27.680 --> 00:09:35.040
like the lt t-shirts that we have and also our community forum which you

00:09:31.680 --> 00:09:35.040
should totally join