WEBVTT

00:00:00.040 --> 00:00:07.080
a VR Gaming GPU for $200 it sounds too good to be true

00:00:04.960 --> 00:00:12.240
doesn't it I mean that's really cheap for a performance metric that has been

00:00:09.280 --> 00:00:17.080
held in this nigh impossible space for so long this car being able to stand up

00:00:14.679 --> 00:00:21.720
to this feet would be wonderful and could be a great Boon for the market

00:00:19.359 --> 00:00:26.930
adoption as a whole of VR so the question is can it perform in VR and if

00:00:24.640 --> 00:00:28.960
so how

00:00:28.960 --> 00:00:31.960
well

00:00:35.120 --> 00:00:41.760
the vitess mouse from Phoenix features a lightweight design and a naago 3310

00:00:40.160 --> 00:00:45.960
Optical sensor check it out to get a chance to win one at the link in the

00:00:43.680 --> 00:00:50.920
video description before we get into how the rx48 itself feels let's talk a bit

00:00:48.600 --> 00:00:55.800
about why VR benchmarking is so difficult unlike traditional 3D games VR

00:00:53.320 --> 00:01:00.800
games have to live within a specific set of parameters to avoid causing serious

00:00:58.280 --> 00:01:06.159
issues such as motion sickness for the user to accomplish this VR applications

00:01:03.840 --> 00:01:12.200
always have vertical refresh sync or vsync on and the current headsets are

00:01:09.360 --> 00:01:15.320
capped at 90 Hertz meaning that results of higher frame rates shown by

00:01:13.799 --> 00:01:22.000
traditional benchmarking tools like fraps aren't always necessarily

00:01:18.920 --> 00:01:24.479
representative of an increased user

00:01:22.000 --> 00:01:29.960
experience as you may or may not know most in-game settings for VR experiences

00:01:27.079 --> 00:01:35.880
are variable to accommodate for the necessary 90 HZ frame rate and avoid

00:01:32.880 --> 00:01:38.360
issues like motion sickness this is why

00:01:35.880 --> 00:01:41.880
we can't provide traditional Benchmark results like 72 frames per second on

00:01:40.880 --> 00:01:47.560
ultra settings part of your run may be at

00:01:44.360 --> 00:01:50.520
Ultra but certain settings may also drop

00:01:47.560 --> 00:01:55.640
to high or even medium at some points in order to maintain your frame rate this

00:01:53.200 --> 00:01:59.880
is a good thing I'm not complaining but it does have a side effect of more

00:01:57.560 --> 00:02:03.880
complicated benchmarking procedures the there are some ideas about what we can

00:02:01.719 --> 00:02:09.840
use to test VR applications like motion to Photon latency frame times Mist

00:02:06.640 --> 00:02:12.160
frames CPU and GPU time used in each

00:02:09.840 --> 00:02:16.720
frame and specific components like response latency of specific

00:02:14.120 --> 00:02:21.360
head-mounted displays but current programs don't have the proper

00:02:18.720 --> 00:02:25.920
capabilities needed to measure these stats well Luke I know how you can bench

00:02:23.920 --> 00:02:30.280
VR score-based benchmarks why yes Keen viewer

00:02:28.239 --> 00:02:35.239
score-based benchmarks for VR do seem like the cat's freaking pajamas

00:02:32.920 --> 00:02:41.400
and they are coming future Mark's VR Mark which is based on directx11 and

00:02:38.160 --> 00:02:44.319
measures display latency and persistence

00:02:41.400 --> 00:02:50.000
via external hardware is apparently coming soon whatever soon means and

00:02:47.680 --> 00:02:56.280
basemark VR score which is being made in cooperation with krych supports both

00:02:52.560 --> 00:02:59.159
vx12 and dx11 and features multiple test

00:02:56.280 --> 00:03:03.599
types including interactive VR static VR and VR spatial audio it's sort of

00:03:02.040 --> 00:03:08.480
available but still needs to be validated by a large sample set from a

00:03:06.239 --> 00:03:13.280
number of reviewers and the community the inherent problem with synthetic

00:03:10.120 --> 00:03:16.120
benchmarks is they are not realworld

00:03:13.280 --> 00:03:21.080
performance metrics we need to see how relatable these results are to the

00:03:18.879 --> 00:03:24.879
actual experience of using various setups and make sure the scale actually

00:03:23.000 --> 00:03:28.720
works properly however some of the massive advantages to score based

00:03:26.599 --> 00:03:33.920
benchmarks like the ones listed above are that since they're automated they

00:03:31.200 --> 00:03:38.799
are very easily repeatable they may become widely used and cited thus

00:03:36.319 --> 00:03:43.120
becoming a great resource overall collectively for prospective VR headset

00:03:41.400 --> 00:03:48.200
buyers so if you're concerned about system compatibility you can actually

00:03:45.360 --> 00:03:53.239
see how your system May perform before buying a headset the only additional

00:03:50.599 --> 00:03:58.079
problem I see here is when new harder to- Run games come out hopefully gaming

00:03:55.560 --> 00:04:03.920
companies will be good about review copies and let reviewers see what

00:04:00.720 --> 00:04:06.280
Hardware people will need for various

00:04:03.920 --> 00:04:12.360
experience levels before the release of the game with all of that said how did

00:04:08.720 --> 00:04:15.319
the rx480 Faire actually quite well

00:04:12.360 --> 00:04:19.720
especially for a $200 card now I can't check myself how the settings change and

00:04:17.560 --> 00:04:25.960
stuff throughout the games but watching Jon play raw data it was fine he didn't

00:04:23.120 --> 00:04:30.639
notice any big problems and everything was okay now we have seen certain things

00:04:28.600 --> 00:04:35.280
from sites like PC perspective where they put cameras inside and you can see

00:04:32.479 --> 00:04:39.800
the frame timings and stuff isn't as good on cards that aren't as powerful so

00:04:38.000 --> 00:04:44.880
we will see how it fares once VR benchmarks come out but if you want to

00:04:41.720 --> 00:04:47.000
just be able to game in VR V rx480 can

00:04:44.880 --> 00:04:51.520
get you there man and for 200 bucks that's pretty sweet they passed this

00:04:49.360 --> 00:04:55.560
test now when AMD was launching the rx480 they had a little meeting with a

00:04:54.080 --> 00:04:59.400
bunch of the reviewers talking about how benchmarking VR wasn't really a thing

00:04:57.560 --> 00:05:03.520
yet and I hope to have a meeting with the NVIDIA relatively soon to talk to

00:05:01.800 --> 00:05:09.199
them about the situation soon so hopefully soon enough I'll be able to

00:05:05.479 --> 00:05:11.680
give Real Performance metric-based VR

00:05:09.199 --> 00:05:17.800
benchmarking result information but right now it's just a she did pretty

00:05:14.440 --> 00:05:18.840
good it's good enough style review so

00:05:17.800 --> 00:05:23.520
hopefully that's good enough for you guys hopefully you like this video you could press the like button if you did

00:05:21.520 --> 00:05:27.880
or dislike the video if you didn't get subscribe check out the uh Amazon store

00:05:26.400 --> 00:05:31.600
if you're interested in checking out an rx480 also check out Link in the

00:05:29.880 --> 00:05:35.639
description to buy a shirt that's cool and talk on the Forum about what you

00:05:33.400 --> 00:05:39.000
think about VR benchmarking check out this video if you want to see the actual

00:05:36.960 --> 00:05:42.919
review of the rx480 and I'll see you guys next time