#virtual-reality
1 messages ยท Page 24 of 1
vive is ยฃ220 more and i have no space to play moving around
Unless you really like dirt rally or elite I'd wait it out then. Seated controller based VR just isn't the same
At least wait and see if anything happens at connect in a week. I highly doubt it, but there could be a price drop or more likely some kind of bundle announced
either with the touch or with a load of games
lol @dry fjord
Vive fan I am guessing
Touch is coming out next month
so, no more xbox gamepad needed
It's hard to see that the touch will be as accurate as the vive
plus touch is a way, way better ergonomically that Vive's controllers
Crazy talk
as per vblanco, it's as precise
I don't know anyone who's spent extensive time with them both that doesn't feel that way. It's not an anti-oculus thing, but just as an overall package the vive has a (slight?) edge
I do like the ergo of the touch way better; not sure i agree with the tracker comment
well, next month as people get Touch put to real test we can see the feedback
sure, but some of us are very biased ๐
don't get me wrong. as i say the layout blows away the vive IMO
layout?
the button config
(I don't use any desktop VR, but pro-Vive vibe is felt here, where clearly it's not the best VR system, if you don't count room scale setup)
I dunno, I have a soft spot for controllers shaped like a hilt
it more closely matches what I really want to use it for
the pro-vive vibe comes from the fact that it's quite good and oculus is just mired in crap and bad experiences of the past
why is it "clearly not the best VR system"?
there's no "clearly it's not the best system", we haven't all had a chance to try both, but room scale is also a big deal
like as someone who's tried both you admitting you've never tried either and telling us how it is is a little funny. I've spent probably hundreds of hourse in both
ultimately to us we don't care, we'll develop for both HMDs
clearly not the best because design is ugly, SDE is more pronounced, more expensive, less good software for it, software for HMD is worse what what Oculus offers. So overall, to a degree it's not as nice from a consumer point of view.
I tried Gear VR (developing for it) and I can see some points being valid without me trying either
sure, someone might like Vive's controller more than Touch, someone has plenty of room to take advantage of room scale, and for b2b applications I'd go with Vive too.
but from consumer's perspective, Rift is what I will buy
the vive is shaped like it is so the controllers can be attached to other things
there's a reason behind it
maybe for R&D Vive is more fun
maybe borrow or rent one and spend a week with it
So design is subjective, but i agree i like the feel/look of the oculus more. SDE is worse on the vive, but godrays are way worse on the oculus. It's a tradeoff. Which one i like more is actually pretty game/app dependent. More expensive only because it has the controllers, oculus touch + rift will be around the same. I disagree it has better software. I like oculus home more, but i think the games on the vive are better by far atm.
form a valid opinion
Manus VR is the only other thing that I know of where you can attach controllers to. However, who on earth will spend another $250 for gloves?!
I was thinking more like assault rifle setups
or mounting to gimbals for virtual joysticks
one thing I'll say is having using room scale for a couple of weeks, it's changed what I'd do in a mobile VR game. it makes you think about VR differently.
even if a mobile device doesn't have room scale, you start to become more familiar with what really works on other platforms too
Oculus has Carmack who, I am sure with the team of other geniuses working at Oculus, created async timewarp. Vive has nothing like that (based on vblanco's experience he shared here). Ultimately you can squeeze more visual goodness with Rift than with Vive and keep fps steady.
why not use proper joysticks?!
x1000 better to use real deal that fool around with Vive's controllers
stop being influenced by what other people think and try one yourself
and by real deal I mean steering wheels and HOTAS
I wasn't talking about replacing a hotas
use your imagination
you can do so much with them
I'm in danger of breaching an NDA. I'll shut up now.
I do think for myself.. Rift was my initial choice. And with all the folks writing reviews and devs talking about, my belief that Rift is a superior platform in a long run was reinforced.
Again, for b2b I'd use Vive. But for games and consumer apps - Rift.
business to business?
yeah
what in the fuck
for archviz, research, edu I'd use Vive
that's not b2b
well, it is
no it's not
it's not end-user business when you offer services to other businesses
so, just go try both for an extended period of time...
as @dry fjord says don't base your opinions on others, just go try them both for a good while
I'm told i should get ยฃ200 for my turbo chameleon, not bad
that's a cute device
what's turbo chameleon ?
its an fpga cartridge designed to go in a c64, but if you run it alone it becomes a souped up c64, an amiga and lots of other old hardware
I have a Vive and a Rift at home. My short version is that the Rift is more comfortable to use as a consumer and that the Vive is more important to have as a developer.
I use the Rift more but if I could only keep one it would still be the Vive.
how is everyone handling player 'death' scenarios in VR? what techniques and approaches are you using? fading screens? pinning a static hud or image?
I go to their house with a black sack and when they die I yank off the headset and bag them
it's much more immersive
it makes up for the vive having too much backlight
heh. like will ferrell in old school?
hehe
its funny i thought about your question and realize few of the 'games' i play in vr result in dying
yea, made me think
i guess hover junkers it fades, i think eve it might fade or 'crack glass'
most of them let you continue to move your head, but teleport goes off and "YOU HAVE DIED" appears on the screen
our thing is an app not a game so it doesn't have a death state
I recommend taking control of the camera and spinning it 90 degrees and slamming it on the ground for highest immersion. Then having it slowly float into the air like a ghost. Trust me players will luv it.
put an animation of vomit on their screen too as encouragement for what to do next
hehe
oh, in all seriousness.. i'd remove all saturation in post process and tint it a slight blueish gray
and then fade to black
maybe make any elements of their body go transparent
grieving relatives should run at you wailing
or worse
nobody cares
you die and nobody cares at all
chromatic aberration until the user takes off the HMD in disgust
@abstract gale I add a slowdown in the game until timing reachs 0, then i froze everyone, and do a fade to black
ive also tried a ghost
wich is very common
important, the slowdown/pause does NOT freeze either the controllers or the head
that on my new game, but im going to use GodMode for the demo
but after it ends, it gives you a score where it subtracts if you got hit
my VRMultigames game uses pausing
just pausing
so, had my first demo of my game on a conference this weekend
didnt expect it to be:
- this much awesome (regarding the responses)
- this much destroying standing around and just talking stuff (omg, sleeeeeep)
Hope you wore good shooooes
well, the shoes were not the issue
i still fill like dead
but at least i smell better now
Where were you demoing? EGX?
pixel vienna
haha, my game ๐ VR Archery stuff
Does it have a name so I can mercilessly google it?
Or is it actually VR Archery Stuff - because if thats the case, thats the best name.
that would've been a cool name actually ๐
though, you wont find much right now about it
its called "Franz"
and its about 30% finished now
aww
(old version, without hands and less quality, etc: https://youtu.be/5DTfHpKPxho)
(and also with the "codename" ๐ )
i have this one released for free
there is also that other one with unity, wich is also free and really high quality
there are too many bow games, and very good ones are free. If your game is just standing there shooting, there are free alternatives that are better than your game, so you have no chance at all
so, actually my game features 10 levels, magic spells with gestures, teleporting around in later levels with enemies running directly at you and a background story to everything
and if you mean "QuiVR" with the "high quality" unity game; Nope, thats no real competitor
and, i am yet to see a bow shooting game for VR with realistic graphics and not that stylized/comic stuff
there is my game, no bow though, but it has an arena with free movement and spells and guns
the physics are a bit....low gravity ๐
and honestly, if i would've taken all the "recommendations" that i have no chance with my type of game, i wouldnt have some free oculus+touch and wouldnt be in discussion about some nice funding right now.
i got my oculus + touch after i got a vive and published a vive game
i tried before but got ignored
so im kind of late to the party, i only had vr stuffs since july
i sent them a video and a link to a demo of my game. 2 days later they wanted the address to ship 2 rifts+touch to
i should really find funding for this game
im currently funding it from my pocket, but i just dont know how to find funding at all
but as said. i have gesture recognition spellcasting stuff already in, just not demoing it yet, since for a 5 minute expierence on a conference, thats just too much for the people there
and until now i funded it from my own pocket too (doing it since may)
my spellcasting doesnt use gesture recognition, but you need to do the motions of the spell
its with the buttons or trigger
but you have to toss the fireball, or block the bullets with the shield, or aim with the ray
i did prototype gesture based, but i didnt find it good on my game wich is fast paced
well, think about doing a circle with your hand on the ground and then "slam" the controller from top to bottom -> firewave with pushback for enemies
well, for fast-paced stuff it can get difficult; which is the reason why i have only very simple gestures
but if you want, i could probably help you finding some ways for funding
since i have at least a bit of experience with that right now ๐
a good point its that im on playstation awards, as finalist
and i do have a published game with 25.000 installs
with 97% good reviews
thats a good start, since then you have a more trusted "relationship" already, since you delivered in the past
my paid steam game, deathwave, has 800dollas in sales right now
early access, made in 3 months with 0 budget
and no marketing at all
guess how much marketing i did ^^
i did it for learning, anyway
i started mine for fun. like 1 1/2 months ago it started getting more serious
how are you finding funding, btw?
its a multitude of nice and surprising coincidences actually
it helped really much to demo it a the pixel vienna conference
im doing that in 10 days
i even had a guy from activision playing it and having much fun
i hope to find a publisher or something, im making my game a "demo mode" of 10 mins gameplay highly polished
if you want a simple hint: look at #introductions ๐
sent them stuff
not all
but a lot of it is
its a mixed architecture
i use c++ for the "core" stuff
but then each enemy has their habilities and AI on blueprints
similar with weapons
its a good workflow for speedy development
for a truly big development, go 90% c++
blueprints are problematic to mantain
I assume the bow is sideways for aesthetics? looks better than being held vertically
its not, im just handling it that way
the bow is attached vertically to the Vive hand
but it is inmense
its a longbow
i think its like a meter and something high
around meter and half
in horizontal is easier to strike enemies with it
i had that idea from one of the earlyest testers
"hey, wich such a big bow, might be cool to hit enemies with it"
so i tested it, found it hilarious, and left it in the game
the firs version of the bow game was much, much worse
it was static
enemies just walked to you
i then added melee and movement
i then added enemies with guns, flying drones, and time slow when teleporting
is the scenery all your gfx or using some epic assets?
I will be testing a rift on wednesday
from the talk in here i expect it to be a lot better than gearvr
it's definitely better than GearVR
though I still find seated, controller experiences to be a bit shit
standing and with motion controllers is the way forwards
standing is ok, moving about isn't ... no room
testing nvidia 4.13 multires
doesnt work well with forward
that on my VERY high density map wich has nearly 2 million polygons and a thousand drawcalls
the basepass seems to take too much due to that high amount of polygons and drawcalls
so deferred runs way faster there
also becouse multires deferred runs real good
maybe the basepass doesnt do multires?
might be that
nods
whats multires
yeah theres a unity daydream vr beta, only seems to work if you have a nexus 6p though
it laughed at my nexus7 tablet and refused to do anything
Daydream seems kind of daft
you need one phone to use as the screen, and another one to use as a controller
however, a phone has no haptic support whatsoever, so it's a shit controller
it only supports Nexus 6p
there's not much point trying to develop for it though, because like I said, there's no controller yet, you have to use a second mobile phone
which has no proper IMUs, so essentially it's new Cardboard SDK
and yeah, second phone for controller sucks - I only have 1 phone
no, because there is no Daydream hardware available
and I am not willing to get sick using Cardboard
plus, SDK will be partially in 4.13.1 and fully in 4.14
hmmm cardboard doesn't bother me
other than it looksl ike shit cos I'm using a nexus7 tablet
it sucks that there aren't whole a lot of Gear VR devs and none of the existing ones is on Unreal Slackers :/
that might be because nobody sees much of a future for Gear VR games
I doubt that
people making money on Gear VR market already
except that they all use Unity
people may be making money, but they're sure not making good games
I am not making a game for Gear VR either
as a matter of fact, what sells best is experiences
and I find myself looking for good experiences more than for games in VR
the problem is Gear VR just doesn't suit games
what makes you say that ?
gearvr looks far too locked down, daydreamvr has more scope long term
anyway I shall see proper vr on wednesday
not fogged up gearvr
locked or not, it has nothing to do with platform being suitable for a particular type of content
it has very limited options for meaningfully interacting with content
I mean, basically all you can do is look at things and tap a button
or use gamepad as with Rift
except that people don't own mobile phone compatible game pads, so you can't rely on that
or a clever game design can still use touchpad and be fun (but no one have implemented it yet) ๐
it's like making PC games that require trackballs; largely pointless
there are PC games that can do with mouse/kb, but require HOTAS or steering wheel for proper gameplay
yeah, niche games, and you can usually still play them just fine with a gamepad
but those are established genres, with a hardcore following
and they're the kinds of games you spend a considerable amount of time in
I can assure you that games with touchpad only controls can be made fun
if people don't make those it doesn't mean it's not possible
plus we shall see what Oculus unveils at OC3 - maybe controller, maybe positional tracking, maybe both
I always thoguht GearVR was perfect for getting the mobile VR porn market on the go. "Just gonna pop into this toilet and go have a samsung, brb".
VR porn is meh
at least the way they film it and the way they prep actresses/actors
I did browse some ( for educational purposes you understand ) and I thought "ugh they just made a 360 video from POV" and thought they could do much more with interactivity.
Imagine being the guy/gal that has to program genital physics for VR Porn games...
ha
I patented some gloves for VR porn. They cum with Foreskin feedback ๐
lol
@pallid echo I think the VR dev that worked on this has some experience in that area: https://www.youtube.com/watch?v=g6frKRN2HzI
Anyone here doing cross PSVR/PC dev? We're getting lots of issues with SDK compatibility. For example, leaving on the Move Controller and Morpheus plugins will make motion controller tracking broken in editor on PC.
Another issue we have come across while using VR preview is odd lighting. Ambient cubemap lighting specifically seems to drop out occasionally, resulting in dark black shadows for us. Occasionally VR preview results in an odd green effective which masks a lot of the screen, though the scene is still discernible.
if anyone here is interested in positional tracking for mobile VR should check out these guys https://twitter.com/VRTracker
The latest Tweets from VR Tracker (@VrTracker). VR Tracker is an Open Source 3D position tracker to improve Virtual Reality experiences
why is it that if you parent a component to a motion controller, it moves with the controller with zero-latency. but if instead the component is parented to the pawn, and its transform is updated to match the motion controller in a tick() function, there is quite a bit of latency between controller and component
I donno but I get that too, I guess of course
the wobble of things attached in that way
annoying
yes. makes me think tick isn't called every frame.. just close to every frame
In some cases it may make sense to use a Vinterp to smooth out the movement. You could also grab the tracker data manually and update the main motion controller in the same way so both objects match and are smoothed, though late compared to the native tracking info
also, anyone try out forward rendering recently in UE4 on VR?
@frosty geode do you know what's different between attaching vs. updating transform in tick? is the engine-level code taking some shortcuts for attach that aren't possible in game-level code?
I'm not sure what's happening lower level unfortunately. I found a thread that mentions GatherLateUpdateComponents though
Hello everyone,
So I've been running into trouble with the "Low Latency Update" functionality of the motion controllers in VR. Obviously I want my controllers to have low latency updates because there is a very noticeable controller lag introduced without it (this is especially unacceptable when asking players to do anything with precision, like aim a gun). However, as the Low Latency Update option states "render transforms within the motion controller hierarchy will be updated a second time
seems like a good starting point cheers
@frosty geode the oculus forward renderer is working great for me in 4.11. however during vr preview traces do not work - that is the only nuisance
sorry i mean to say the traces work, but they do not render properly
I'll download Oculus' 4.12 branch and give it a shot. Are there still issues with translucency?
why not 4.13?
is there a 4.12 or 4.13 branch? where are you finding it?
anyone have print string tips so you can see it when VR headset is on?
even on the PC view it prints too high so unless its printing a lot of stuff you cannot see it
@normal thorn by default it will also log those prints to the console which you can bring up with the ` key
and those are pretty visible if you're just occasionally printing
also you could detatch the output window from the editor
oh wait nm that wont help you
I don't see a 4.13-ofr branch, I'm assuming ofr stands for Oculus Forward Renderer. There are 4.11-ofr and 4.12-ofr branches available.
@frosty geode I think you'd be better off going with the new forward shading features in promoted or master.. i've used both of those oculus versions and find the new forward rendering in Epic's source less buggy. For example spot lights don't work at all from what i've tried with the OFR
thanks @fleet veldt I know it says prints to log but I never found that log.
@frosty geode disabling low latency update on the motion controller made it so the controller has the same (high) latency as everything else... so it's definitely that ๐
now going to see if I can get my component to use low latency update without being attached to the motion controller
no I still never see the print sting log. Message logs tab?
This is rendered with the UE4 forward shading in 4.14 promoted. https://youtu.be/kajzgd014ZM
nice
Dave, you could add a text render component to your motion controller and update that, instead of printing.
yeah thanks, I found the log in the output log but I swear sometimes its not there. Anyway. just bashing my head against a wall, you know. VR development
tmek that looks badass
@fleet veldt how big is the team making that?
after last weeks hires we ahve a total of.. one.
you mean its all you?
hah wow!
still, looks great
I'm actually a fan of re-used marketplace assets. makes everything feel more real when you see stuff you recognise. like each game could be set in the same universe.
and thanks for encouraging feedback!
@fleet veldt cinematography reminds me of demoscene. sick.
for sure ๐ let me know what your impressions are! ghostvector.com/builds/
it was built with oculus so it only has very limited vive support. but it does work. you use the triggers to teleport and interact
I wish the whole vr hardware thing was more transparent on UE
any of you guys ever make a look at trigger? Specifically I'm trying to make a look away, look back player pawn flashes to help you find it.
heh, it's funny we both have sunglasses on in our profile pics lol
right now I made a frustrum geo w/ collision then made a overlap event
that's because sunglasses are cool
that should work but when I parent to camera overlap events constantly trigger
when I set location/rotation to camer no event trigger
@normal thorn it sounds like you're talking about taking control of the camera orientation away from the player.. that's generally considered bad in vr because it can make people uncomfortable/sick.
nothing like that
when player looks away from their 4rd person plane I want to help them find it when they look back
its moving so sometimes for some people its hard to find
you mean an indicator?
I want to highlight the plane after people stop looking at it and return to find it
frustrum geo overlap event should, in my mind, work
but parented to the camera it fires continuously
while looking
do once, and reset
plane as in two dimensions of axis?
but the end overlap also fire continuously
or planes as in reeeeow plane?
airplane
gotcha
rc airplane
you're gonna need to draw a picture
might just have to to a sphere trace or something,
@normal thorn I see. sometimes you can just stick a big collider in front of the camera.. for example a big long capsule to detect when thinks are right in front of it. That's how epic had the moving security cameras detect when the player walked in front of them in the BlueprintContent sample.
but yeah if you wanted to make sure you covered the entire view you'd have to check a frustum volume
Top down Scifi its one of the absolute best packs
its also all on 1 texture, so you save a lot of memory and drawcalls
everything is 1 master material with instances, that uses the same texture
its main problem is that it is very low poly with a lot of detail in the normal maps, so there is stuff that in VR, looks completely flat
like the doors or the pipes, wich are a literal plane with the normal maps
MAKE VR GREAT AGAIN
makes sense
they dont want a swarm of bad VR at the start
becouse literally every amateur vr is probably going to try to put shovelware there
yeah
Crossbows in VR are the bomb: http://imgur.com/gallery/e6JPaGV
Second Crossbow in the game, I could play with this mamajama all day long. Doesn't pack quite as much of a punch but it's small and nimble. More info at www.playyore.com
@pearl grove lookin good ๐
@pearl grove Looks awesome
Man the last 5% of shipping a VR game is absolutely killer
we have a few hitches in the framerate and we can't for the life of us figure out whats causing it ๐ฆ
might ship at r.ScreenPercentage 130 to overcome it in the short term ๐ฆ
@spring pond Were you at 150?
140 (default for vive)
with 130 we hit 90fps almost everywhere on a 970, but there are some weird hitches
try nvidia multires
4.13?
Already on 4.13
there is multires for 4.13, but i tried it and it instacrashed
guess its too early
it also increased the time in forward renderer
the depth prepass takes quite a lot
i guess its becouse its doing the depth prepass WITHOUT multires
3 reasons I can't use multires: i didn't find the builds as stable (minor), Doesn't work on AMD (major), doesn't work with post process blendables (dealbreaker)
@spring pond I only have oculus but r.screenpercentage doesn't seem to affect the Rift, only HMD PD 1.3 .. how do you handle both ? or are you only supporing vive currently?
only supporting the vive for this launch; we will update for touch launch soon
(it works fine with the touch and oculus, but we haven't changed tutorial gfx for touch and fixed some haptics and button mappings)
what is your game's website/youtube I'd like to take a look since you're so close to shipping ๐
https://www.youtube.com/watch?v=oxvDI5aDx7A Ive been doing this today
a bossfight for my game
still balancing it becouse the boss is taken from a desktop game
@spring pond.. you're not marketing your game before release? o.O
nope, its free
I see
@spring pond Free games should be marketed too.
mine wasnt, so it got people, but not that many
i mean it will be marketed, but we're fine with just dropping it and letting it grow organically
(If you want people to find it and play it)
@spring pond Take it from me, that's a myth.
Organic growth is extremely rare, you should definitely invest time into marketing if you plan to have people playing it.
it has to be the most memeful thing or the most fantastic thing to happen
so people tell other people
marketing is by far the most important part of gamedev
we will, but we are fine with the marketing coming right at release time, its also more compilcated than that, but i can't get into it
you can do a shitty game, market it well, and get your money back
look at those big mobile f2p games
All right, well, just my two cents.
if you do a good game but not good marketing, no one ever knows it exists
unless its literally the best thing ever made
I agree with you all, but again its slightly complicated
@wicked oak that looks good.. i'm sure you get asked this a lot but is that the only locomotion method won't that make a large protion of playerbase get sick?
I'll be sure to post a trailer/steam link when it goes up
ill have blink teleport
and also one where you press the button and aim, and you move in that direction
it definetly looks exciting visually
if you aim UP, you get the ark to teleport to higher places
right now rotation is on 90 degree intervals
you press the grab button in one hand to teleport, and if you press the grab button in the other, you rotate
its surprisingly intuitive once you get it
good for vive wich has limited buttons
as im using touch, ill just add it on the joystick too
and joystick movement
it really gives me a similar vibe I felt from when games like quake3 first came out (which is a good thing)
im reading the pre-demo build
i demo this the day 6
im going to try putting it on PSVR, thats why im not adding joystick movement
the psvr doesnt have a joystick, it has 5 face buttons, a trigger, and select/start buttons
wich is honestly a lot of buttons
@wicked oak is there anything tricky to posting videos on twitter (recomended bitrate etc?)
its really really tricky
normally it has to be 30 fps, youtube quality
with that you should be fine
but its picky
its REALLY picky
in youtube you upload whatever and it accepts it
not twitter
and it will lose your time and then complain with an odd error
@wicked oak what assets are you using in that prototype project?
custom assets and infinity blade pack
@wicked oak Just so you know the face buttons on the PSVR controllers (Move) are pretty small and not the most comfortable things to use
i know
i plan to use them for extra stuff
main control is trigger and central button
Yeah, totally, it just took me by surprise to see how small the buttons actually are. The trigger and Move button feel pretty great though
also remember that PSVR is front facing only so make sure you have some sort of orientation control as well
ah, nice
Anyone have any pointers when incorporating sound in VR? Any pro tips?
mono with positional ๐
Sounds just played directly to the headphones sound strange compared to ones that aren't
How about ambient/BGM ?
Don't really have a ton of that in our thing
we do play a fan noise right to the headphones
Haven't had any issues with playing music on mine.
But if it feels weird, maybe you could put an audio component on the player's pawn
Above it, in the sky, that blasts music
Could feel more natural for... "music from the heavens'" or what have you
At least then it has a position. @granite jacinth
Definitely one of those new things in VR that need some experimenting from devs to find the best practices
Cool, thanks for the tips @still marsh
@granite jacinth Just to be clear I have not tried those myself. Just ideas. ๐
Speaking of audio, has anyone had any success with using the different sound modes on PSVR? By default it puts this strange short reverb on everything that presumably is supposed to help with positional, but it sounds awful on music and ambient FX.
I saw that there was a method to turn it off in the SDK but I'm not sure
oh wow, the repo I was about to link wasn't official one
but from google
I'd like to see price of HMD and price of controller, and if all that works with Galaxy S6 (when it gets Android 7 update).
Any idea why my swarm coodinator PC wont show the right IP address https://infusestudio.gyazo.com/32901c5b09cf86f37888a5cb2149734d I have tried to fix this for a couple of days now
Samsung haven't mentioned any daydream compatability
infuseStudio: that's the pc its running on, the other ip could be any pc on the intranet...far as I know
it shouldnt be 127.0.0.1
thats not the right IP address
And there isnt a way to change that information in swarm coordinator
Do you guys use multiple computers via swarm to bake lighting?
I only have 1 pc
sounds like tango
I like that they did a 3d model of van gogh's bedroom
maybe I want a massive cord hanging out of the back of my head because matrix
also I'd be incredibly cautious of objects it CAN'T see
or things that may confuse it, like wall mirrors or things that are reflective in IR
or simply getting messed up from sunlight filtering through the curtains
my ears will prick up when I hear "IR, visible and UV image tracking from six sources"
test it on lots of different headphones!!
Btw, I wonder if Daydream would allow usage of 2 controllers simultaneously
you know what would be cool for vr? if the vr had an enviroment scanner and would make a cartoon version of where ever you go, like if you walk around your house it scans and saves it as the level info that way you can interact with the real objects in your house but you can add npc's and fight them etc. that way you can really walk in your house but fight zombies or what not on the theme of game play you want. and not have to stand still
so.... hololens?
yea like a hololens
but say if you pick up a cup it always scans the area that way you can make a house for gamers to go in and have rubber cups/ plates and actually feel the objects in their hands and throw them or use them as wepos etc
๐
REAL OBJECTS used in game. like have a chip in the cup that tells the vr what its weight should be and the damage it can do etc to make interactable inviroments!
just a thought ๐
who wouldnt love going into a lazertag type arena and fight off zombies using real world objects?
there actually is something like that, it's not done yet though I think
I'll look it up, it was a while ago... they basically made a warehouse scale VR playground with wooden walls and all that
im mean think about it. we have motion controlers, motion capture devices, lazer scanning tech and wireless tech. it could be done if there was enough resources. maybe ill pull out my heath kit and start working on some type of wireless chip that can talk to a plugin. to send variables to unreal or something etc
https://thevoid.com/ I think it was this one
At THE VOID you will walk into new dimensions and experience worlds without limits. The Future of Entertainment exploring with Virtual Reality.
i have an old broken wireless camera i can play with its components since it sends data over a wireless router !
WOW impressive. this is exactly what i meant! but with objects that can be picked up off the ground etc
and so people can experience it in their own homes too not just in a big facility,
im looking for people with oculus touch to test my prototype
PM me if you are interested
Do people still buy apple phones? ๐
lol, wut?! $500 for sensor?
So if you combine that with a phone that the instructions say not to put next to your head, wireless earphones that add more EMF to your head, a sense that will most likely add more EMF to your head. You should be ripe for a good dose of cancer by 40-50. Well played Apple. Well played.
on top of that there is no rotational tracking, so you get horrible laggy Cardboard crap :/
sounds like if VR dies, it's the same reason Flash died. Bad devs making it look terrible :p
so, $500 for sensor, $700 for the phone (let's say it's what top Daydream phone would cost), $80-100 for HMD, $100 for controller. Almost the same as PC + Rift.
It is a developer kit for sure.
it's expensive for sure
Had a go on the PSVR at EGX
Easily the best one I've used so far
(not used a Vive though)
PSVR is okay, but the controllers are very limiting
you can't do room scale, so that kills it for me
room scale is the thing for VR, imho
lol, what if 90% of users do not have room for room scale VR ?!
Depends what you're doing I guess, I played Battlezone VR and it was perfect
(I don't)
I've used the Vive in a very small UK house with no problems
But its a game where you're meant to be sat down
room scale is as large as your room is
@balmy inlet most of VR things never make it out of dev kits stage
if your space is 3mx3m or less, you can still do room scale
my space is my chair and less than 1m x 1m if I move it out of the way
@mighty carbon sure. But you have to start somewhere.
I get a 6x6 room for Vive development next month. My only problem with that is, I need to keep switching rooms when I want to test things. I think I much prefer sitting on my arse with a DK2 :p
I used my kit in the kitchen / dining room because it was the biggest available space
doesn't matter where developers use it - it's for end-users and if they don't have room to walk around, experience might be as good as seated.
I'm talking about as an end user
if I want to develop for it, I have a pretty big office space
so, why bother?! PSVR provides positional tracking, so even if you stand/sit, it will help with immersion a lot
because it isn't the same
well, if you aren't walking around 15ft x 15ft room, it is exactly the same
no, it isn't
in 'flat' terms it's the difference between a side scroller and a first person shooter
mechanically speaking you have considerably more freedom
poor choice of analogy
is it though?
if I can stand and make 1 small step forward/back/left/right, bend/tild/squat with PSVR, it's exactly the same thing as if I was wearing Rift/Vive
no, it isn't
lol, yes it is
I'm beginning to think you haven't actually used a Vive
yeah its really not the same
why isn't it? due to latency / accuracy ?
so PSVR isn't 360 deg ?
nope
thats a killer for me
hmm..
you're mechanically constrained to looking forwards
So even worse than Rift
correct
may as well call it "360 video headset"
I see now
but this is fine for what PSVR is aiming for
the PSVR isnt half bad
it still has moderate tracking for the headset if you turn around 180 degrees
but its true that the controllers are its biggest limitation
becouse they are just normal PSMove
but it cannot track the controllers, because it only has one camera and they are just Move controllers
yeah, its a forward facing experience
yup
can still make some good games for it, but it isn't going to be the same experience
its a HUGE difference compared to the oculus with a xbox controller
my interest is that there is a load of money in that platform
no competing with steam vr shovelware, and way, WAAAAAY bigger userbase
that for sure
it does look like the user base will be bigger
but there's no guarantee it will persist
PSVR could be the next Kinect, heh
it's gonna be the biggest VR scam since Oculus priced their controllers :p
its the only VR system for the japanese
as the japanese have no gaming pcs
lots of strange japanese vr games on it for launch
Kiki Kiki Chapperone Simulator
I really dont get japanese games even though I love the culture ๐
psvr also has a high entry fee for users
compared to pc vr its less than half
ps pro and ยฃ400 for vr kit
but it runs better on PS4 Pro ๐
of course, on a pro it can run FAR better
yes right, within 6 months it will be ps pro or crippled joke vr
i speculate that it on a PSpro its probably going to be near GTX970 quality
i know it has instanced stereo stuff + "MultiRes" + more hardware/driver optimizations
and looking at how far they go on PC with the nvidia branch
they are no joke
instanced stereo halves the drawcalls, multires cuts pixel amount by half
I am gonna wait for PSVR to take off and then make an educated choice for a VR platform to play and develop for. I have a feeling that Vive and Rift will remain niche for quite some time
sure, but it's not just about the best platform. It's about costs too.
instanced stereo isn't a magic bullet and in many cases will make performance worse
i have tried all
rift could kill itself with the locked down attitude
Now, if Project Scorpio will support Rift or some special version of Rift, that will change everything
multires is great though
The rift will be ok once they bleed out some more tech for it. Vive is currently pretty damn awesome. OSVR is catching up slowly.
the ps4 has higher drawcall efficiency
early adopters are more likely to be nerdy types who want full access
@wintry escarp PSVR is even more locked, but somehow no one cares.. So no, Rift won't kill itself by being locked
they'll go to vive
not sure about that
PC isn't a console
the Rift + Touch is awesome
i have both oculus and vive
and i develop for them at the same time
so i switch beetween them constantly
me too
smilertoo is just being political
the software on the oculus is just so much better
vblanco: i don't know about that
like FOSS folks - ideology tops practicality
@wicked oak what makes you say the software is better? I agree it looks more polished from a consumer facing standpoint, but it has some wacky config shit
that too - I don't want to f#ck with configs. I just want to plug and play, like with Gear VR
I like the oculus timewarp more than the vive reprojection, i'd say that, i thought you meant oculus home and such
the main beef i have with vive, is that it only gives you 8.5 out of the 11 milliseconds you have
if you go over 8,5 it locks to 45 fps
meanwhile, oculus async timewarps it
wich reduces latency a ton
and it looks smooth even if you are lagging to 40 fps
if you are at 80 fps then its timewarping it anyway
so it still looks really smooth
vive just throws you to 45
The headset reprojection in the vive is pretty good at making the head motion look smooth, but the hand controllers look jittery as shit
they work better on oculus
becouse while its timewarping, its timewarping at far higher frequency than 45 fps
in my game, you can drop to 50-60 fps with a ton of explosions around, and poeple still will get tricked with it being smooth
normal gameplay its 90 fps
but becouse it is allways around 10 out of 11 avalible milliseconds
it goes right to 45 on vive
the game runs noticeably better on oculus
people have been porting timewarp to the Vive anyway, hehe
the vive will pump 45fps content to 90fps and reproject tho, its not like the framerate crawls down to 45fps in headset. You can still look around smooth, you just get horrid jitter on the controllers
I haven't actually paid much attention to how timewarped frames look with the touch. it jitters enough all by itself ๐
ah
yeah, we are trying really really hard to hit 90fps all the time on a 970 and are really close
with well written software i think a lot of people try and make this huge gulf between the two. When set up right both HMDs/controller sets are pretty fantastic
I find the oculus more polished feeling and the vive more reliable
@wicked oak are you saying you effectively have to hit what is 117 fps on the Rift to get a smooth 90fps on the vive?
the opposite
you have to hit 110 fps on a vive to get smooth 90 fps
oculus gives no craps, it runs at whatever your frametime is
and timewarps it
yeah that's what i meant (and thought i said lol)
hi
well that's disheartening as i've been developing on the Rift and thought it would be an easy transition once I got a vive
hi zenbauhaus
o/ @real needle
i was wondering if any of you encountered the same problem as me when packaging a vr project
What would that problem be?
you doing 4.13 forward render
?
there is a bug that wont let you package on 4.13 forward render
a damn shame
I was going to package a 4.14 fwd rendered version of our project under the "beta" tab of our steam app, but too busy making the production one work ^_^
Anyone have a definitive list of the render features supported / not supported with the forward render? I downloaded the latest promoted branch to experiment with and have been having a number of shader compile bugs. I did find out it lacks SSAO and cascaded shadow maps, making it pretty unsuited to fully dynamic games atm.
Actually for that matter, anyone else NOT using lightmass on a VR game? heh
maybe its targetted for 1080 platforms?
would you believe that we are hitting 60 on a PSVR? : D
yeah because I am imagining a screen that says "PSVR @ 60fps". You put that in my head though
60ps is just so lax for VR. Whoever thought that was a good idea should be employed at Apple
honestly it's what most devs are doing for PSVR - I don't know any UE projects targeting 90 or 120
oh its probably the limit, I wouldn't doubt. It's just when I hit 60fps in a rift or vive, I feel it. And i'm a dev thats been working with it for a year or so
the reprojection to 120 tech makes 60 not a total vomit comet on PSVR
fair enough. I haven't had chance to check out PSVR yet. I missed Gamescom by a whisker otherwise I would.
i would say it potentially provides an even smoother experience over OVR / vive due to running at 120, but the lack of detail and artifacts from reprojection bring it down a bit
but that's why I am looking into forward rendering
I have a feeling we'll be shipping with deferred on all platforms though, due to our dynamic requirement
hey john, how is the PSVR performance compared to a PC?
im going to develop for it, im talking with sony for it, already became a playstation partner and signed the agreement and stuff, just need to get the developer portal credentials
@hidden dust
well it's tough to say, do you have something running on a PS4 already?
I would like to do the same but I think we are exclusive on "the other one" for a non-VR version
Do I need to be incorporated and have office space to get ps4/PSVR devkit?
@mighty carbon it usually helps. I mean I tried in the past to get devkits as a solo self-employed dev and pretty much get laughed at
yes and no. they technically only require a static IP address
which you can get around...
:( so much for Sony + indies...
don't dismiss it too quickly. we technically don't have an office and we have 2 devkits and 4 psvrs haha
psst - http://strongvpn.com/
StrongVPN.com - Providing high speed, unlimited bandwidth, multiple countries VPN accounts for over 100,000 users. Since 1995
don't tell anyone ; )
Hmm.. interesting
Are Sony's devkits expensive? (For example XB1 devkit is ... basically XB1 console)
yes they are, but you can go through their indie program to get them for free
I thought they no longer do that
we met with them first at indiecade and got hooked up from there
this was 2 years ago so i dunno
Aye
and to be fair we had quite a bit of game already because we were first planning on just coming out on PC
I wish MS had clear road map for VR with Scorpio. MS seems to be friendlier and less conservative than Sony.
yeah, actually we have a few xb devkits but no current plans with them
US company, basically
motorsep they just renewed their indie distribution program so you might be in luck.
I will say for a super small team working with multiple platforms in general is a headache
ive given up hope of MS opening xb1 up like they promised
there will be no new xblig
with xblig I could just plug in my x360 and do what I wanted in xna
id@xbox needs their permission and unlock
Well, you had to sign up with Epic to get UE4 ;)
I think we have a 3 dynamic spotlights in our entire thing that only effect 2 actors
not in our game : D
hmmm seems to be a huge win10 update out
yeah
it raped my machine overnight
thanks windows, I had work open
it also reset the default browser to edge, took chrome off the taskbar and put edge on in first place
still on win 7 for deving
@dry fjord that happened to me. POS operating system. Luckily I didn't have work open because I have had this happen before
we're running enterprise
it's meant to keep its filthy paws off of us
I may organise a meeting with a local microsoft rep and rough them up
Now that's networking !
if this is your first windows update night, you have to fight
During a nighttime update, nobody can hear you scream.
it's as if there's someone at MS thinking up ways to annoy users
"How can we be evil today?"
think Microsoft mentioned they were adding linux into windows at some point. I think that's pretty damned evil.
odd really considering powershell is pretty much the same thing less the python support
or was it perl
i forget
now we can really bash windows
Ba dum tis
Lol
powershell is impossible to use though
it's so arcane
whereas most devs already know bash
yep, powershell is one more thing to learn. would prefer to just avoid it and use bash so its the same on the local as on the server
anyone done carpet in VR? I've tried POM (Too expensive), Bumpoffset (Looks crap) and now I've made a faked Shell and Fin technique, thought it would be cheaper, but turns out 25 stacked masked alpha planes hits performances again (Screenshot), thinking maybe tessalation might work. Any suggestions?
Ah fair, Thing with POM i found that even adjusting the amount of stack still didn't help. I'll try removing some planes
yeah POM isn't the fastest thing ever
with POM the smaller the surface the better
you could use your rug in a more discreet way
cut out parts that are occluded
yea but check this
carpet still in view
only tanks when looking straight down
suppose less is in view
also looks wose now with less planes...
that carpets a bit too shaggy too
use a finer heightmap
also, try enabling pixel depth offset
that would help
so grab the offset from the PMO and plug it intop that output
your things will look better when they sink into it
Wait, if you increase the heightmap resolution, will that increase the FPS? @dry fjord
Don't think so, I think Masked alphas are the issue
Gonna give tesslation a go, heck
Last time I did something similar I simply had a height map that was stark white dots on an all black background. Then I used that to multiply the colour of the grass and as a bumpmap. I only needed one tiling texture for that then.
noobs: nah, it's still jsut sampling a texture
I do pretty much the same thing here: https://www.unrealengine.com/marketplace/advanced-fur-shader
High fidelity fur and grass
except with more features
tesselation won't make much difference either
it's just an expensive pixel shader
oh
you CAN adjust the slices depending on the viewing angle
so if you look straight on you can show less slices
that'll make it a bit faster
the slices don't have to be consistent over the whole surface
there's an example floating around but it's fairly simple
You mean bitmapping it?
not sure
POM already reduces the slices based on view angle but you can make it more aggressive
That's what I did have in mind
Closer the camera less alpha
ah ok
as far as i know it'll draw all masked or translucent materials in z order
might do some stencil or tile optimizations, but that's it
So what is the cheapest POM
less slices ๐
Yea, but through which implementation?
manually removing planes didn't work that well
unless I need to have like 5
in the shell/fin you basically add more surfaces to draw that can't; be optimised by occlusion culling
in POM you add more steps to draw that can't be optimised by occlusion culling
it just moves it from one call to the other
with POM you can turn down the number of steps it makes
so it's slightly easier to tune
so in effect they both perform the same
and because it's per-pixel you can eliminate unnecessary steps if you know the camera angle
yeah pretty much, with POM being slightly better
with shell/fin you do less work cutting the edges of the object
Okay, so the node built into ue4 should be the best method
or is there a custom one that works better
it's just up to raw pixel pushing power now
that image I linked above, try to adapt their view stuff to yours
Thanks for the info
uise it to set the max step count
POM defaults to 8 minimum but from directly above you could get away with 3, which is the real built-in minimum
but you want it to be angle dependent
Hurrah! Performance is good ๐
anyone figured out a reliable way to start in windowed mode?
I thought i could change from ini and i tried the "SetRes 1280x720w" but it seems to need to be precisely timed
Soz, internet decided to cut out. In the end I settled again for POM, but this time stripped out a bunch of stuff. Previously I had much more complex stuff going on, this time realised I didn't need that
astonish: run a console command in your game mode?
mind you, I hate that method
res flipping back and forth
that's awesome @hard lagoon
Happy days ๐
try to tweak your settings to get down to under 8ms a frame
then you'll be perfect for VR
only 3ms to go
gosh, I wish
I might tweak some stuff
But I've got to migrate the entire project first, got packaging issues, then I'll go through and figure out what r.insertsettingname here to enable/disable
are you using the recommended VR quality settings?
someone linked me to it a few days ago, it helped a lot
the defaults are not so great
I've seen several recommendations, what are the ones you've gotten?
I've seen the showdown settings and the recommended ones on the docs
just the epic ones
Information over developing for VR.
yeah that's them I think
the showdown ones
it's such a shitty repetitive task that I might make a VR optimiser blueprint node to run commands
yea I've used those
I even went through each one
try MSAA and forward rendering too
oh nice
I know at Uni we had to kill the quality of our VR game haha
not sure if I'd tried that
just entered every command possible to get it to run in vr
I guess I have if its in that list
but that was over a year ago, I think vr rendering has improved ten fold
I wonder if we're still waiting on the nvidia multiscale rendering stuff
I've got one dynamic light
1.5 I should say
fortunately I can get away with it, since our game takes place in a room
cached shadows helped out thought, nice new feature for 4.13
Yea, we have it so you can hit the lamp, plus there are tons of moving pickups, and makes baking not necessary
ah gotcha
well the less stuff that it's rendering in the dynamic pass the better
definitely do bake what you can
get a baseline static lighting setup going and add dynamic lights from there
anyone know off hand the name returned for GetHMDDeviceName for the vive?
will look up one sec
i'm 99% sure it's "SteamVR"
well fudge, i just upgraded promoted and now my UE4 build won't start ๐
@urban shell if you make a VR template project and search blueprints for GetHMDDeviceName you'll see the names it campares for vive, oculus and ps4
ah, cool. i didn't know that, and actually have one of those projects set up. will check it out. thanks!
yep, SteamVR
https://www.youtube.com/watch?v=cOXGglrFu1w << finally got ISMs working. Even with ISMs and aggressive LODs performance is horrible on Gear VR ๐ฆ
Galaxy S6
Gotta try merging actors instead of using ISMs and if that doesn't help, I guess I'll have to simplify that "crypt" turning it into some teleport with a few columns or something like that
Found some issues with culling ISMs using cull distance volumes last night
I prefer for Epic just either drop Gear VR support, or make a statement about state of the support, as half of the sh#t that should work - doesn't. And if it will work for Daydream but not for Gear VR, then that will be pretty bad :/
oh yea, Dynamic lights, even if they don't cast shadows destroy performance in VR
Has anyone tried the mobile renderer for a desktop VR project?
with regular projects we see yyuugggeee increases in performance but I'm sitting thinking it's gotta be too good to be true for VR or else it would just be written somewhere as a standard practice in big bold letters
i would be interested on that
how would you do it?
so i can test it
becouse my game has no fancy graphic features, it would run on the limited phone mode
just tried here, no diff in performance
I've heard that on projects with complicated terrain on 4.13 it makes a diff, but just tested and I get the same results with mobile/desktop
well it was worth a try
Project settings - target hardware
oh
but that doesnt use the mobile renderer, i think
it just sets some basic features off
the other mode just enables everything
ah yea, now I have to figure out how to force the forward renderer
forward renderer is on the rendering settings
its a checkbox
on 4.13
that is the DESKTOP forward renderer, wich is a lot more complex than the mobile one
as i dont really have many lights and 99% of my geometry is static, its probable using simplified mobile style forward would run quite well
yea, I think so as well. I think any style could benefit, the only issue that might arise is lower quality AA
I think they only way to really test is to switch the target platform and make a build
mobile gearVR has msaa
Did that get put in in 4.13?
ah, cool.
by default, gearvr has used msaa
Are you runnin the Nvidia branch?
really? have you tried it?
yes
yea, true
oh yea? that's pretty snazzy
yea, it's all worth it for multi-res
we see 30-50% increases in fps with that on
shame it doesnt seem to work well in forward
well not yet I guess