PCVR/Link Development
Recently active
The OVRMainMenu script locks the mouse. This can give unexpected behavior. It took me a while to find out why my mouse wasn't working. Please give an option on this script to disable this feature.
http://imgur.com/Cu4Bz14http://imgur.com/0OjrU8pSince I ordered my DK2 in May I've been learning Unity engine development in my spare time. I'm a full time web programmer so I've had a head start with the scripting. I'm pretty much new to everything else involving 3D game development. My goal is to create DDR experience for the rift where you feel like you are actually hitting the arrows with your feet when you're playing on a mat, and flying through various scenes that fit the mood of the music weather it's space, a creepy swamp, tropical reef, etc.Right now it's very crude because I'm terrible at game development. But I'm hoping to spark interest in the development community.When I complete this project I should be able to convert any existing stepmania track to VR format. I've got the matching of arrows to the music down without having to recreate the original patterns. The existing game data works!It would be nice to have some 3D models for the arrows instead of flat planes with an
Hello all :)I would like to know if one of you guys would be interested to develop a VR Planetarium for educational purpose ? and of course share the benefite at 50/50 I allready got some customers interested in education Business, and they wait for a demo. I would like to develop it on my own but i have no time as i work on other part of my Project.i've got also Domain Name for it, and i wait for the DK2 with lot of impatience to make demos :)If you are interested or got questions please feel free to contact me.Have a nice day and soon we will all received the DK2 :D :D :D
I saw today in the news a guy is trying to open the vr cafe in Boston before the consumer version is released, and it brought up a few questions. Mainly a good point was made by some feedback to the article " You wear this thing strapped to your face, pressed against your skin. Sharing such intimate equipment with countless strangers is yucky, just one step above sharing toothbrushes."It is a very valid point and for those of us considering opening one of these cafe's, which I have been for quite some time, are there any plans to make the rift clean for multiple users to use in a cafe type setting? Maybe this is a question best suited for the new Team Carbon Design to have the answer for?
Hi there, I don't know why, but the camera is upside in one project - any advice to fix this? I didn't change anything.Thank you! AD
What will happen, how do we choose which to use in the SDK without having to unplug them all the time?
Conventional Rift engine design currently appears to focus on rendering the entire scene twice, once from each eye. The only alternative that seems to get discussed is stereo reprojection using a depth buffer, but that's universally agreed to be unsatisfactory.But actually, there's no need to render your entire scene twice in every circumstance. Thanks to the Rift's display's low angular resolution, there's a fairly low distance threshold beyond which stereoscopic depth cues are no longer distinguishable. We can calculate this threshold distance thus:Average interpupillary distance (IPD) ~= 0.064mBest angular resolution = 1280px horizontal resolution / 90 degrees horizontal view angle = 14.22 pixels per degreeSmallest angular size of a pixel = 1 / 14.22 PPD = 0.07°∴ Distance at which stereoscopic effects are smaller than 1 pixel = 0.064 / tan(0.07°) ~= 52mAdditionally, as FredzL pointed out in IRC, distance stereopsis is not a primary depth cue at longer distances:One technique to t
_As we need More horizontal field of view, than vertical one._As the screen have to be splitted in two for each eyes.Why not having the screen disposed vertically ?So we would have 1080p*960 per eyes instead of 960*1080That would be 10% more horizontal pixel, with the same screen.
I want enable VR in an aircraft but I can't use OVRPlayerController because the rigidbody interferes with the aircraft. OVRCameraController works fine but it doesn't support mouse look, which I need.Is there any way around this?
The short length of the cable between the control box and the Rift itself is limiting for anything but seated or standing-in-place.The obvious solution is to put the control box in a carry pouch. This solution is fine for the video input and USB cables. as these can be extended easily. The issue is with power, as it's not desirable to have to carry around a bulky power adapter and extension.I presume that we could extend the length of the power cable using some extension cable. I've measured the size of the adapter plug as 2.1mm (inner ring) and 5.5mm outer ring, which would indicate that this might do the trick?: http://www.amazon.co.uk/Ex-Pro-Quality-2-1mm-Socket-Extension/dp/B004FVHD7I/The other (better) option is to extend the cable running from the control box to the Rift itself. This might be a bit trickier... has anyone attempted it?
Hello...I was wondering if anyone would like to team up and pool our talents to make something in time for the launch of the Rift at the end of the year or early 2015?My name is Damian Piesse, and like me I am sure there are people out there who feel they have some of the skills needed to get something really cool made for the rift's launch. I feel that we need to strike soon if we want to get something made in time. I am so passionate about this that I have made a site called Urift.com (literally this weekend June29th) to help collect people together in order to try and make something. There probably wont be any money involved (if there is it would shared equally), instead I see this as a labour of love - a stepping stone to forming a company in the future perhaps. Meanwhile we need to get a team together, a team of people who share this kind of idea. We need an awesome idea/inspiration then a set of developers/programmers/artists etc. Perhaps you have an amazing idea burning i
Hey,I've been browsing the forums, and there are some really impressive looking DIY projects and startups over the last few years (since Oculus 'resurrected' VR) focusing on solving the problem of human interaction.However, With the DK2 release coming closer, there's one question on my mind: What's the best off-the-shelf input device for the rift?By off-the-shelf, I don't mean that I can pick up at a supermarket, but are there any data gloves that are available for purchase (not pre-order) now?Preferably, it'd have to be affordable (as I intend to have it as the primarily supported input device for my project). Data gloves seem to look promising, To my understanding, the only one that can currently be bought is the 5DT data glove, but the prices there seem to be extortionate (Just within my budget, but not at all available for consumers)I've also looked into the Kinect, and whilst the full body tracking is a plus, it's simply not a viable choice.So: What input device did you use for th
Hello,I understand that if you have a wireless receiver Dongle hooked to a Windows PC, the XBox controller should "just work" (being Microsoft). For a Mac, you have to download a driver from here:http://tattiebogle.net/index.php/ProjectRoot/Xbox360Controller/OsxDriverAnd (depending on the games) buy a gamepad configuration app off of the Mac App Store.Has anyone had any success with this configuration as an input device for the Oculus? Right now, I'm looking to either order a wireless receiver or a more compatible controller.Thanks,- Scott
Hello rift community.Currently i don't have a rift yet. I am eagerly waiting for my DK2 to start developing (and also checking ut demos and games).I have been thinking about controllers for VR. The Razer Hydra seems to be a nice choice but not really available anymore.I do have a Steelseries SRW-S1 here. The SRW is a steering wheel that is held in the air. It's axes are controlled by motion sensors. It is very accurate and you can play racing games very well with it. It wont be as good as a 400 $ wheel and it will make you tired after long periods of playing. It is for people who don't have the possibility to mount a wheel to a desk or simply don't want to.I just played asseto corsa with it today and at this moment i thought that this device could be quite cool when playing with a rift. The SRW is cablebound but the cable is very long wich means you aren't very restricted in movement.I will report back how it works together with the rift once i get my DK2.
Heyup,I want to open a discussion on modeling the physical characteristics of the Oculus device and the eyes for ray-tasting / ray-tracing / ray-marching.1) Would a one-to-one representation of the device configuration be desirable, even feasible? e.g. Modeling the lens characteristics, its position from the eye (taking into account the eye relief mechanism, lens type), the position and size of the screen... 2) ignoring the technical limitations of eye tracking, would it be possible to modeling the actual eye itself. Pupil, cornea, iris, ... modeling ambient luminosity, focal distance, eye position and direction, even eye conditions, myopia, hyperopia, presbyopia. 3) what shortcuts and optimisations could be taken : I'm thinking a reverse barrel-distortion-type shader.4) How best to represent the distortion to generate ray. Thinking displacement map / normal map type of offset computation. Then the rays for each fragment can be computed quickly. 5) problem : But how would you take into
The absolute best possible controller solution for Oculus is actually to not give us a controller at all. Instead of re inventing the wheel again and running the risk of coming out with something sub par why not just instead give us a cheap positional tracking object that we can attach to a controller we already own.You cant make a controller that everyone is going to like, but you can make the controller that each individual likes work with the sensor. why would they do this?- controllers are expensive! an add on would cost a couple bucks rather than 60$.- I get to use a high quality input device that I like and already own.- controllers are risky and can make or break consoles.- base package is cheaper which most likely means more people choosing to be an early adopter and more money in their pocket to spend on games for the rift.what do you guys think? good idea or crappy idea?
I was wondering if the concept of the Myo could work as a data glove alternative?https://www.youtube.com/watch?v=oWu9TFJjHaMhttps://getmyo.com/
I posted this on Reddit here on the idea of orientation data from either a camera that took the video or from the Rift that was used to record a Let's Play video. That orientation data is then used to shift the screen viewing the video in VR. This creates a pseudo-stabilized view of the video where the rotation of the camera or player's head will not disorient the viewer. The viewer instead will see the screen is moving to the right, so the viewer just looks right to keep up with the action. I've tested this using image stabilized gifs and the result is pleasant to watch in VR. What it looks like is a smaller video projected onto a larger dome or spherical screen then that video is moved around the screen. Wasn't sure the best term for a viewmode like this. Flashlight or Headlight view makes since it would seem as if the view is projected from your center. Being John Malkovich viewmode could also work as a term even though the movie is more like watch replays of VR games on YouTube (sc
I've been brushing up on my GL skills, and I was wondering about the flow through the GL distortion pipeline in the SDK. I'm fundamentally unhappy with the idea of glFlush()/glFinish() being used for synchronization of any kind and even more so with the busy-wait mechanisms designed to block the client from beginning a new frame until the SDK decides the time is right. It seems like one of the fundamental problems you're trying to solve is to ensure that when the next frame starts you have a decent idea of when it's going to be rendered, and thus can provide the most accurate head pose information. Right now this is done by flushing and busy-waiting in endFrame(), but this demonstrably causes dropped frames and a lower overall framerate on my desktop machine. I'm thinking there's a better way.In the distortion rendering, you could instantiate a fence sync object after doing the draw, or performing the buffer swap, and use a queue of unsignaled sync objects as a count of pending fra
It's not out yet, but Sixense have released some info on their upcoming sdk for VR.Apparently at release it will support Hydras and Stem, but they plan on working with other hardware makers to make this a common api for vr.The sdk allows any number of sensors placed anywhere on a body. Inverse kinematics are used to make non tracked bones behave sensibly.The sdk has Unity and UE4 integration, plus has a cross platform C api.The software can take input from headsets such as the rift (and other hmds) for head control and use that with the inverse kinematics to control the skeleton, or fuse both hmd sensors and a stem sensor on the head.Apparently it will even work without and Sixense devices (hydra or stem), all you need is the hmd sensors for a minimum level of support (a body that matches the head).http://sixense.com/sixensevrsdkhttps://www.kickstarter.com/projects/89577853/stem-system-the-best-way-to-interact-with-virtual/posts/823494
I am sure this has been asked, but I can not find it anywhere. Has anyone posted a basic human avatar with instructions on how to put it into unity for the Rift? Just a basic look down and see you legs walking when you move. In the steam game "The Forrest" (sweet game BTW) is another example. I assume the user mounts the camera where the eye sockets or neckbone is located? Any advice would be greatly appreciated thanks!
Im running a Meetup out here in Sacramento CA, shameless plug http://www.meetup.com/Sacramento-Virtual-Reality-Meetup and am wondering how can I get a hold of some Oculus swag? I would be willing to pay and think it would be a great way to spread the word. Im talking about shirts, stickers and what not. cheers, Bullardo
Hi,I've written some programs with the Oculus Rift SDK 0.3.2 and everything worked fine. I ran them with both a 'real' Oculus and the 'false' one, aka created with ovrHmd_CreateDebug(ovrHmd_DK1)But here I've come to an unexpected issue: the following code throws the exception:hmd_ = ovrHmd_Create(0); if (!hmd_) { hmd_ = ovrHmd_CreateDebug(ovrHmd_DK1); if(!hmd_) { throw std::runtime_error("Cannot create the debug dk1"); } }The code for ovrHmd_CreateDebug is quite simple (located in CAPI_HmdState.cpp):OVR_EXPORT ovrHmd ovrHmd_CreateDebug(ovrHmdType type){ if (!GlobalState::pInstance) return 0; HMDState* hmds = new HMDState(type); return hmds;}So I suppose GlobalState::pInstance is null, but how? why? Until now I did not bother with any GlobalState::pInstance...Maybe someone knows the cause of the problem or experienced it before?Thanks for your help
Does anyone have a DK1 they would like to sale? thanks!
http://www.bbc.co.uk/blogs/internet/posts/Virtual-Reality-the-future-of-TVFor our american friends, news filmed in 360 degrees as an experiment, available to download for the rift..Sorry if this is in the wrong section
Already have an account? Login
Enter your E-mail address. We'll send you an e-mail with instructions to reset your password.