Razer Hydra driver for Android/iphone

[Kosen]

Hej VRGeeks from the world,
Back from the Global Game Jam I had the idea, about porting the raze hydra controller to a mobile device.
Do you know if something like that already exist.

I remember I might have seen something similar in the Project Holodeck because there was no wires between the player and the games station.
I you have any idea about how the console system used to do that, I'll be really pleased to understand it.

Anyway do you know what are the prebuild drivers, or if any driver for Android/iphone device already exist with the razer hydra.

One more question : I'm actually working with the HMZ-T2 HMD from sony, do you know what is the hack to get rid of the main current wire and processor Unit.

Thanks for your help.

Kosen

[Christophe Gouet]

Hi Remi,

Could you please be more precise about the Razer Hydra? Do you want to receive it's tracking values in a phone application ?

--
You received this message because you are subscribed to the Google Groups "VR Geeks" group.
To unsubscribe from this group, send email to vr-geeks+u...@googlegroups.com.
Visit this group at http://groups.google.com/group/vr-geeks?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[Remi Cambuzat]

I thought it was enought precise, sorry for the misunderstood.
The idea is the following one, I want to be able to use the Razer Hydra in a phone application without using any other device, such as an external computer.
Everything should be on the smartphone. Instead of plugin your Hydra inside the computer usb, you just plug it in the micro usb slot on phone and then you use it.
It might need a bit of calibration but the idea is here.

2013/1/29 Christophe Gouet <chric...@gmail.com>

[Sébastien 'Cb' Kuntz]

Hi,
Kyle Johnsen from Georgia Tech did exactly this.
I don't remember exactly if he plugged the hydra directly on the phone or if it had an interface in between.
I remember there was an additional but don't remember if it was only a battery or a converter.
Maybe the Android VRPN server can work directly ?

Anyway, he had the hydra base and phone on his hips, so he was completely free to walk.
He used the accelerometer from the phone to detect pace and the magnetometer from the phone also to detect the orientation of the body, so you could walk in the direction of your body.

You can try to contact him directly: http://www.cise.ufl.edu/~kjohnsen/
Let us know about your progress !
cb

[Jan Ciger]

Hello,

On Tue, Jan 29, 2013 at 8:20 AM, Sébastien 'Cb' Kuntz <sebasti...@nowan.net> wrote:

Hi,
Kyle Johnsen from Georgia Tech did exactly this.
I don't remember exactly if he plugged the hydra directly on the phone or if it had an interface in between.

Strictly speaking, it could be doable, because many Android devices have the USB OTG port - it can act as both host and device. However, you would need to implement a HID library for Android to read and decode the data from the Hydra.
On the other hand, you will still need an external power supply - the phone wouldn't be able to power the Hydra alone.
 

I remember there was an additional but don't remember if it was only a battery or a converter.
Maybe the Android VRPN server can work directly ?

Doubt it. Android is not Linux, you don't have many of the APIs that VRPN uses accessible. There probably could be a VRPN server written for Android, the question is what would be the benefit. If you are connecting Hydra to the phone, you likely want something to run on the phone itself, thus no need for VRPN.
 

Jan

[Cédric Syllebranque]

Hi geeks.

Does anyone know if an augmented version of the Occulus Rift is planed,
even in very early prototype stage?

Thanks.

C�dric

[Jan Ciger]

On Tue, Jan 29, 2013 at 10:32 AM, Cédric Syllebranque <syll...@gmail.com> wrote:

Hi geeks.

Does anyone know if an augmented version of the Occulus Rift is planed, even in very early prototype stage?

Uhm, augmented in what sense?

Do you mean a see through version for augmented reality? That I doubt, it would be next to impossible to do with the distorting optics in the Rift. You could probably attach wide angle camera(s) to it and use it as active see-through, but that would still require a lot of processing power to correct out all the distortions. I don't think that they have anything even close to that onboard - powerful DSP or an FPGA for doing the image processing alone would cost more than the entire Rift.

Jan

[Cédric Syllebranque]

Exactly, for active see-through. I think distortion can be efficiently compensated using a precomputed Look Up Table...
Trivisio made an augmented version of one of it's HMDs...

I doubt they planned this too, but if you never ask...

Cédric

--
You received this message because you are subscribed to the Google Groups "VR Geeks" group.

To unsubscribe from this group and stop receiving emails from it, send an email to vr-geeks+u...@googlegroups.com.

[Jan Ciger]

On Tue, Jan 29, 2013 at 4:41 PM, Cédric Syllebranque <syll...@gmail.com> wrote:

Exactly, for active see-through. I think distortion can be efficiently compensated using a precomputed Look Up Table...
Trivisio made an augmented version of one of it's HMDs...

You need to think embedded hw, not something that you can run on a big ass i7 PC. I don't think it would make sense for them to sell an HMD with a webcam attached and claim it is a solution for AR with actually getting the image from the camera to the HMD being the buyer's problem ...

Moreover, the geometry gets complex fast, because you would need expensive fish eye camera(s) (one distortion that has to be removed) in order to be able to cover the huge FOV of the Rift and then you need to pre-distort in real time for the Rift's lenses. That's a lot of calculation to do in real time for HD video, especially if you have two camera streams that have to be merged into a single image needed by the Rift. Even with look-up tables, you are still running through tons of pixels on every frame. It is certainly doable using some high-end FPGA (we are talking about processing video streams at Gbps speeds here, that is what common LCD displays use ...), but it would be a very expensive and a very niche product, IMHO. It makes very little sense commercially if you are going after the mass market.

You could always bolt a camera on your Rift yourself and then do all the processing on the PC - with much increased latency, though.

Jan

[Maxim Lysak]

2013/1/29 Jan Ciger <jan....@gmail.com>

You need to think embedded hw, not something that you can run on a big ass i7 PC. I don't think it would make sense for them to sell an HMD with a webcam attached and claim it is a solution for AR with actually getting the image from the camera to the HMD being the buyer's problem ...

Unfortunately this is what everybody does...

About the subject - considering that Oculus team so focused on fighting latency issues, I cannot imagine they put into device something as slow as webcam at all.

[Jan Ciger]

On Tue, Jan 29, 2013 at 6:13 PM, Maxim Lysak <maxim...@gmail.com> wrote:

2013/1/29 Jan Ciger <jan....@gmail.com>

You need to think embedded hw, not something that you can run on a big ass i7 PC. I don't think it would make sense for them to sell an HMD with a webcam attached and claim it is a solution for AR with actually getting the image from the camera to the HMD being the buyer's problem ...

Unfortunately this is what everybody does...

Unfortunately yes, but when a decent FPGA costs in the order of several hundreds of EUR (just the chip!) and the time of the engineer who has the skills to design the circuitry and firmware for it would cost some tens of thousands of EUR, so it isn't too surprising. The price could go down only if you managed to get a custom chip built - but then you have sell hundreds of thousands units to recoup the costs. The economics just doesn't work ...
 

About the subject - considering that Oculus team so focused on fighting latency issues, I cannot imagine they put into device something as slow as webcam at all.

Yeah, my thoughts exactly. It would be a stupid gimmick.

Jan

[Fivos Doganis]

Sorry but I have to disagree: a video see-through HMD is far from being a gimmick, however "geeky" it might look. I'd even seek to attach two cameras if possible!

As for the processing, if the distortion correction for the 3D rendering can be performed in real-time to cope for the exotic lenses of the Occulus, I don't see why distorting the image captured by the camera would not be possible. And yes, even tethered to a powerful PC with spare GPU time to keep latency as low as possible, I believe that such an add-on would be really interesting for some AR applications!

Fivos

--

[Jan Ciger]

On 01/29/2013 09:09 PM, Fivos Doganis wrote:
> Sorry but I have to disagree: a video see-through HMD is far from being
> a gimmick, however "geeky" it might look. I'd even seek to attach two
> cameras if possible!

You have misread my comment, Fivos. I was referring to the possibility
of Oculus simply attaching a dumb webcam to the Rift and selling it as
an "AR solution" (similar to what Vuzix does) as a gimmick. Not to the
concept as a whole.

>
> As for the processing, if the distortion correction for the 3D rendering
> can be performed in real-time to cope for the exotic lenses of the
> Occulus, I don't see why distorting the image captured by the camera
> would not be possible. And yes, even tethered to a powerful PC with
> spare GPU time to keep latency as low as possible, I believe that such
> an add-on would be really interesting for some AR applications!

If you do it on PC, you don't need Oculus (the company) for that - you
can do it yourself, really. They would just be selling you the camera(s)
on top of the HMD.

I was more thinking along the lines of an integrated solution, where you
actually get a working HMD with correctly working pass-through out of
the box, similar to the way the older i-Glasses and other AR goggles
worked (sure, those were much simpler with optical pass-through). Not
something to "finish at home" i.e. a camera and glasses bolted together
and that's it.

That is a much harder problem with the Rift because of the optical
properties. Vuzix has that easy, because they use fairly narrow FOV on
the displays, so the distortion is minimal and simply hooking up a
cheapo webcam is enough.

Jan

[Fivos Doganis]

Thanks Jan for your detailed explanation!

Is therefore your only concern the ability to correct the distortion from the camera(s) in real time using embedded hardware?

If so, is this not already achieved by the Occulus? If not, with all the OpenGL ES capable mobile platforms around, wouldn't it be possible to use shaders for this correction?

[Jan Ciger]

On 01/29/2013 09:50 PM, Fivos Doganis wrote:
> Thanks Jan for your detailed explanation!
>
> Is therefore your only concern the ability to correct the distortion
> from the camera(s) in real time using embedded hardware?

That's what I was thinking about, otherwise it would not be any
different from the existing products on the market.

>
> If so, is this not already achieved by the Occulus?

Well, I am not in any way affiliated with them, so I cannot say this
with 100% certainty, but I don't think so. The hw in the Rift is only
the tracker/IMU and then a second board which is the VGA/DVI/HDMI to
LVDS driver board for the LCD. Nothing more. The driver board is pretty
much what you would find in a normal LCD monitor as well, at best it can
do some simple things like image scaling (to support different
resolutions) and rotation/image flipping. All that is typically handled
by a dedicated chip, those are mass-manufactured and pretty standard.

I believe they have a 3rd party supplier that they are buying the
displays and the drivers from - that's why they had to change to the
larger panel recently, because the original vendor has disappeared (or
the panel was discontinued, don't remember). You can actually find
similar modules (LCD + driver board) on eBay easily, e.g.:

http://www.ebay.co.uk/itm/7inch-N070ICG-1280-800-IPS-lcd-panel-with-HDMI-VGA-DVI-AUDIO-lcd-driver-board-/181044273324?pt=UK_Computing_Laptop_Screens_LCD_Panels&hash=item2a271460ac

> If not, with all
> the OpenGL ES capable mobile platforms around, wouldn't it be
> possible to use shaders for this correction?

Sure, it would, but you still need a fairly high speed connection for
the cameras. I am not sure that something like a Cortex A9 could handle
two cameras at full frame rate at once without using a dedicated hw.
Maybe it could, haven't tried it.

Moreover, you don't really want to do this in software if you can avoid
it, because the latency between capturing images from the two cameras
does matter. Just streaming over USB adds a fairly significant delay to
the signal. That's why I spoke about using an FPGA - all the capturing,
undistortion, fusing of the camera images and pre-distortion could be
done in hardware. Shaders can handle only the image warping, but none of
the rest.

Jan

[Sly]

You don't need to distort anything,
just put the same lense on the camera as the one on the occulus and you are good to go !
Or a least one that distort in the apropriate way.

Sly.

2013/1/29 Jan Ciger <jan....@gmail.com>

Jan

--
You received this message because you are subscribed to the Google Groups "VR Geeks" group.

To unsubscribe from this group and stop receiving emails from it, send an email to vr-geeks+unsubscribe@googlegroups.com.

[Palmer Luckey]

That would be an ideal solution in theory, but it has some problems in practice.

Different people will have slightly different distortion parameters based on eye relief, focal distance, facial geometry, etc, to say nothing of people who modify the headset with different viewing optics.  Fixing the distortion optically gives no flexibility to change for any of those factors.

To unsubscribe from this group and stop receiving emails from it, send an email to vr-geeks+u...@googlegroups.com.

[Jan Ciger]

On Wed, Jan 30, 2013 at 3:32 AM, Palmer Luckey <palme...@gmail.com> wrote:

That would be an ideal solution in theory, but it has some problems in practice.

Different people will have slightly different distortion parameters based on eye relief, focal distance, facial geometry, etc, to say nothing of people who modify the headset with different viewing optics.  Fixing the distortion optically gives no flexibility to change for any of those factors.

Yup. And never mind that the lens which is good for viewing will likely make a really poor lens for a camera. Having such lens custom made would likely cost more than the device itself (off the shelf CCD camera lenses cost ~$100 already).

Jan

[Lorne Covington]

On 1/30/2013 4:59 AM, Jan Ciger wrote:

... (off the shelf CCD camera lenses cost ~$100 already).

Nope.  $25 for six standard M12 mount lenses - $4.17 each.  I use these on modified PS3Eye cams (the M12 lens mount is $20). 

Cheers!

- Lorne

-- 

http://noirflux.com

[Jan Ciger]

On Wed, Jan 30, 2013 at 4:27 PM, Lorne Covington <noir...@gmail.com> wrote:

On 1/30/2013 4:59 AM, Jan Ciger wrote:

... (off the shelf CCD camera lenses cost ~$100 already).

Nope.  $25 for six standard M12 mount lenses - $4.17 each.  I use these on modified PS3Eye cams (the M12 lens mount is $20). 

That is not exactly the kind of lens I meant. Industrial/research cameras use C-mount (e.g. PointGrey products). You don't want to use a hacked webcam for a serious product like that, do you?

This is more like it:
http://www.thorlabs.de/newgrouppage9.cfm?objectgroup_id=1822

The cameras are around $300-$800 each, depending on type, framerate and resolution. Add the lenses, around $120-$150 each and now we are talking. And these are still off-the-shelf components, nothing custom.

Jan

[Lorne Covington]

On 1/30/2013 10:41 AM, Jan Ciger wrote:

On Wed, Jan 30, 2013 at 4:27 PM, Lorne Covington <noir...@gmail.com> wrote:

On 1/30/2013 4:59 AM, Jan Ciger wrote:

... (off the shelf CCD camera lenses cost ~$100 already).

Nope.  $25 for six standard M12 mount lenses - $4.17 each.  I use these on modified PS3Eye cams (the M12 lens mount is $20). 

That is not exactly the kind of lens I meant. Industrial/research cameras use C-mount (e.g. PointGrey products). You don't want to use a hacked webcam for a serious product like that, do you?

I thought we were talking about something to go on a head-mounted display, not industrial/research cameras!  I don't want to try and wear something using those lenses.

And in any case you're making a straw man - you can certainly find all sorts of expensive cameras and lenses, but many devices and most phones have hig-res cameras in them nowadays, and those sensors and lenses are nice and compact, and as a result of volume production available inexpensively.

I'm struck by how this thread has been characterized by what seems to me to be a lot of nay-saying and pedantic thinking instead of creative problem solving.  I think it would be GREAT to have cheap, fast, hi-res cameras to go on the Oculus!  C'mon, necessity is the mother of invention, and constraints are the fire of creativity!  Let's hear some positive ideas about what MIGHT work instead of all the reasons something WON'T.  I thought this was the VR-Geeks list, not an internal engineering review at IBM.

Cheers!

- Lorne

[Jan Ciger]

Hi Lorne,

On Wed, Jan 30, 2013 at 5:16 PM, Lorne Covington <noir...@gmail.com> wrote:

And in any case you're making a straw man - you can certainly find all sorts of expensive cameras and lenses, but many devices and most phones have hig-res cameras in them nowadays, and those sensors and lenses are nice and compact, and as a result of volume production available inexpensively.

Right, but I haven't seen a phone that has fish-eye lens with ~120 degrees FOV, did you? Such lenses are really hard to produce for small sensors (optical/physical constraints) without complicated tricks and are not exactly common. Sure, there are the cheap-ass lenses like this one: http://iphonephotogear.com/?p=1 but you get what you pay for ... That's why I was more thinking about a "pro" camera + lens, that are actually available. Sure, you wouldn't want to wear two CCTV cams on your head, but the prices for would be likely similar for a nice board camera + decent lens.

Still, even if you find camera/lens, none of the other things outlined before are addressed. The cost of the cameras was really a tangential issue to the entire discussion.
 

I'm struck by how this thread has been characterized by what seems to me to be a lot of nay-saying and pedantic thinking instead of creative problem solving.  I think it would be GREAT to have cheap, fast, hi-res cameras to go on the Oculus! 

I agree.
 

C'mon, necessity is the mother of invention, and constraints are the fire of creativity!  Let's hear some positive ideas about what MIGHT work instead of all the reasons something WON'T.  I thought this was the VR-Geeks list, not an internal engineering review at IBM.

I think it is quite important to be realistic with what is actually technically (and commercially!) doable and what is fantasy and wishful thinking. There is also a big difference between what would work as a neat hack in a lab (I love PS3 Eye cams too!) and what is actually usable in a commercial product (where you would likely use a board cam, with better resolution, frame rate, external sync to be able to merge image from two cams, etc.). Otherwise we are risking to devolve to the level of Gizmodo with their sensational headlines about "paradigm shifts" and how this or that tech is going revolutionize computing ...

And yeah, I am known to be a naysayer, especially when I see something that wasn't thought through to the end. So feel free to call me pedantic, but if something is to actually work in the end, those "details" will not disappear by themselves. I am an engineer doing this for a living, not a manager that just handwaves the issue away and lets the engineers figure it out somehow, no matter how crazy the requirements are.

I also see a lot of people saying here "Oculus should do this", "Oculus should provide that", well, they don't, sorry. If someone thinks that such and such thing is a good idea, feel free to design a product and start a business - I am sure it would be a market hit and sell well.

I am going to end my rant here, but it tends to tick me off when people attack me with being negative about something instead of actually addressing the arguments being brought up.

Regards,

Jan

[Naem BARON]

Hello world,

the idea is interesting, I was thinking of it like a possible addon to the Oculus Rift and not necessarily a built-in solution.

Like Lorne, I think a "phone-camera" would be affordable for this purpose.

There is not even a need for a HD cam, the OcRi (personal abreviation) has 1280*800 panel, so 640*800 per eye.

Two 640*800 (or 1280*720 due to ratio restriction) cameras could do the job.

For the real-time processing (distort, merge, predistort), since the OcRi is not wireless, we could use the computer.

==> Yes, bad and ugly latency will come ! But it's always difficult to have the cake and eat it !

The main question would be:

Does the latency is small enough to allow a good experience ?

If yes, we got our cake.

C U

Naëm

---

De : vr-g...@googlegroups.com [vr-g...@googlegroups.com] de la part de Lorne Covington [noir...@gmail.com]
Date d'envoi : mercredi 30 janvier 2013 16:16
À : vr-g...@googlegroups.com
Objet : Re: [VR Geeks] Augmented Occulus

[Naem BARON]

re-Hello world,

I forgot about the lenses... I am not an expert at all regarding this, but what kind of inconveniences are added by something like that : http://iphonephotogear.com/?p=1 ?

C U

Nems

return 0;

---

De : vr-g...@googlegroups.com [vr-g...@googlegroups.com] de la part de Naem BARON [ba...@et.esiea-ouest.fr]
Date d'envoi : mercredi 30 janvier 2013 17:42
À : vr-g...@googlegroups.com
Objet : RE : [VR Geeks] Augmented Occulus

[Jan Ciger]

On Wed, Jan 30, 2013 at 6:42 PM, Naem BARON <ba...@et.esiea-ouest.fr> wrote:

Hello world,

the idea is interesting, I was thinking of it like a possible addon to the Oculus Rift and not necessarily a built-in solution.

Like Lorne, I think a "phone-camera" would be affordable for this purpose.

There is not even a need for a HD cam, the OcRi (personal abreviation) has 1280*800 panel, so 640*800 per eye.

Two 640*800 (or 1280*720 due to ratio restriction) cameras could do the job.

Um, isn't that HD cam already? Anything above 640x480 or PAL resolution is called HD usually. 1280x720 is a 720p cam ...
 

For the real-time processing (distort, merge, predistort), since the OcRi is not wireless, we could use the computer.

==> Yes, bad and ugly latency will come ! But it's always difficult to have the cake and eat it !

The main question would be:

Does the latency is small enough to allow a good experience ?

I guess that would have to be tested. However, from my experience with cameras, you have noticeable "delay" (latency) between moving the camera and the image moving on screen, even without any processing. It is noticeable even at 60fps with the PS3 Move cam. I am not sure what would that do regarding the motion sickness if you are using that inside the Rift, with the huge FOV and no outside visual references. This sort of latency is usually not a big deal with a normal, optical see-through, where it affects only the displayed augmentations but it could be a big issue with a camera-based see-through. Most likely you would want to limit yourself to slow movements - the blur can be very disorienting even in a normal HMD like eMagin z800 that I tried something like this with.

Fortunately, it is easy to test - connect the image from a webcam to the HMD and see whether you can live with the latency or whether it feels like turning inside of a jar of jelly.

Undistortion/pre-distortion is usually doable within a frame on a reasonably powerful PC these days, the capture/transfer/display panel latency is more significant.

Regards,

J.

[Guillaume Depestèle]

Hello,

Except the camera need to have quick shutter and fast processing to send the image to the screen who need to be also fast. I don't know the application of that but it's seems very expensive.

Wait for fast oled with alpha :-D

[Jan Ciger]

On 01/30/2013 06:49 PM, Naem BARON wrote:
> re-Hello world,
>
> I forgot about the lenses... I am not an expert at all regarding this,
> but what kind of inconveniences are added by something like that :
> http://iphonephotogear.com/?p=1 ?
>

I haven't seen this particular gadget first hand, but it is a lens that
doesn't have a mount, for starters - you glue/stick it on your phone. It
could work for a prototype, though, if you manage to attach it to your
camera somehow and manage to get the image in focus - the distance
between the sensor and the lens is critical, so it may not fit
everywhere (but that is true about any lens, it is a little simpler if
you have lenses meant for your camera in the first place).

Also, who knows what is the optical quality of it - it likely doesn't
have first class glass inside, considering its price. The photos taken
with it look OK. At least it seems to have the antireflex coating and
the lenses are actually glass, not plastic.

If you want a very cheap fish eye lens, then door viewholes
(http://www.asia-manufacturer.com/whousepic/a48/38715/pb_93ns1331108039.jpg)
were reused for this purpose before - you unscrew the two halves apart
and one (the one that sits on the outside of the door) has a fish eye
lens. Of course, this isn't really photographic quality lens, but may be
easier to get.

Jan

[Palmer Luckey]

The problem with all these fisheye lenses is that they were made with one thing in mind: Large field of view.  Distortion does not matter, and it does not even matter if the manufacturing tolerances result in 10% differences between different lenses.

Going to a lens maker and asking for a wide field of view lens is easy.  Going to a lens maker and asking for a wide field of view with a very specific distortion curve is harder.  Asking them to make every one of those lenses identical to one another is harder still.

Technically speaking, all of these methods are very possible.  An FPGA or a custom ASIC would probably be "best", but the volume required to make those cost effective is enormous.

--
You received this message because you are subscribed to the Google Groups "VR Geeks" group.

To unsubscribe from this group and stop receiving emails from it, send an email to vr-geeks+unsubscribe@googlegroups.com.

[Cédric Syllebranque]

Wow, I did not think this thread would go so far in every directions when I started it.

When I asked "Do you know if an augmented occulus is planned?" (Jan, I did not say "The occulus should be..."), I had two ideas in mind and I am glad that both went in the field by themselves. This proves that VR geeks is, at least, what I needed because debating about such precise considerations is nearly impossible locally.

First, I was thinking "prototype". It is indeed quite fast to develop an augmented first prototype. I did it for eMagin with some ugly webcams, but result was not so bad, even with a huge offset between cams and lenses! This tend to prove that, okay, it is good to be as close as possible to the sight line, but human brain can correct a lot of parameter by itself and many labs are studying it. The real question is "will my application need ... low latency, low distortion, etc...".

Secondly, I had indeed an "industrial" version in mind. I agree here that considerations are far from above, because manufacturer can not focus on a given application but has to be the more "generic" possible, and of course, need to take volume and manufacturing cost into account. I am glad to see that this is not unfeasable, but maybe the market for mass production is still not there. Thus, maybe there is room for "addons", more dedicated ones, which, depending on their goals, can be more expensive if they are "sold" with the given applications.

I hope everybody will keep beeing creative and that Jan (or anyone else) will make dreamers stay on the ground, because both are indispensable to innovate and create good and durable products...

Cheers geeks...

To unsubscribe from this group and stop receiving emails from it, send an email to vr-geeks+u...@googlegroups.com.

Visit this group at http://groups.google.com/group/vr-geeks?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.

--
You received this message because you are subscribed to the Google Groups "VR Geeks" group.

To unsubscribe from this group and stop receiving emails from it, send an email to vr-geeks+u...@googlegroups.com.

[Jan Ciger]

On Thu, Jan 31, 2013 at 10:39 AM, Cédric Syllebranque <syll...@gmail.com> wrote:

Wow, I did not think this thread would go so far in every directions when I started it.

When I asked "Do you know if an augmented occulus is planned?" (Jan, I did not say "The occulus should be..."),

That wasn't really addressed at you, Cédric, I am sorry if it looked like that. There were other threads here and also a discussion on Twitter about hand tracking and what not.
 

Secondly, I had indeed an "industrial" version in mind. I agree here that considerations are far from above, because manufacturer can not focus on a given application but has to be the more "generic" possible, and of course, need to take volume and manufacturing cost into account. I am glad to see that this is not unfeasable, but maybe the market for mass production is still not there.

Yup, my thoughts too. This is a super niche market, unfortunately, so it keeps the costs high.
 

I hope everybody will keep beeing creative and that Jan (or anyone else) will make dreamers stay on the ground, because both are indispensable to innovate and create good and durable products...

Well said :)

Jan

[Jan Ciger]

On Wed, Jan 30, 2013 at 11:06 PM, Palmer Luckey <palme...@gmail.com> wrote:

The problem with all these fisheye lenses is that they were made with one thing in mind: Large field of view.  Distortion does not matter, and it does not even matter if the manufacturing tolerances result in 10% differences between different lenses.

Going to a lens maker and asking for a wide field of view lens is easy.  Going to a lens maker and asking for a wide field of view with a very specific distortion curve is harder.  Asking them to make every one of those lenses identical to one another is harder still.

Yup, you would need to calibrate/measure each lens and the correct the distortion in hw/sw. That is more hassle and problems, of course - either it would have to be done during production (extra costs on every unit) or by the end-user (hassle, possibly poor user experience if something goes wrong ...) Probably not something one would want in a consumer product.
 

Technically speaking, all of these methods are very possible.  An FPGA or a custom ASIC would probably be "best", but the volume required to make those cost effective is enormous.

Yup. Custom ASIC of the size needed for a job like this would likely never pay for itself - you would have to sell 100k units or so to offset the development and production costs. That is just not realistic at all with such product. Of course, unless Apple decides that the next hip must have are white iGoggles with a camera and integrated phone on them and everyone will run to buy them in order to not be seen as backward and behind the curve when talking to their friends.

Regards,

Jan

[Ryan Pavlik]

I wonder if HIDAPI would work on Android - that's the only dependency of the VRPN driver for the Hydra. If it would, then this should be no problem. You might try posting on the VRPN mailing list.

The only catch is that the power for the Hydra is from USB, so you'd probably need to use a powered hub in addition to the USB OTG connector.

You don't need to run the vrpn server as a separate app, either - you can use it in a single process (see client_server.C in server_src for example) as just a device driver library - "server" and "client" in the same process and thread. Works great. No sense in re-writing drivers :) Actually, a lot of times when I'm writing a "nicer" server app, I'll include some client code as well so the server app shows the state of the device, so effectively I do this all the time :D

Ryan

On Tue, Jan 29, 2013 at 4:45 AM, Jan Ciger <jan....@gmail.com> wrote:

Hello,

On Tue, Jan 29, 2013 at 8:20 AM, Sébastien 'Cb' Kuntz <sebasti...@nowan.net> wrote:

Hi,
Kyle Johnsen from Georgia Tech did exactly this.
I don't remember exactly if he plugged the hydra directly on the phone or if it had an interface in between.

Strictly speaking, it could be doable, because many Android devices have the USB OTG port - it can act as both host and device. However, you would need to implement a HID library for Android to read and decode the data from the Hydra.
On the other hand, you will still need an external power supply - the phone wouldn't be able to power the Hydra alone.
 

I remember there was an additional but don't remember if it was only a battery or a converter.
Maybe the Android VRPN server can work directly ?

Doubt it. Android is not Linux, you don't have many of the APIs that VRPN uses accessible. There probably could be a VRPN server written for Android, the question is what would be the benefit. If you are connecting Hydra to the phone, you likely want something to run on the phone itself, thus no need for VRPN.
 

Jan

[Jan Ciger]

On Fri, Feb 1, 2013 at 11:37 PM, Ryan Pavlik <ryan....@gmail.com> wrote:

I wonder if HIDAPI would work on Android - that's the only dependency of the VRPN driver for the Hydra. If it would, then this should be no problem. You might try posting on the VRPN mailing list.

HIDAPI on Linux depends on libusb. I am not sure whether that is available on Android, along with the required syscalls interfaces in the C library.

Moreover, permissions for hw access would be a problem - it would work most likely only on rooted devices, because Android doesn't seem to have any way to make devices directly accessible (via sysfs or /dev) to user applications, so your app wouldn't be able to open the device file.

Jan