Meant to be Seen

FatMcNasty wrote:

davidmi58 wrote:I wonder is there some where on the board I could link off to power 3 IR led's. Then I could use free track. for head tracking.

You can get 3.3V from C502, or you may use the the AVR-ISP pinout on the other side, it has 3.3V on pin marked VCC (against GND). Do not forget to use a properly calculated series resistor on each diode.

rozsnyo wrote:

FatMcNasty wrote:

davidmi58 wrote:I wonder is there some where on the board I could link off to power 3 IR led's. Then I could use free track. for head tracking.

You can get 3.3V from C502, or you may use the the AVR-ISP pinout on the other side, it has 3.3V on pin marked VCC (against GND). Do not forget to use a properly calculated series resistor on each diode.

I was looking at that while you were replying! and was just thinking hmmm could I use that. LOL Ill have to fire up the 3d printer to make a bracket and get some 1.2v IR led's ordered. And a resistor.. I can snake out a mini JST connector so I can pop it off to rebox the Rift. Thanks rozsnyo for the fast response. When I get around to doing it ill post pic if the others want to see it.

Hopefully he will make me a backup for my DK1 and CV1.

I'm fairly certain the LEDs on the DK2 are actually sequenced / pulsed.

It has driver chips in there with a total of 80 output channels!...

https://d3nevzfk7ii3be.cloudfront.net/i ... yDbouAWsCN
http://www.st.com/web/catalog/sense_pow ... 3/PF250094

https://www.ifixit.com/Teardown/Oculus+ ... down/27613

Soon after my DK2 arrives, I'll be doing some experiments to see what the pulsing sequence is.

They are probably synchronized to VSYNC or the camera sync?

OzOnE.
EDIT: Interestingly, they've left the touch digitizer on the front of the Note 3 display - I wonder if the contrast could be improved by buying a panel without the digitizer?

OzOnE2k10 wrote:I'm fairly certain the LEDs on the DK2 are actually sequenced / pulsed.
Soon after my DK2 arrives, I'll be doing some experiments to see what the pulsing sequence is.

Might be just patterned, to know that it is tracking exactly its head and not somebody others.

OzOnE2k10 wrote:They are probably synchronized to VSYNC or the camera sync?

That would be hard without direct link. USB is quite poor when it comes to precise timing.

OzOnE2k10 wrote: Interestingly, they've left the touch digitizer on the front of the Note 3 display - I wonder if the contrast could be improved by buying a panel without the digitizer?

That is rather just a cheap protection against touching the surface of the panel, which is not made in a way that it should be exposed. This is also the reason you must never touch the surface of the RiftUP LCD. You would scratch it while cleaning very easily, not to mention you can wash away some important coating You might get about 5-10% more brightness, I do not think it has any effect on contrast, which is in OLED infinite anyway

You can not get cheap cover glass and it would take lot of effort to couple it with the screen properly, so you do not have the dust there... so use of the whole touch which is made in large quantities is still cheaper.

That's true - I forgot it is OLED on the DK2.

Still having a few issues with my RiftUP kit btw - it didn't want to work the last time I tried it, just a mild grey flashing?

I think I still need to shield the cables a bit more, but it's music festival season, so will be a while before I get around to it.

I heard something about the DK2 having a sync signal for extra cameras, or for mounting another camera on the Rift itself, so that might be what the 2.5mm jack is for?

So, you could be right in that all the LEDs are pulsed at the same time.
The multiple chips is just an easy way to drive them and keep the relative brightness the same?

Although, they are run from SPI, so they could be using some interesting patterns to ID each LED.
That would mean the camera only seeing a few LEDs per frame so the latency would be pretty bad.

Probably just pulsed all ON at the same time then, and that's what seems to be happening in the YouTube vids I've seen.

I'm intrigued by the integrated latency tester too - I bet they just take into account the fixed latency of the panel controller, then detect when the first pixels comes out of the HDMI receiver. :p

I can't see a light sensor in there at least, so the above idea sounds feasible (apologies if this has previously been discussed, I haven't been keeping up with all the threads.)

OzOnE.

Sorry if this is OT, but this is really cool...

https://www.youtube.com/watch?v=FJwljGamIGw#t=57

They NEED something like this on the CV1, and cameras on the front for AR (or simply so you can see the damned desk / keyboard / mouse / drink more easily. lol)

OzOnE2k10 wrote:That's true - I forgot it is OLED on the DK2.
I can't see a light sensor in there at least, so the above idea sounds feasible (apologies if this has previously been discussed, I haven't been keeping up with all the threads.)
OzOnE.

Especially at OLED you can easily measure consumption Or they just might use the VS/TE signal which is normally used for syncing the touch array (latency up to 1 frame, or lets say it correctly: phase).

While I remember - Mike did an awesome vid on the MIPI protocol last year...

https://www.youtube.com/watch?v=7TedIzmguP0

I never really looked into it properly at the time, but it looks like it will be relatively easy to generate a MIPI output directly from the FPGA (with a few resistors)...

http://www.electricstuff.co.uk/nanohack.html

Obviously, the smaller / cheaper FPGAs are still fairly limited on the maximum pixel clock output due to the DDR / DDIO stuff, but might be usable for 720p and maybe 1080p with a bigger FPGA?

Then again, the Solomon chip is cheap enough, so probably not much point generating MIPI directly on one chip?

OzOnE

Some extra MIPI bits here too...

https://mega.co.nz/#!iopFwYhK!9FNduE1TI ... VdjKfEtuQ4

Here's another project which uses a similar method to Mike for the MIPI level conversion...

https://wiki.hackerspace.pl/serial-disp ... d:hardware

OzOnE.

OzOnE2k10 wrote:While I remember - Mike did an awesome vid on the MIPI protocol last year...

https://www.youtube.com/watch?v=7TedIzmguP0

I never really looked into it properly at the time, but it looks like it will be relatively easy to generate a MIPI output directly from the FPGA (with a few resistors)...

http://www.electricstuff.co.uk/nanohack.html

Obviously, the smaller / cheaper FPGAs are still fairly limited on the maximum pixel clock output due to the DDR / DDIO stuff, but might be usable for 720p and maybe 1080p with a bigger FPGA?

Then again, the Solomon chip is cheap enough, so probably not much point generating MIPI directly on one chip?

OzOnE

You might also just read the specs - which gives you a little more insight than just randomly look if you can do it. E.g. there is a CRC and without specs it is not so easy to determine what is the algorithm and what is subject of CRC, and corner case treatment. You would need bruteforce for that... why for if you can just read it and implement straightforward?

I know, I just thought it was a very interesting vid.

When Mike first started, he didn't have any of the specs, so it was good to see some reversing techniques.
Yes, finding the proper CRC and other stuff would have been very difficult without the specs though.

The FPGA stuff I thought might be handy though, as it might allow us to further simplify future designs for the Rift and make the new panels more universal.

OzOnE.

Oh - Daniel, did you receive the other RiftUP! kit back from me OK?

I completely forgot to let you know that I sent it back weeks ago.

Unfortunately, I didn't get chance to wrap the box before my Father posted everything, so I hope it reached you OK?

I've been wondering on how to drive an oled 1440p screen. The 1440p panels I've seen use 2ch 4-lane MIPI. I assume that is the same as 8-lane 1Gbps MIPI.

Could two 4-lane chips be combined with a displayport receiver?

I assume if you could get it working you could get maybe 75-85Hz refresh rate at 1440p.


I'm also curious on getting a single 1080p OLED running at 120Hz. Could 2 MIPI chips be used for 2ch 2-lane and be overclocked? Each channel outputting to 1080x960 or 540x1920 @120Hz?

Inscothen wrote:I've been wondering on how to drive an oled 1440p screen. The 1440p panels I've seen use 2ch 4-lane MIPI. I assume that is the same as 8-lane 1Gbps MIPI.
Could two 4-lane chips be combined with a displayport receiver?

Get me the screen and datasheet, We know how to drive it.

Inscothen wrote:I'm also curious on getting a single 1080p OLED running at 120Hz. Could 2 MIPI chips be used for 2ch 2-lane and be overclocked? Each channel outputting to 1080x960 or 540x1920 @120Hz?

Overclock by 100%? That will not work.

Back again (been on holiday for a week).. I was wondering if anyone else has managed to get any of the "Direct to Rift" versions of demos working with the RiftUp running the patched (or unpatched) OVRService? I've not had any luck so far, everything just shows up on my main monitor nothing seems to activate the rift's display. The only thing that appeared to work was the original Oculus World Demo, the later (0.4.1) Unity Tuscany Demo doesn't work either.

Also, if I hit the space bar for the stats I see that it claims the resolution it is displaying is 2364 x 1641, which seems a bit odd, particularly as the window it is displaying in just fills my 1080p monitor.

Finally, I cracked, having cancelled my original DK2 order, I decided a couple of days ago my life was incomplete without one, so I've ordered one, won't ship until September however. I couldn't resist the lure of higher refresh, low persistence, improved optics and positional tracking (that last one was the straw that broke the camel's back, particularly in Elite Dangerous).

rozsnyo wrote:This is also the reason you must never touch the surface of the RiftUP LCD. You would scratch it while cleaning very easily, not to mention you can wash away some important coating

That would have been good to know sooner.. I've been at it with a soft cleaning cloth a couple of times to remove pesky dust specks (still have to have another go.. there are more). Is there a recommended way to clean the screen in the event of dust specks?

DrBeef wrote:

rozsnyo wrote:This is also the reason you must never touch the surface of the RiftUP LCD. You would scratch it while cleaning very easily, not to mention you can wash away some important coating

That would have been good to know sooner.. I've been at it with a soft cleaning cloth a couple of times to remove pesky dust specks (still have to have another go.. there are more). Is there a recommended way to clean the screen in the event of dust specks?

I use my wife's cleaning set for glasses. You shall be fine with that sort of liquid and the soft material it comes with. Or just wipe the dust with paint brush / system used for cleaning DSLR sensor chips. If you do not want to get dust back, do it in front of a fan.

It would be awhile before I can get a 1440p OLED panel. A datasheet would be hard to come by for me unless leaked. Thanks for the offer though. I was just curious on how it can be driven. I assume it'll use something like this chip I posted before:

Inscothen wrote:rozsnyo- I think for 2560x1600@60Hz, you may want the solomon SSD2858.

http://www.sid.org/Portals/5/pdf/Solomo ... DW2013.pdf

What is the limitation keeping a 1080p OLED from being driven @120Hz? The panel's driver ic? Inability to combine displayport or dual link dvi receiver >> two MIPI transmitters >> single panel?

The DK2's panel has a resolution/refresh available in windows for 1080x960@120Hz. Worst case would be 2 panels side by side in portrait like Valve'a prototype each running at 960x1080@95Hz.

Hello

I am troubled by the following matter.
Please let me know good solution.

1.I connected the riftup! to PC.

2.The riftup! has been recognized by PC.

3.The riftup! was connected with HDMI to HDMI.

4.However, the property of a screen is displayed as DVI.

5.The screen of a riftup! revolving 90 degrees.

PC VGA is nvidia GTM555

OS win7

Thank you.

Hi, saru099,

I wouldn't worry about the DVI thing. Some graphics cards just show it as that, as it's electrically very similar to HDMI anyway.

You will have to change the screen rotation for the RiftUP! display though.
It should be in your nVidia settings somewhere?

If not, try downloading the latest drivers for your gfx card, or look in the Windows display properties for the rotation option.

You should see when you have the correct rotation - it's usually Portrait / Flipped on my GTX 670.

If your graphics card / drivers doesn't have the rotation option, then it's a bit of a problem. Most of them do these days though.

OzOnE.

Thank you OzOnE

I tried the following.

Installed nividia newest driver.

use Windows display properties for the rotation option.

However, it is not improved.

LED is green during connection with HDMI to HDMI.

I use laptop PC.

DK1 working no problem.

Hi,

If the screen is showing an image fine, but it's still rotated, then it can only be done to a software issue really.

Is there any change to the image at all when trying to rotation settings on the RiftUP output in the graphics driver?
If it's rotated correctly, but the image is flipped the wrong way, then there's normally a setting for that too in the nVidia Control Panel?

This is the same for the DK2 stock panel I believe - it still requires the rotation to be done by the GPU.
EDIT: The RiftUP itself does not do any processing to the image as such, so the gfx card must output to that display in Portrait / Flipped mode.

If it's a laptop, then there may be some quirks of the drivers which don't allow proper rotation.
That's what it sounds like to me at least, 'cos all the desktop card nVidia drivers for the last few years should handle rotation / flipped just fine?

OzOnE.

If it is a laptop, it usually prefers output cloning. You might want to disable that and output image only to RiftUp.

Thank you OzOnE2k10,Thank you rozsnyo

I tried all the methods of advice.

When a clone option used

1.The riftup! screen is not unified(Right-and-left division)

2.Change of resolution is invalid.

I would like to know the meaning of a diode green color.

LED colors:
- RED: hdmi input expected, no signal
- ORANGE: dvi input expected, no signal
- GREEN: signal found. Here the screen backlight is enabled and image from PC should appear

@saru009 - supply photo of the screen inside and screenshot of your desktop when showing resolution configuration by email to riftup@rozsnyo.com

Screen resolution is an evident improvement.
But I'm quite bothered by the "ghosting" effect (not moving, everything is clean and with clear cut, just moving a little the head, not very quickly, makes object edges blurred one moment, until they finally become clear when I don't move). I may make a video about this.

I don't know how is the ghosting on DK2 but if it's better is it possible to adapt their solutions to the RiftUp! ? (I think it's a solution like software-based shutters where blurred images are not displayed)

ThePhilosopher wrote:Screen resolution is an evident improvement.
But I'm quite bothered by the "ghosting" effect (not moving, everything is clean and with clear cut, just moving a little the head, not very quickly, makes object edges blurred one moment, until they finally become clear when I don't move). I may make a video about this.

I don't know how is the ghosting on DK2 but if it's better is it possible to adapt their solutions to the RiftUp! ? (I think it's a solution like software-based shutters where blurred images are not displayed)

The difference to DK2 is hardware related, they use OLED displays, not LCD. You can't fix that with software *.

*) overdrive can be done in sw, but it might introduce another set of image quality problems.

New Vorpx DK2 update is out....Anyone having any success in getting Vorpx to recognize their Riftup?

willibeu wrote:New Vorpx DK2 update is out....Anyone having any success in getting Vorpx to recognize their Riftup?

Nope, just doesn't seem to work.. guess it isn't prepared to deal with a "DKHD2 prototype".. gutted as I was really looking forward to playing Bioshock Infinite with proper stereo rendering. Interested to hear what others with DK2s make of it, I will look in envy until my DK2 arrives (whenever that might be )

Dr Beef ....others who have DK2 's are getting ' the device not found ' message also....so I'm thinking there may still be hope. Ralf posted ..all vorpx does is initialize the rift and reads its name. Any way to change its name?

I doubt we can solve this one. VorpX doesn't get the HMD info direct from the rift, it connects to the OVRService which returns the details of the HMD back to the calling process. The problem is the ovrsevice uses a number of things to figure out what the connected HMD is, the first of which is the EDID of the display, in the case of the RiftUp this is still OVR0001, for a DK2 it is OVR0003, so we would need to flash New firmware to the riftup controller board. It also uses screen dimensions, that's why it identifies as a DKHD2 prototype rather than a DK1, I don't think faking a DK1 would be good as you'd get lots of rendering issues like we saw with the previous VorpX.
If we had the source to build our own OVRService we would be able to do something, but I am fairly sure there's little we can do I'm afraid. I will have a more detailed look as I would love to see if working, but don't get your hopes up. It might be easier to convince Ralph to let VorpX recognise the DKHD2 as a DK2.

OK, I Had a bit of a play about (read: Hack), my assertion was incorrect.. VorpX (and in fact, the Oculus Config Utility!) reads the string returned from the OVRService to identify the headset. Using the trusty hex editor, I change the "Oculus Rift DKHD2" to read "Oculus Rift DK2".. and surprise surprise, VorpX now works!

Also, opening the config util, it now reports me as possessing a DK2!, the Demo Scene button was enabled, though I haven't tried it out yet.

I think the problem is, because it now thinks it is a DK2, which has a smaller screen, it isn't using the screen size we flashed to the rift, but those hard coded into the OVRService. The result is a little bit of cross eyed ness, not as bad as it was with the RIftUp originally, but I could feel something. Hopefully fiddling with the VorpX IPD might sort it out.

If you want to try it yourself edit your OVRService with a hex editor and search for DKHD2 and change it to "DK2" inserting 0s after it.

I don't have any more time to mess about with this now, but would be interested to hear if anyone else has any luck with this.. seems we have a solution of sorts anyway.

Great ...Thank you Drbeef.....Will try that later today....the next play was to talk to ralph and see if he could insert the name of the riftup into Vorpx like you said. ... Thanks again.

Works like a charm....that was huge!!! Just played Dishonored for a little bit and it looks fantastic. FOV needed adjustment in the vorpx settings and I turned up the 3D. For others just remember in your Hex editor to add four Zero's after changing DKHD2 to DK2 otherwise the exe will not run. Dr beef ...Well done sir!!!

willibeu wrote:Works like a charm....that was huge!!! Just played Dishonored for a little bit and it looks fantastic. FOV needed adjustment in the vorpx settings and I turned up the 3D.

Glad it is working for you too.

BTW.. I would recommend you try Dishonored using the latest version of Vireio, I'd rate it way above the VorpX experience, particularly as you can't end up looking straight ahead but with the camera pointing at your feet (thanks to VRBoost).

Anyway, that's slightly off topic.. at least we have a solution that works for now, though it seems that anytime someone updates something, a new hack is required to make the RiftUp work.

I will try Vireio. I just need the Riftup to get me to CV1...Drbeef when you get your DK2 give us a little review comparing the screen of the Riftup and DK2.

I did the comparison of RiftUp and DK2. For me RiftUp is better. The difference of pixel pattern is visible, for me too screendoor efect is more visible on DK2. FOV is +- the same or little better on RiftUp but the main problem in DK2 are the lenses - especially huge chromatic aberation, Really HUGE. Yes it can be compensated by software, but it is not fully predictable the eye position, and then it is all doing to be too complicated, too calculation of compensation of this aberation is not so simple and more power of processor or gpu is needed ( i think). The lenses cannot by changed for old ones and i think the old lenses are far better. DK2 is 50g heavier as RiftUp and it is noticeable, too cable loolks no so good. Then the synchronisation cable....2 USB.... Setup was too not so simple...i think they have a lot of work in front of them...but i think the lenses have to by changed. Interesting positional tracking was not so especially important for me.
Thx Daniel for RiftUp! I think in this time it is the best hmd for us normal peoples!

Small update after am evening of experimenting.. screen size is not an issue, the string might read DK2, but all other details remain the same as they were before. Returning DK2 from OVRService just means VorpX works and you can view the demo scene in the config utility (it is a pointless scene without positional tracking though, but does work fine.

Bioshock Infinite looks absolutely stunning on the Riftup in true stereo 3D. A far cry from the DK1 with only z-buffer 3D! If you own VorpX and BI and a riftup you must try it, it's gorgeous.

Tweaking the OVRService doesn't seem to have broken anything else either, so a bit of a result.

As soon as a get my DK2 and have taken it for a test done I'll offer some opinions,might be a couple of months off yet though.

Any experts want to do a step by step on how to edit the ovr service to recognize the RiftUP? Would love to Vorpx to work again.

Also, I can confirm that the RiftUP's display looked better to me than DK2, IMHO. I hate SDE and RiftUP's was much less noticeable to me than DK2. If it were not for the positional tracking, I would still be on the RiftUP, I believe. Come on CV1!