Re: [osg-users] Fwd: Unstable passive stereo on Quadro using Quadbuffer

[Jan Ciger]

On 16/11/16 01:06, ZJ Tian wrote:
> Hi, Voerman, 
> 
> Glad to hear your idea.
> 
> I previously made a mistake to assume that there is an automatic frame
> synchronization between the video card and the monitor. Now it appears not.
> 
> I think it do will work by using the monitor's SDK. 
> Thanks for your reminder.
> 

You may not need the SDK initialization (depends on how the display
handles the stereo signal), but at the very least you need the
synchronization cable. Without that it won't work properly.

J.



signature.asc
Description: OpenPGP digital signature
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org

Re: [osg-users] Fwd: Unstable passive stereo on Quadro using Quadbuffer

[ZJ Tian]

Hi, Voerman,

Glad to hear your idea.

I previously made a mistake to assume that there is an automatic frame
synchronization between the video card and the monitor. Now it appears not.

I think it do will work by using the monitor's SDK.
Thanks for your reminder.


Regards,
TianZJ
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org

Re: [osg-users] Fwd: Unstable passive stereo on Quadro using Quadbuffer

[Voerman, L.]

Hi TianZJ,
we have a zSpace screen (closely related to your HP screen), and found that
to get stereo working the screen needs a usb cable and the following calls
to the zSpace SDK must be made:

ZSError error;
// Initialize the zSpace SDK. This MUST be called before
// calling any other zSpace API.
error = zsInitialize(&_zSpaceContext);
if (error != ZS_ERROR_OKAY) {
char errorString[256];
zsGetErrorString(error, errorString, sizeof(errorString));
// MessageBox(NULL, errorString, "ZSError", MB_OK | MB_SETFOREGROUND |
MB_TOPMOST);
_zSpaceContext = NULL;
}
ZSHandleg_bufferHandle = NULL;
error = zsCreateStereoBuffer(_zSpaceContext, ZS_RENDERER_QUAD_BUFFER_GL, 0,
_bufferHandle);
if (g_bufferHandle) {
zsSyncStereoBuffer(g_bufferHandle);
zsDestroyStereoBuffer(g_bufferHandle);
}
//I do this at program shutdown, but suspect this can be done at once if
you're not useing the headtracking/pen tracking.
// Shutdown and cleanup the zSpace SDK.
zsShutdown(_zSpaceContext);
_zSpaceContext = NULL;

I don't know if HP has created it's own SDK or refers you directly to
zSpace for the software, but I think you need a similar sequence.
btw. The signal is send as frame sequential stereo - commonly referred to
as "active stereo signal" trough a single video cable.
Referring to this system as passive stereo just because the glasses are
passive confuses a lot of people. (the display switches circular
polarization at 120 Hz)

Regards, Laurens.

On Tue, Nov 15, 2016 at 2:24 PM, Jan Ciger  wrote:

>
>
> On Tue, Nov 15, 2016 at 11:45 AM, ZJ Tian  wrote:
>
>> Hi, Jan,
>>
>> You lightened my way. That's the answer I think.
>> I will try to work it out though you guessed correctly the display device
>> I am unluckily using.
>>
>
>
> You are welcome.
>
> Sadly, these problems are common. The hardware vendors will claim that
> their gizmo is perfect, does everything out of the box and, of course, plug
> and play. Only when you actually buy it you will discover that the sales
> materials neglected to mention that the function you need requires a very
> expensive accessory sold extra and made out of unobtanium (because nobody
> stocks it - like the mini-DIN connector brackets for Quadros) or that the
> function actually works, but it is severely limited to the point of being
> unusable because some fool somewhere decided to save $0.1 by eliminating a
> connector or shipping non-working firmware ...
>
> I have had these experiences even on very expensive pieces of hardware.
> Especially with VR you do have to always test before buying because the
> sales brochures and salesmen are never telling the whole story if not lying
> outright.
>
> J.
>
>
> ___
> osg-users mailing list
> osg-users@lists.openscenegraph.org
> http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
>
>
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org

Re: [osg-users] Fwd: Unstable passive stereo on Quadro using Quadbuffer

[Jan Ciger]

On Tue, Nov 15, 2016 at 11:45 AM, ZJ Tian  wrote:

> Hi, Jan,
>
> You lightened my way. That's the answer I think.
> I will try to work it out though you guessed correctly the display device
> I am unluckily using.
>


You are welcome.

Sadly, these problems are common. The hardware vendors will claim that
their gizmo is perfect, does everything out of the box and, of course, plug
and play. Only when you actually buy it you will discover that the sales
materials neglected to mention that the function you need requires a very
expensive accessory sold extra and made out of unobtanium (because nobody
stocks it - like the mini-DIN connector brackets for Quadros) or that the
function actually works, but it is severely limited to the point of being
unusable because some fool somewhere decided to save $0.1 by eliminating a
connector or shipping non-working firmware ...

I have had these experiences even on very expensive pieces of hardware.
Especially with VR you do have to always test before buying because the
sales brochures and salesmen are never telling the whole story if not lying
outright.

J.
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org

Re: [osg-users] Fwd: Unstable passive stereo on Quadro using Quadbuffer

[ZJ Tian]

Hi, Jan,

You lightened my way. That's the answer I think.
I will try to work it out though you guessed correctly the display device I
am unluckily using.

Thanks.



- - -
TianZJ



在 2016-11-15 18:14:59，"Jan Ciger"  写道：



On Tue, Nov 15, 2016 at 10:40 AM, ZJ Tian  wrote:

> Hi, guys,
>
> I encountered a trouble with passive stereo using quadbuffer.
>


Duh, I am an idiot, should have read more carefully.

If your system uses passive glasses (the non shuttering type) then the
problem is most likely in the setup of your display. If the display is
using sequential ("active") stereo from the Quadro (that's what I assume
since you are using quadbuffer) and converting that to passive, do check
for the presence of the synchronization signal. If you aren't using HDMI
1.4 (or newer) connection, you may have to connect an extra mini-DIN or a
USB cable somewhere. Some Quadro cards require buying of an extra bracket
with a connector for this (like this one:
https://www.bhphotovideo.com/c/product/838382-REG/PNY_Technologies_930_50764__000_Full_Height_Bracket_Stereo.html
 )

If it is this display:
http://h20195.www2.hp.com/v2/GetDocument.aspx?docname=c04784690=quickspecs=EN_US==us=en

Then you may be actually out of luck. The DVI-D and DP 1.2 ports don't
support the synchronization signals required, AFAIK. You may want to
contact HP support about it and verify with them.

J.
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org

Re: [osg-users] Fwd: Unstable passive stereo on Quadro using Quadbuffer

[Jan Ciger]

On Tue, Nov 15, 2016 at 10:40 AM, ZJ Tian  wrote:

> Hi, guys,
>
> I encountered a trouble with passive stereo using quadbuffer.
>


Duh, I am an idiot, should have read more carefully.

If your system uses passive glasses (the non shuttering type) then the
problem is most likely in the setup of your display. If the display is
using sequential ("active") stereo from the Quadro (that's what I assume
since you are using quadbuffer) and converting that to passive, do check
for the presence of the synchronization signal. If you aren't using HDMI
1.4 (or newer) connection, you may have to connect an extra mini-DIN or a
USB cable somewhere. Some Quadro cards require buying of an extra bracket
with a connector for this (like this one:
https://www.bhphotovideo.com/c/product/838382-REG/PNY_Technologies_930_50764__000_Full_Height_Bracket_Stereo.html
)

If it is this display:
http://h20195.www2.hp.com/v2/GetDocument.aspx?docname=c04784690=quickspecs=EN_US==us=en

Then you may be actually out of luck. The DVI-D and DP 1.2 ports don't
support the synchronization signals required, AFAIK. You may want to
contact HP support about it and verify with them.

J.
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org

Re: [osg-users] Fwd: Unstable passive stereo on Quadro using Quadbuffer

[Jan Ciger]

Hello,

On Tue, Nov 15, 2016 at 10:40 AM, ZJ Tian  wrote:

> Hi, guys,
>
> I encountered a trouble with passive stereo using quadbuffer.
>

That sounds like your display and glasses are not being correctly
synchronized to the graphic card. For stereo to work in a stable manner the
glasses must know which frame is being rendered by the GPU at any given
moment. That is usually achieved either by the old school 3 pin mini-DIN
cable from the Quadro, synchronization signal being sent using a USB dongle
(like Nvidia 3DVision) or the signal being digitally "tagged" with the
left/right frame tags (recent HDMI/Display Port interfaces versions).

Some systems are "dumb" and don't use this sync signal - typically cheap
DLPLink projectors - and rely only on the stability of the input signal to
not swap the left/right frames while in use. The glasses are only told to
switch between left/right but don't really know which image is being
actually displayed by the GPU, so you have to manually "invert" the stereo
at the start, typically by pushing a button somewhere. That is good enough
for something like Bluray video but not for interactive application where
the framerate is never 100% stable.

I am not familiar with the display and glasses you are using so you will
need to check the manual for your hardware how to set it up correctly.
However, in general, simply enabling quad buffer stereo *is not* sufficient
in itself to obtain a reliable stereo without these glitches. The
synchronization between the GPU and the display/glasses must be in place as
well. Be careful that synchronizing the glasses to the display alone (using
radio, infrared signal, display flashes ...) may not be sufficient if the
display is not actually synchronized with the GPU - that's the case of the
common DLPLink systems, but may exist elsewhere too. If the glasses are not
connected to the GPU directly and are driven from the display (common with
3D TVs and projectors), I would expect that this could well be the case.

Regards,

Jan
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org

[osg-users] Fwd: Unstable passive stereo on Quadro using Quadbuffer

[ZJ Tian]

Hi, guys,

I encountered a trouble with passive stereo using quadbuffer.

The 3D effect is not stable
sometimes I can get a correct 3D vision, but other times it seems that  the
left frame and the right frame had been reversed, which can be corrected by
reversing my polaroid-glass. The result randomly varied every time when
restarted the app.;

Here is the environment I am using:
Hardware
GraphicCard  Nvidia Quadro M2000
DisplayHP Zvr Virual Reality Display
Glass   Polaroid glass

Configuration
Display Hz   120Hz
GraphicCardStereo Enabled
   Stereo Type Common Stereo
   3D Display Enabled

Software Config
osg::DisplaySetting  stereo enabled
 stereo type QUAD_BUFFER
GUI  QT5.7


I am confused, since it should NOT have happened: The OSG knows the left
buffer and right buffer, and it should send them to graphic card correctly.
And the card should not mistake these two buffers when sending them to the
display device either. But then which is the problem?

I tried google but found nothing relative.

Anybody who has confronted this trouble?
Thanks for any hints.




- - -
TianZJ
tianz...@gmail.com
2016/11/15
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org