Arbitrary switch between normal and reversed stereo when switching from mono

I have Quadro 4000 on Linux 64 bit, running on a Le Novo C20x workstation. Driver version in current use 290.10

I use the stereo to look at biological macromolecules (proteins, DNA) using commercial molecular modeling software.
For reference purposes, let us call correct stereo as “front to back” and incorrect stereo as “back to front”. Due to left handedness aka chirality inherent in biological molecules, looking at incorrect stereo is worthless for my purposes.
When I switch from mono into stereo, the image does not always come up correct. As part of my work, I do have to switch between mono and stereo quite often. There is only a 50% probability that the correct stereo is displayed. Sometimes I have to flip back and forth several times or use a software switch available within the software to manually reverse the stereo to get the correct image. Have other users experienced similar reversal of stereo images (or at least arbitrary choice between correct and incorrect displays when stereo is initiated). Thanks for any answers.

Does this problem still occur with newer drivers?

We are slated to try the latest version, 310.44, released early this month, in the next few days. Back in October 2012, while trying to set up the system, we found that we had to work with an earlier version (290.10) to have the stereo operate in in our set up. (We have 3 installations, and all are running linux 64bit on LeNovo workstations (slight variations in type, and all have the Quadro 4000, and none could use the latest drivers at that time). This time, I will be more careful to note down the problem with the latest version, since I can compare it to what we have had for the last 7 months. The test of the new driver will be on a Le Novo C20X workstation.
Lakshmi Narasimhan

I have an update. We installed the newest version of the graphics driver 310.44 dated 2013.04.02.

Short reply: The performance was worse.

With this driver, the switch between normal and reverse stereo happens every 1 to 2 seconds and without much input from the user. This type of rapid reversal of correct/incorrect versions of the molecules is a no-go. Now, I remember that this is the reason we did not use the currently latest graphics driver back in Oct. 2012.
More on this problem. When this reversal happens image that is on the screen does not change when this flip happens. But one can’t be sure of that because it change could happen at such a rate that it is not perceptible to the naked eye. The emitter might be sending a signal to the stereo glasses that is out of phase with what is on the screen. In order for the correct stereo to be there all the time, the emitter should signal for “right eye shutter” to close when the “left image” is presented on the screen and vice versa (at least as I understand it). So with the new graphics drive, the emitter is frequently (frequency is 1-2 seconds) changing the phase of the signal to the emitter in relation to what is on the screen.
We have reinstalled version 290.10. All things being the same, with version 290.10, the phase-sync between emitter and glasses does not change once it is initiated. So if one starts out with the correct stereo, until stereo is turned off I can do my work. It is only when stereo is initiated (when going from mono to stereo) half of the times the phase starts out wrong. But it stays unchanged. All graphics software I use provide a software switch (usually buried down a few menu levels) that can “reverse the stereo”. If I use it to set the stereo back to correct image, then until the next time I switch back to mono, I don’t need to worry because the phase-sync is stable. What would be nice is to have a stable phase-sync, coupled with a “function button” that can be a toggle to switch the phase. Then, as soon as I see the reverse stereo, I can press the function button and I will have a stable phase-sync in correct stereo. In earlier devices I have used, (I have been doing this for over 22 years), there was a flip switch on the emitter to correct the phase.
I am hoping there are others who have wrestled with this problem and have found some solutions. (May be there is a setting in some file that needs to be reset). Would appreciate the help.
Nvidia might be able to compare version 310.44 with 290.10 and see what causes the frequent phase change. Then we would be able to use the latest drivers instead of hanging on to an old version that works.
Thanks for any help.

Lakshmi Narasimhan
Pfizer Inc.

Can you please also try 319.12, and follow the sticky posts’ instructions on how to generate and attach an nvidia-bug-report.log.gz file?

Beta version 319.12 rocks!

Here is the list of key benefits I see with 319.12:

It gives stable stereo vision. No constant flipping between correct and incorrect stereo. (see posting on version 310.44 in reply 3 above).
More robust than the stereo in version 290.10 (which is the version I was using until yesterday). The stereo in 319.12 is not susceptible to losing the correct stereo image because of interruptions (looking at a different screen, leaving the desk for sometime etc.)
I also love the fact that I don’t have to cling to an older version of the driver for keeping the stable stereo capability.



Lakshmi Narasimhan
Pfizer Inc.


I am experiencing the same problems with a Quadro 4000 card in a D30 Lenovo workstation. I have tried the latest two drivers, both of which are giving me the arbitrary eye-flipping problem (reversing stereo and making it useless, with a random chance of getting the normal or flipped stereo.

Drivers 319.17 and 319.23 both have the same issue. I am now going back, back, back to find a working driver. Apparently it is corrupted once again (?).

Hi again,

I have not heard any reply, but I am still experiencing the arbitrary reversed stereo-viewing (i.e., the left and right eye views are reversed, making stereo viewing unusable). I have now tried all the latest drivers 319.17 up to the latest 319.32 certified driver for 64-bit Linux. I am using a Quadro 4000 card in a D30 workstation, under the latest updated RedHat Enterprise Linux 6.4 OS.

Is there a solution to this issue? It is a critical problem in our workplace.

Thank you for your support.