This forum is monitored by iZ3D Support Staff and Stereoscopic Veterans.
However, please keep in mind that for iZ3D to find a solution, we need as much information about the issue as possible.
You may also fill out an iZ3D Support Ticket here.
Joined: 07 Oct 2005 Posts: 3573 Location: Airplane
Posted: September 15, 2008, 01:59:31 PM Post subject:
Hi, Chris!
It is a bit later here - 2AM , but let me talk with our developers tomorrow about specs changing - I think it is possible, but need to double check. We also want to try another idea how to hack thread - it may request some performance drop, but would be fine for powerful cards.
Posted: September 15, 2008, 03:55:33 PM Post subject:
BlackQ wrote:
Hi, Chris!
It is a bit later here - 2AM , but let me talk with our developers tomorrow about specs changing - I think it is possible, but need to double check. We also want to try another idea how to hack thread - it may request some performance drop, but would be fine for powerful cards.
Wow, sounds good! Thanks Vadim.. In terms of the performance of this driver on higher-end cards, if there were ever a reason to upgrade this is it.
Posted: September 15, 2008, 10:59:51 PM Post subject:
Hugo wrote:
I think itґs not a real good solution to have the stereo-output in dependence to game-fps.
With shutter glasses, i need at least 100 Hz. With titles like Crysis and perhaps all newer games, it would be very hard to get such high values.
Obviously yes that is a problem, but you have come right out there and stated crisis which is the biggest FPS killer of them all. Many people are using shutters with DLP projectors which run a maximum of 85hz anyway, and from the tests I have done with a GTX280 this kind of framerate is easily acheivable with the vast majority of newer games at XGA resolution. The marker spec is already there, it would just require a little tweaking to enable it to be used with the Stereographics Stereo Enabler (for example) to give a shutter mode that actually works pending a fuller solution.
Ultimately, this would still be no good for things like watching 3d dvd's and movies, as you're stuck with the glasses refresh at whatever the video is recorded at... headaches and nausea all round - so hopefully BlackQ's investigation into the render hack will yield a positive result!
Joined: 07 Oct 2005 Posts: 3573 Location: Airplane
Posted: September 16, 2008, 06:43:34 AM Post subject:
Hi, Chris!
We can change xml specs - it is possible, but will take time. Could you, please, explain why such specific marking may hep more than existing one - I did not understand you well. This explanation will help us to understand the best way how to do it
Posted: September 16, 2008, 08:20:23 AM Post subject:
BlackQ wrote:
Hi, Chris!
We can change xml specs - it is possible, but will take time. Could you, please, explain why such specific marking may hep more than existing one - I did not understand you well. This explanation will help us to understand the best way how to do it
Hi Vadim,
The purpose of me proving the example was for this reason, to clarify what was needed. For instance, from a post above byTril it appears your xml marking spec only allows a mark to be specified starting at the top line. The Blue Line Code of stereographics is on the bottom line, and for other solutions it may be elsewhere in each frame. In terms of the blue line code (which the cheap Stereographics Stereo Enabler uses) I am confirming the exact spec with them but I think it is something like:-
Left eye: left 3/4 of bottom line is blue..
Right eye: right 3/4 of bottom line is blue.
If you use the marking spec example I have shown above, you can specify any pixel (or group of pixels) on any line to be any color, and have mutiple lines of different colours if you wanted. Typically, the left eye frame will have one set of markings, then the right eye a different set. Then you just loop LRLR obviously.
I have also included in the example a different group for each possible screen resolution - as obvioulsy the marker lines would be longer for higher resolutions, and not being the full line length a differnt number of pixels would need to be specified according to resolution. The driver can then determine which set of markers to use according to the resolution.
Does this make sense? Its pretty simple stuff really if you think about it.
I'm a software engineer and you could easily modify your XML parser to do this in less than a day - you can easily determine the current screen resolution too to pick which set of markings to use.
However, this might turn out to be a wasted excersize if you think you may have another solution to explore re: interrupting the rendering thread as you say above. Is this something you still need to look into or is it a dead end? I'm curious (as I expect everyone else is) how this new idea has come about - did you consult an expert?
They all seem to have excellent knowledge about nVidia and/or ATI drivers
Another v good suggestion - these guys have been working low-level with these cards for years so if anyone could give a definitive answer it'd no doubt be them. What do you think BlackQ? Also, was it clear re: the marking specs now?
Joined: 07 Oct 2005 Posts: 3573 Location: Airplane
Posted: September 17, 2008, 09:31:07 AM Post subject:
Hi, Chris!
Regarding XML changes: we can try to do it only on Monday. The guy who can do it has another urgent task at the moment.
We tried interruption today and yesterday - no success - we have 10-15 fps after that .... unfortunately looks like interruption blocks render process also ...
Posted: September 17, 2008, 02:17:56 PM Post subject:
BlackQ wrote:
Hi, Chris!
Regarding XML changes: we can try to do it only on Monday. The guy who can do it has another urgent task at the moment.
We tried interruption today and yesterday - no success - we have 10-15 fps after that .... unfortunately looks like interruption blocks render process also ...
OK Vadim thanks - if you could go ahead and do this next week that'd be good - I'll confirm it is working and purchase the stereo enabler dongle to perform some tests.
I still think this really needs to be addressed at the lower level but this is a good stopgap while you continue research, hopefully some of those other guys can help.
Posted: September 19, 2008, 03:43:48 PM Post subject:
BlackQ wrote:
Hi, Chris!
BLC is added to release candidate already. Rest part will be analyzed on Monday as promised.
Fantastic Vadim! You mean BLC will be in the next release right? Did you get the exact specs on this then? If you can analyze this Monday that would be great, I've been conducting more tests on the lastest games the last couple of days and I think with high end cards at xga resolution there are very few modern games that fall under 45fps (stereo), making this a viable solution for 85hz projectors at this time. I've got a Stereographics Enabler on order in the meantime which should hopefully be delivered in the week, and will be able to extensively test VESA glasses such as edim and revalator with these games and post my results imminently.
Thanks again for the continued attention to this important aspect of yor driver, I've been extremely impressed with the general performance thusfar and it'd be great if we can find a happy medium to satisfy the needs of SG users pending a fuller solution (not to mention this addition would abstract from the need to have h/w pageflipping buffers).
As an aside, did you say these drivers have not tested successfully under SLI as yet? In particular I'd like to try with a pair of GTX 280's to see what is possible, but wont bother investing in the extra card if this is not a viable option.
Joined: 07 Oct 2005 Posts: 3573 Location: Airplane
Posted: September 20, 2008, 01:44:09 PM Post subject:
Hi, Chris!
This is the same BLC specs like other used before. for last line pixels from 0 to X are blue for Left and from X to Y are blue for right (sorry I forgot X and Y value).
Yes, BLC will come with next release (Monday or Wednesday).
SLI has to be fine - only iZ3D has problem with SLI because of one output - shutter should be fine - you can easily check it with anaglyph - SLI impacts to core only - not to output
Posted: September 21, 2008, 09:17:49 AM Post subject:
BlackQ wrote:
Hi, Chris!
This is the same BLC specs like other used before. for last line pixels from 0 to X are blue for Left and from X to Y are blue for right (sorry I forgot X and Y value).
Yes, BLC will come with next release (Monday or Wednesday).
SLI has to be fine - only iZ3D has problem with SLI because of one output - shutter should be fine - you can easily check it with anaglyph - SLI impacts to core only - not to output
, but let me talk with our developers tomorrow about specs changing - I think it is possible, but need to double check. We also want to try another idea how to hack thread - it may request some performance drop, but would be fine for powerful cards.
