Gabriel, yeah, at first I thought (hoped!) it could be triple buffering, but after some pretty extensive checks (searching the windows registry, driver settings, etc), it looks like that isn't the problem. Also, the 1 frame latency is consistent across different boards with different drivers on different computers. It would be unlikely that they all had triple buffering enabled. Andrew, as for scene complexity, yeah we've been keeping it pretty simple with just a single stimulus (plus a large bar underneath for accurately setting the background colour) in a single viewport. Otherwise, at 200Hz we start to drop frames. I recall starting to drop frames when trying to play back two movies (two textures) at the same time (your suggestions for that by the way are still on my to-do list, along with a bunch others...). On the other hand, our Nvidia card is a not exactly new Geforce 4MX, but for our stimuli it has the same lag as our new Radeon 9800. Perhaps the drivers have improved in terms of latency since you last tested? Was your latency a constant 10 ms or did you notice it varying? When I get a chance I'll try turning off sync to swap, I guess the latency will become more gradiated and will correlate better with the complexity of the scene. Cheers, Martin Spacek On Mon, March 7, 2005 10:37 am, Andrew Straw said: > Dear Gabriel -- there is no triple buffering going on, so I'm pretty > sure this is not the issue. > > Dear Martin -- IIRC, I was actually getting a 10 msec delay on somewhat > older graphics cards (nVidia GeForce 4 Ti). So if this is a "pipeline > latency" perhaps it is dependent on the complexity of the scene (a bit > higher for my stimuli, I think) and the speed of the graphics card. > > If you turn off "sync swaps to vsync", is the latency reduced? (Yes, I > know this causes other problems, but perhaps they could be largely > worked around, especially if you have feedback from a DAQ card about > when your frames are actually being drawn.) > > I wonder if DirectX would have this same issue. I'd be surprised if it > didn't, but the driver code paths may be substantially different, so I > suppose it's a possibility. > > This is really the only issue that's holding up submission of a methods > manuscript on the Vision Egg. I keep wanting to do "just one more > calibration" to get to the bottom of this issue, but I simply haven't > had time without it being my #1 priority. Anyhow, it's good to keep the > issue on my radar screen -- I hope I'll get back to it myself before too > long. > > Cheers! > Andrew ====================================== The Vision Egg mailing list Archives: //www.freelists.org/archives/visionegg Website: http://www.visionegg.org/mailinglist.html