forums.naturalpoint.com

TrackIR is increasingly taking on OptiTrack camera capabilities, with grayscale and all. Does TrackIR 5 do the 'final object building' itself like the V100 or is it still being done on the PC? I was just wondering if the grayscale video might put an increased load on usb bandwidth and maybe even cpu, something that has never been an issue before. Is it also possible to read in grayscale using OptiTrack?

Hello:

We added in the Precision Grayscale mode to increase the resolution of the marker tracking, similar to OptiTrack, as you point out, but with a different algorithm.

The TIR5 grayscale mode slightly increases USB bandwidth usage, but not by very much, the overall data rate is still very low. Due to our implementation, we haven't measured any increased CPU demand. Yes, the OptiTrack V100 cameras are capable of computing the marker centers in the camera, and this is important when you have 24 cameras working on the same PC.

Yes, you can read in grayscale from the OptiTrack cameras, if you mean can you read in the raw uncompressed image. You can do it at 100FPS at VGA resolution.

So the improved subpixel precision of TrackIR 5 can simply be put down to pixel brightness now being included in the weighted center calculation? Or are you doing something fancier?

Hello:

It is more complex than that, as we can also understand the edge conditions of the marker and other fun tricks. I would love to get into more detail about how it all works, as I find it quite fascinating and fun to discuss, but it is also part of the "secret sauce" so to speak. One of the challenges was how to work all the processing into the FPGA, which can be more complex than software coding.

I hope that helps some.