So I had a little success with the point grey cameras using libdc1394 version 2, only I’ve hit a few roadblocks and hoped this would be the place to ask for advice.

I can get the 6 images from the ladybug but because it’s something like 4 million pixels it’s painfully slow. As I also mentioned in another post because the image is tilted 90 degrees the bayer filters that come with libdc don’t work.
– Is there any way to optimize the process of reading an image from memory?
– What about a very efficient way of turning the image 90deg with in the jitter SDK before applying the bayer filter?

I’ve used the same principles to craft jit.bumblebee, but I haven’t been able to get it to output a matrix. I think the problem might be that the pixels are RAW 16bit interleaved, but I’m not sure if Jitter can handle that the 16bpp? Or maybe is there a way to deinterleave the stereo pair in Max so I don’t have to do that from scratch before I can even test it?

The last thing I can’t get working properly is the open and close messages. How do I initialize the camera on the open message if the init has to happen in the jit class?

Thanks.