[dullard]

Hello,

I have a puzzler for you...

I have an intel iMac and a G5 (both brand new). Both have external displays. I would like copies of what is shown on those two displays to be superimposed on top of each other (using something like a chroma key), and that composite signal to be added to an audio track and recorded as DV or MPEG on a hard drive. And all this has to happen in real time.

Here's my solution. The DVI out from each computer is converted to VGA using Apple adaptors. That signal is split. One signal goes to the external display, the other is converted from VGA to s-video. The s-video signals (from both computers) go into a video mixing board (like this one). The mixing board uses a chroma key to superimpose the two signals, and outputs to s-video. That analog signal is converted into firewire DV using a converter. The firewire goes back into the G5 and iMovie records it.

I *think* that will work, but please let me know if you see any problems. It does, however, seem like a bit of a hackneyed solution, since that poor video signal goes through 5 different conversions/manipulations.

Why am I doing all this? I'm building an eye tracking lab. The iMac is showing stuff to people on that external display. The G5 is calculating their gaze position and creating a crosshair that moves with their eyes. I want to put together what the subject sees with where their eyes are into one composite image.

Any help would be very much appreciated!

[Uncle Skeleton]

Why does it have to be merged in real time? Do the subjects need to see the output of the whole crosshair thing? Because if not, you could use a screencapture app like snapz pro to record both computers and merge them afterwards (and you don't even need to record the crosshair's screen, only its coordinates, then render it afterwards). And unless the imac's output is interactive, I don't see why you'd need to record it, instead of just using the original source again later on...

In short, no I don't think you'll find an all-digital capture program that includes features like superimposing two sources. But you could merge the two sources twice, once for the real-time requirement (no need to convert back to digital, right?) and again later with just the coordinates from the G5's software, to try to save some video quality.

[dullard]

Originally Posted by Uncle Skeleton

Why does it have to be merged in real time? Do the subjects need to see the output of the whole crosshair thing? .

The subjects don't, but the experimenter does. The superimposed image allows the experimenter to check that the cross hairs line up with the stimuli, and recalibrate the camera if not.

Someone told me recently that they had used a mixing board that allowed you to fade between to VGA inputs (so that you could switch a projector between two presenters' laptops). They said that I could leaving the fading half way, and so superimpose the two images. They couldn't remember anything else about the name of such a box, however, and I can't google them....

[nerd]

Sounds like it would work but if you leave iMovie up so the video window is visible you'll probably get some video feedback. I would use a 3rd Mac for recording the mix.