Steve, I was once working on a project similar to this, but dealing with real time point cloud data. When we wrote our tool set we originaly planed on using maya but ended up using motion builder because it had more provisions for dealing with streaming mocap data. I could see it work with a callback mechanisms like that but I think it might get hairy dealing with dg/time updates. I might checkout the mocap section in the devkit. They have an example of a mocap server that reads system time and updates a rotation based on seconds, minutes, and hours. You might be able to reuse that code to read from your stream and send it into maya as a device using the defineDataServer command. I have never worked with a mocap stream to maya implementation so all this info is part speculation.
Matt On Wed, Feb 11, 2009 at 8:48 PM, Steve Wart <[email protected]> wrote: > Sorry to reply to my own note, but I've had a look at the viewCallbackTest > example plugin > http://me.autodesk.jp/wam/maya/docs/Maya2009/API/view_callback_test_8cpp-example.htmland > it seems I should be able to do everything I need to do with the Maya > API. > > I built the example but when I install the plugin I get the error "Error: > does not match the current architecture" - presumably this is because I'm > using the 30-day trial version (at least according to this discussion thread > > https://collada.org/public_forum/viewtopic.php?f=12&t=1022&sid=51f6f06403a8a8b76f177593e8254145&start=30 > ) > > Can anyone confirm that this is in fact the case? > > Steve > > > On Wed, Feb 11, 2009 at 5:49 PM, Steve Wart <[email protected]> wrote: > >> I'm very new to Maya, but I've been reassured that what we have been asked >> to do is possible. Hopefully someone here can help point us in the right >> direction. >> >> We have an external camera which is going to be sending location, pan, >> tilt and zoom data over the network. We would like to use this data to >> manipulate a camera in Maya in real-time and render the results on the >> screen. This will be used to preview a video signal combined with the >> rendered image. >> >> There is no need to save the rendered results to a file, in fact, if that >> will slow us down, we would prefer to avoid it altogether. We are thinking >> it might be better in fact to send the video signal directly to tape, >> although we need to save the animation keyframe data so we can do a final >> render for post-production. >> >> We have some simple test harnesses in python and C++ that can read the >> location and timecode data from the network, and I've managed to move the >> camera around on the screen using another script. One problem I'm having is >> that the script process seems to be running in the same thread as the GUI, >> so it hangs until it has finished processing all commands. >> >> Is it possible to have a script running in the background that continually >> updates the camera attributes, progresses an animation according to the >> timecode data in my script, and renders the resulting image in real-time? It >> doesn't need to be a high-quality render. I've searched through heaps of >> tutorials, but there seems to be little about driving animation in >> real-time. >> >> Also, I am hoping it would be possible to texture a video frame from the >> camera within each animation frame, I suppose in an orthographic projection >> in front of the rendered model. I can get the frame data from the video >> signal, but is it possible with the C++ API to interact with the frame >> buffer in Maya during the render? >> >> Hopefully my description isn't too confusing. If you have any suggestions >> as to a high-level approach (or warnings about specific technical pitfalls), >> I'd love to hear them. >> >> Thanks, >> Steve >> > > > > > --~--~---------~--~----~------------~-------~--~----~ Yours, Maya-Python Club Team. -~----------~----~----~----~------~----~------~--~---
