I have seen several people asking for this and I thought it might be cool as well; extend the Flancedoor format. Instead of forcing certain points to connect in certain ways to specific other points, change it so we can connect arbitrary lines together any way we want. This would allow everything from pairs dancing, to stick fighting clones, to people drinking at the bar, to all kinds of abstract stuff and lord knows what else. I think it would be very cool and it doesn't seem that far a stretch from what we have now.

Thanks,

Umm... Someone in the know (Toby probably) will correct me if I'm wrong, but it seems to me that something like this would require re-writing the Flancedoor interpreter which is built into the Empeg visualization libraries (which are closed source). Unless we've got someone out there who can disassemble Toby's code and know enough to implement the code which actually processes the dance moves, I can't see how this could be done.

Edit: Nevermind, I just realized you are probably talking about this as a feature request rather than a 3rd party project. I guess I thought we were still in the programming forum.

Dear All,

The format of the Flancedoor files is far from perfect as many of you are already noticing. In suggesting an improved format the following should be considered:

Needs to be scalable for any number of vertices.
All vertexes should be modeled in 3 dimensions rather than two such that physical constraints such as limb length can be applied to the model which will make the task of modeling much simpler. This will also mean that the dance engine can cope with much of the movement itself (such as rotation, the dancer can then be viewed from any angle).
How are we going to sync the model to the sound? Varying the number of interpolation frames to be a multiple of the beat seperation and aligning the start point of each to the beat phase is my best idea for this problem yet.

I do intend to further develop the Flancedoor visual, so don't give up and keep the ideas coming.

Toby.

Well, as we know, synching the dancer to the sound is a problem that's been tackled before. So you wouldn't necessarily have to re-invent the wheel. But they were not particularly successful, as I recall. The WinAmp plugins that do this seemed (to me) to just animate the dancer through some predefined loops that only marginally matched the music.

You can mark key frames with a beat point attribute. Not all keyframes would qualify as a legitimate beat point.

Synchronizing the beat points should be to the nearest beat point, backwards or forwards. (Maybe)

If the beat point is too many frames away, then ignore the next beat.

I think there should be genre or otherwise tag based dance sequences. So if a Waltz comes on, versus hip hop or new age, it should behave differently.

Calvin