I'd like to control two separate projectors playing separate pre-
recorded video in synchrony (dual head). Don't know which elements
are best to achieve this. Interested to know the components which I
should use as there are a few choices.
I've prototyped some simple video control in GEM, on an OpenGL 3D
surface as described at http://www.agnula.org/documentation/
dp_tutorials/pd_gem/ and the rendering and control of frame speed is
perfect. However for the real thing I'll need dual head rendering of
two video files, and presumably rendering the frame sequence to an
OpenGL pixmap on a texture mapped surface is unnecessary and itself
could create some limitations for dual-heading.
THE ALTERNATIVES
The basic questions are...
Video rendering : Gridflow vs. GEM vs. X. Serial comms : comport vs. serialctl vs. X.
Both dual head rendering and cross-platform serial communications
seem to be contentious issues, so I think it might be useful to have
ideas from those with experience of either or both of these issues,
and can recommend the libraries most likely to work for both my
prototyping and deployment environment. I hope to work out the
details myself, once I have an idea which libraries should work in
principle.
Also if anyone has suggestions as to the strategy I should use to get
serial control data in from a Basic Stamp microprocessor, and map two
output video streams to two screens simultaneously, this would be
very useful.
THE PROJECT
I want to emulate the front and rear window of a car, (footage
recorded from two DV cameras in a real car), on opposing projection
screens. The reason I'm keen on using PD is because I need to
control the speed of the video in real time as if controlling the
speed of the car's accelerator, from zero mph to full video speed and
above. Tried this with processing and it was very glitchy.
I'm prototyping on Mac OS X but will deploy probably on an intel
machine running debian.
Welcome your contributions.