I have been tasked with programming a GUI for the Adaptive Use Musical Instrument project (AUMI). The GUI is a musical creation tool for people with limited mobility. The current GUI takes data from a video camera, displays it on screen, allows the user to select a point to track (using openCV) and uses this data to cue midi notes/audio samples.
A workable implementation currently exists in MaxMSP, but it has become slow, difficult to manage, buggy and cross-incompatible over time (and size). I will be re-implementing it in a different framework, though I have not yet decided on which. Processing, OpenFrameworks, Juce and Qt have all been brought up as possibilities.
Requirements:
- Final GUI must work on Windows and Mac.
- Computer vision should run as smoothly as possible, even on low end machines (we're just doing point tracking now, though hopefully fancier stuff in the future).
- The GUI must require basically NO technical expertise from its users, this is used in schools mostly. I realize this depends a lot on design, but I also don't want a build with a million files that the user needs to navigate.
Thoughts? Basically I'm trying to find the best balance of programming time (I don't have a ton) and final functionality. MaxMSP was great to get a prototype up and running, but we need something more scalable. It would be really nice to start in the right place.