AhoStage: Development of an automatic manager for scenic shows
Final Master work
Bilbao School of Engineering, University of the Basque Country (UPV/EHU)
Euskadiko Ekoizle Eszenikoen Elkartea
This project aims to create an automatic system that is able to identify at each moment the point of the script in which the show is located and notify the different scene systems that control the music, lights and other effects, so that they can act as planned in their respective escalations.
This will be done using a dubbing system, AhoDubber, which uses Voice Activity Detection (VAD) with multiple input microphones and in real time, so that it can detect who is speaking at the moment. In addition, the system uses the script encoded in an ex-professo developed format to place each intervention of the actors in the context of that script and to issue the corresponding dubbed intervention. This project aims to extend its functionality to enable the application in addition to dubbing recordings, to resolve additional tasks of the technical stage staff (lights, effects, audio) and to be integrated into the set of applications that they typically use.
The following tasks will be performed:
- Review and update VAD algorithms to provide maximum robustness against noise or interference with other actors
- Definition of the recommended platform for the system, indicating both the hardware requirements and, most especially, the software control applications with which the Automatic Manager will interact, such as: QLAB, Show Cue System, MadMapper, DMX drivers, Open Stage Control
- Design and implementation of a sequencing signal generation module that will communicate to the control systems the evolution of the work on the scene.
- Participation in testing the new system in a realistic environment.
- Writing reports and user manuals for the application.