What is GDIF
The Gesture Description Interchange Format (GDIF) is being developed as a tool for streaming and storing data of music-related movements, actions and gestures. Current general purpose formats developed within the motion capture industry and biomechanical community (e.g. C3D) focus mainly on describing low-level motion of body joints. We are more interested in describing gesture qualities, performer-instrument relationships, and movement-sound relationships in a coherent and consistent way. A common format will simplify working with different software, platforms and devices, and allow for sharing data between institutions.
SpatDIF and GDIF workshop 2010
Ircam will be hosting a two day meeting/workshop on SpatDIF and GDIF from Thursday May 20th (starting 2:00pm) to Friday May 21st (until 4:00pm). The initial announcement of the workshop can be found here.
Streaming and Storage
GDIF is intended for both streaming and storage. The following sketch shows an overview of a typical workflow:

Music-related movement is captured by sensors (or any other types of motion capture equipment) which outputs raw data. These are coded into GDIF and sent out in realtime, typically over the network using Open Sound Control (OSC). This can again be used to control a sound synthesis model in realtime. For analytical purposes, GDIF streams can also be stored as SDIF files (or possibly any other binary data storage format). For offline usage, we propose to use XML to create structured files which would simplify analysis.
Namespace for Streaming
We propose a layered OSC namespace, where it is possible to have access to multiple levels of data at any time:
![]()
For realtime purpose, it might be sufficient to use only a few of these layers, but for analytical purposes it might be interesting to have access to all.
Storage
We envisage GDIF as containing multiple streams of movement and gesture data, as well as qualitative data such as annotations and descriptors. This should also be synchronised with other media (e.g. audio and video) so that it is possible to retrieve all relevant data at any point in time.

Publications
- Marshall,M. T., N. Peters, A. R. Jensenius, J. Boissinot, M. M. Wanderley, and J. Braasch (2006). On the development of a system for gesture control of spatialization. In Proceedings of the 2006 International Computer Music Conference, 6-11 November, New Orleans. [PDF]
- Jensenius, A. R., T. Kvifte, and R. I. Godøy (2006). Towards a gesture description interchange format. In N. Schnell, F. Bevilacqua, M. Lyons, and A. Tanaka (Eds.), Proceedings of New Interfaces for Musical Expression, NIME 06, IRCAM - Centre Pompidou, Paris, France, June 4-8, pp. 176–179. Paris: IRCAM - Centre Pompidou. [PDF] [Poster]
- Kvifte, T. and A. R. Jensenius (2006). Towards a coherent terminology and model of instrument description and design. In N. Schnell, F. Bevilacqua, M. Lyons, and A. Tanaka (Eds.), Proceedings of New Interfaces for Musical Expression, NIME 06, IRCAM - Centre Pompidou, Paris, France, June 4-8, pp. 220–225. Paris: IRCAM - Centre Pompidou. [PDF]