Interactive low delay music and speech communication via network connections (OVBOX)

Abstract

The “OVBOX” is a tool for low-delay network audio communication and generic data transmission between multiple clients. Acoustic end-to-end delays of about 30 ms can be achieved, assuming a good internet connection and a distance between clients of less than about 1500 km. For optimal spatial perception, e.g., when using multiple instruments of similar timbre, an individual 3-dimensional room acoustic simulation based on physical modelling is applied in each client. The system is optimised for headless operation using a dedicated single-board computer (Raspberry Pi 4B), but desktop clients are also available for better integration with other audio software. A client-server system allows remote configuration and automatic traversal of network address translation routers and firewalls. With the low latency that can be achieved, the “OVBOX” is used for music applications such as distributed rehearsals or concerts. Other applications include hearing research to achieve interactive speech communication with low delay transmission of head movements for real-time control of virtual reality, and transmission of other biophysical data for online analysis or central data logging. The tool is fully open source.