Room Acoustics Team


WFS Production Chain

Creating a sound scene in WFS involves associating spatial information to the sound signals composing the scene. Virtual sources are used to reproduce both direct sound and room effect following a physical, geometrical or perceptual description.

The sound signals distributed on different virtual sources correspond to the soundcard outputs of a computer equipped with an audio sequencer. This implies that classical sequencer functions such as equalization or compression remain available to the end-user. Stereophonic panning is used in order to distribute the signals over Virtual Panning Spots (VPS), allowing for the creation of phantom sources within a virtual stereophonic VPS imaging area (cf Documentation). Scene description parameters are adjustable via a plug-in giving access to source position (holophonic distance, incidence angle) as well as MPEG-4 perceptual parameters. These parameters can be entirely automated, conferring the possibility of a temporal evolution in spatialization and allowing precise synchronicity with sound events.

Above: Screen shot from a ProTools session in which the Scene Description Parameters have been automated using a dedicated Plug-in.

The author of the sound scene can choose to render direct sound using WFS virtual point sources or panning over the different room effect channels. This allows for the reproduction of a large number of sources without increasing the required processing power. More pragmatically, if the "complete" WFS setup (i.e. uninterrupted loudspeaker distribution) is restricted to the front wall of the listening room, this extends the possible positions for virtual sources to the rear and side walls.

Spatialization parameters are shared and accessible by the other elements of the production chain through a distributed database on the ZsonicNet network developed by sonicEmotion . A large number of parameters are therefore made available at all locations within the installation using an Ethernet type connection. The network displays very low latency (~10 ms), allowing for a global refreshment of parameters in real time from any location within the network. ZsonicNet allows the control of distributed processes from any client inside the network. It enables a synchronous transfer of audio data to all clients and provides a consistent database of parameters. In practice, WFS rendering on a large set-up with different rendering machines can be controlled from one or more audio workstations. This means that WFS can be integrated into existing audio workstations. The audio and control data are transferred from the audio workstation to different WFS rendering machines inside the network. The network itself remains server-less and allows a dynamic configuration with changing reproduction systems.

Thus spatialization parameters can be made available on a ListenSpace interface installed on a portable PC tablet using a wireless Ethernet connection. The author can then modify spatialization parameters in real time while moving around in the sound installation.


Left: The ListenSpace interface, from which the end-user may modify the sound scene according to listening room geometry
Right: Use of a portable PC Tablet with a wireless connection to modify the sound scene in real time while moving around in the listening room

The virtual acoustics processor Spat~ developed by IRCAM has been adapted for WFS rendering. The Spat~ creates a set of room effect channels and transmits the direct sound signals associated to WFS point sources to the reproduction system.
The proposed interface gives access to scene description parameters, as well as a few basic mixing operations (level settings, routing, mute, solo…).
It also contains a multi-channel sound-file player synchronized with a MIDI sequencer. The system can therefore function without the use of the audio sequencer. Spatialization parameters are then translated into fixed MIDI controller values. The MIDI files associated with the multi-channel sound-files therefore form a complete content-coding of the sound scene.

WFS Production Chain: Synthetic view

Copyright © IRCAM Room Acoustics Team,  last modified 10/05/04