Wave Field Synthesis

Eight timed speakers each emit a wavelet (green); summed, they rebuild the red source's true wavefront (blue) so every listener localises the same spot.

What it is

A long row of many speakers fired in precise timing to physically rebuild a sound wave so the source sits at one real spot for the whole room.

Key facts

• Speed of sound in air: 343 m/s at 20 degrees C (about 1 ms per 343 mm, ~1 ms per foot)
• Based on the Huygens-Fresnel principle: every point on a wavefront acts as a source of new wavelets; their sum rebuilds the wave
• Invented at Delft University of Technology (Netherlands) by A.J. Berkhout, 1988
• Speaker spacing sets the aliasing limit: f_alias = c / (2 x d), where c = speed of sound (343 m/s) and d = spacing between drivers in metres
• Example: 0.10 m spacing gives f_alias = 343 / (2 x 0.10) = ~1715 Hz; above that, spatial errors appear
• Real arrays use 0.10-0.25 m spacing, so hundreds of channels: full systems run 192 to 832+ discrete speakers and amp channels
• Doubling correctly-summed coherent sources adds +6 dB (20 x log10(2)); doubling random/incoherent sources adds only +3 dB
• -3 dB = half power; -6 dB = half pressure/voltage; +10 dB ~ perceived 'twice as loud'
• No sweet spot: localisation holds across the whole audience, and a 'focused source' can be placed IN FRONT of the speaker line
• Keep total DSP latency under ~10 ms; the precedence (Haas) effect fuses arrivals within ~5-35 ms toward the first sound

How it works

1. Place a dense line (or grid) of speakers along a wall or stage edge.
2. Define where each virtual source should appear in the room (x, y).
3. DSP computes the delay and gain each speaker needs so their wavefronts sum into the target wavefront.
4. All speakers fire together; the combined ripple matches what a real source at that point would emit.
5. Every listener hears the source from its true direction, no central sweet spot needed.
6. Keep driver spacing small enough that the aliasing frequency stays above the band you care about.

Real examples

• IOSONO / Fraunhofer WFS cinemas and theme-park theatres with rooms ringed by hundreds of speakers
• The Game of Life WFS array (192 speakers) used for spatial-music concerts
• Berlin TU lecture hall demo arrays with 800+ channels for research
• Permanent theatre installs where every seat must hear a voice come from the actor's actual position
• Museum and planetarium soundscapes where objects 'sound' from their physical location

How it helps in live sound

• For touring gigs, treat WFS as fixed-install tech, not a quick rig: it needs hundreds of channels and a calibrated room.
• If you can't do full WFS, fake 'true placement' with delay matrices: 1 ms per 343 mm to time-align sources to stage positions.
• Pick driver spacing for your top band: want clean to 4 kHz? Need d below ~43 mm (343 / (2 x 4000)).
• Budget DSP latency under 10 ms so lip-sync and the precedence effect stay believable.
• Use full-range, matched, closely-spaced drivers; gaps and mismatched boxes wreck the wavefront.
• For a cheaper crowd-pleaser, L-ISA-style object panning gives most of the 'source sits there' feel with far fewer boxes.