Most spatial audio balances loudspeaker levels to suggest where a sound sits — an illusion that resolves at one seat and breaks the moment you move. Holophonic audio reconstructs the sound field itself: a real wavefront, stable as you cross the room, and shared by everyone in it. A field, not a sweet spot.
The loudspeaker boundary divides two spaces: the physical room, where bodies stand and pressure exists, and the acousmatic space beyond the loudspeakers — above the line, or all around the ring — where sources are virtual. Drag the red source anywhere in the acousmatic space; drag the listener in the room. Each source radiates the equivalent of 90 dB at 1 m; the SPL field and the dB grid are drawn in the physical room only — there is no dB in the acousmatic space, because it is not a physical space. In WFS the field is correctly rebuilt only inside the wedge the array subtends from the source — the valid area; beyond it the front simply runs out, which is why a longer array buys a larger shared zone. Drag the WFS source through the array to focus it inside the room. Dashed lines run from each listener to each source; on non-holophonic systems a red line shows where the source is actually heard. Double-click a marker to remove it.
Stereo, surround, Dolby Atmos, ambisonics — different sizes of one idea. Each balances the levels sent to a set of loudspeakers so that a sound seems to sit somewhere between them. The geometry is solved for a single position, the sweet spot, and for a single listener fixed in it.
Stand there and the phantom holds. Step aside and the distances to each speaker change, the level distribution reshapes, and the timing cues diverge; the image drifts and then collapses toward the nearer speaker. These systems are, in the end, personalized — they reproduce what one listener would hear at one point. Everyone else in the room is handed a degraded copy of someone else's seat.
Holophony changes both the aim and the method. Instead of balancing gains to suggest a position, it specifies a desired wave field and builds it in the air — driving many small emitters as if they were elements of one continuous boundary. What reaches the room is not a signal tailored to one head but the sound waves themselves, carrying the correct pressure at each point in space. Localization stays put as you move, and it is the same for everyone at once.
The word shares its root with hologram: holophony is the acoustic counterpart to holography.
The term has carried other meanings. Hugo Zuccarelli coined “holophonic” in the 1980s for a binaural recording method, and some researchers extend it to binaural and higher-order ambisonics. But those remain personalized: they move the space around a single listener rather than fixing it in the room. They reproduce what one point would hear — holophony reconstructs the field.
It does not always take a large array. A single point-source is already holophonic: a voice through a drive-thru speaker is grounded in a place, and everyone in the line localizes it to the same spot. Janet Cardiff's The Forty Part Motet turns forty such points into a choir you can walk among; Cevdet Erek's Room of Rhythms lets the architecture itself fix a single speaker in place. What they share is placeness — sound that anchors to a position rather than diffusing into abstraction, drawing listeners into one shared spatial relation.
The question shifts — from does it sound realistic? to does it hold a shared acoustic horizon as bodies move?
WFS is the most developed way to build a holophonic field. It rests on a principle Christiaan Huygens set down in 1690: around every point of a wavefront, a new spherical wavelet forms, and the envelope of all those wavelets is the wavefront an instant later. Read the other way, it is a recipe — line up enough elementary sources along a boundary, each emitting at the right moment and strength, and their sum is the wave of a source that isn't there.
The physics makes this exact. A field inside a closed surface is fully determined by the pressure and velocity on that surface — by a layer of secondary sources covering it. WFS takes the practical reductions of that result: collapse the surface to a single layer, then to a horizontal line of loudspeakers at ear height. What's left is a row of drivers, each fed the same sound with its own delay and amplitude, together rebuilding the moving wavefront of a virtual source — placed behind the array, or focused in front of it, standing in the room with you.
There is one hard limit, and it is the one that governs all sampling. Spacing the drivers samples the wavefront in space; above a frequency set by that spacing the array can no longer represent the wave, and artifacts appear — spatial aliasing, the acoustic echo of Nyquist. At 12.5 cm between drivers the clean band reaches roughly 1.4 kHz, near where the ear shifts localization from timing to level, so the seam stays largely beneath notice. Closer drivers buy bandwidth; a longer array buys more room before the front runs out at its edges.
Reconstructing a field at room scale takes a dense line of emitters. The AURA is one such instrument — a 192-channel wave field synthesis array built to put holophony into practice, where these ideas stop being diagrams and become a space you can stand inside. There is more on its design and construction at theaura.keremergener.com.
Below, the recipe at work: a virtual source behind a line of secondary sources, each a stone dropped on cue, rebuilding the curved front of a source that exists only behind the line — and a wavelength control to watch the front shatter when the spacing undersamples it.
Drag the red virtual source behind the line of secondary sources, then synthesize: each stone drops on a delay set by its distance to the virtual source — nearest first, farthest last — and the wavelets sum into one curved front, centred on a source that was never in the water. The red arc marks the intended front. Plane wave drops every stone at once. Then shorten the wavelength: the stones sample the front in space, and once λ falls below their spacing the wavelets stop merging — the front shatters into grating lobes that no longer follow the red arc. Spatial aliasing, in water.
A recording captures the sound of a place but not its atmosphere, because atmosphere is not only heard — it is lived, through the body, in the place where the body stands. Played back through a personalized system, sound arrives in a different plane than the one your body occupies: the acousmatic space behind the speakers stays a picture, optimized for one seat.
When the field is real instead of implied, that picture becomes an extension of the room. Sources anchor to positions you can walk around; parallax holds; timbre shifts continuously as you move. Localization stops belonging to one privileged listener and belongs to everyone at once.
The acousmatic space stops being a separate entity and starts working as part of the place — a shared horizon a crowd can inhabit together, with their bodies. Atmosphere becomes something you can put in a room, and something a room can hold in common.
Holophony does not ask the field to look right from one seat. It asks the field to be there — for every body in the room.