← Back to the help

Audio outputs, and what to adapt

The same map does not sound the same on headphones, on a loudspeaker setup or in binaural. This guide tells you what carries over unchanged, what has to be set differently, and the one case where what you drew is no longer used at all.

Menus are named here as the editor shows them in English.

Read this first: in EarMotion, fading is not a distance

In a game engine, a sound fades with a distance in meters. Here, it fades with your position inside the zone you drew: at the center the sound is full, at the edge it is zero — truly zero, not “very quiet”.

What follows from this, before anything else: two zones of different sizes produce exactly the same fade, only traveled in more or less time. Making a zone bigger does not make the sound more distant: it makes the path longer. This is what makes fading something you write, not a simulation of physics.

The four outputs at a glance

The output is chosen in the editor: left column › Output › Audio output › Output mode. It is saved with the map.

Output Is the zone you drew used? What decides what you hear What for
Headphones — stereo yes The zone sets the volume; the left/right position sets the panning. The lightest. The reference while you work.
Headphones — binaural yes Like stereo for the volume, plus a position in the space around the head. Headphones, mobile, app.
Speakers — multichannel
(one speaker per output; 5.1 included)
no The zone of the loudspeaker nearest to the source. See the box below. An installation where each loudspeaker has its own output.
Speakers — Envelope yes The zone sets the volume; the sound is then spread over all the loudspeakers by the Envelope law: its position as seen from the listening cursor is carried into the room, and the loudspeakers nearest to that point carry most of it (a source outside the setup is brought back to its edge). Envelope is EarMotion's own spatialisation: loudspeakers are placed freely, in 2D as in 3D, and each source keeps the same level wherever it is, with no gap or jump from one loudspeaker to the next. With The sound turns with the listener (checked by default), the walking direction is rendered on the wall at the top of the map; unchecked, nothing turns and the top of the map stays the front. Immersive setup, room.

Direct multichannel ignores your zones — and it is the only case

In this mode, each source goes to the loudspeaker nearest to it, and what you hear depends on your position within that loudspeaker's territory — not within the zone you drew around the source. What counts is your position, or that of the visitors seen by the cameras. Outside that territory: silence. One exception: a source set as Non-interactive source plays continuously on its loudspeaker, whether someone is there or not.

In other words, a map written on headphones and then switched to direct multichannel no longer plays your writing: it plays the geometry of your setup. This is not a fault — it is what “one loudspeaker per output” means — but you need to know it before you go into the room.

What to do: if you care about your drawing, choose Envelope, which keeps your zones and spreads the sound over the whole setup. Direct multichannel is tuned by moving the loudspeakers, not by redrawing the zones. To see these territories, turn on Speaker zones in Visualization (right column).

What carries over unchanged

What to adapt, output by output

Stereo the reference

This is the simplest output, and the quickest to write for. The panning follows the direction of the listening cursor: a source straight ahead or straight behind it sounds in the middle, equally on both sides. Stereo does not tell front from back. If you want listeners to sense where a sound comes from, move it to one side.

Binaural check on headphones

Front and back become audible, which stereo could not do. Height too, with the Ambisonic engine, the browser's default (Audio output › Advanced settings › Binaural engine) and always the iPad app's. The Dome engine, the browser's fallback or a choice of the device, renders everything at ear level. Sources that you told apart by volume alone may end up in the same perceived place. Listen, and separate them in space rather than in level.

Changing the output during playback restarts the sources when you enter or leave binaural or Envelope: the sound cuts for a moment and the files start again from the beginning. This is normal: the output path is not the same.

5.1, 7.1 and the other standards layouts, not modes

The Speaker preset of the Speaker table offers the standard layouts: mono, 2.1, LCR, LCRS, 5.0, 5.1, 6.1 and 7.1 (ITU-R BS.775 and BS.2051), Dolby Atmos 5.1.2, 5.1.4, 7.1.2, 7.1.4 and 9.1.6, Auro-3D 9.1, 10.1, 11.1 and 13.1, NHK 22.2; for acousmatic music, octophony with a front pair (in ring order, or in pairs as at BEAST), hexaphony and the ambisonic cube. Each loudspeaker is placed at the standard azimuth, on the channel the standard gives it. Choosing a preset places it at once; ↶ Restore brings back the previous layout as long as you have not changed it.

These layouts play in multichannel: everything said above about direct multichannel applies — and if you want to keep your writing by zones, use Envelope on the same positions. The height layers (Atmos, Auro-3D, 22.2) and the cube are rather a matter for Envelope: multichannel divides the map seen from above, without height.

The low-frequency effects channel (LFE) has no loudspeaker on the map: its channel stays reserved (channel 3, output 4; in 2.1, channel 2, output 3; in 22.2, outputs 4 and 10), which keeps the wiring of a standard interface, and + Speaker does not take it. Zones are not spread to it; only a multichannel source played with Trigger only (without release) can spill into it, its channels going to the outputs that follow its loudspeaker's. EarMotion produces no bass channel: let the bass management of your interface or room processor feed the subwoofer. The older Dolby Atmos 7.1.4, LFE speaker at the centre (12 speakers) preset still places an "LFE" loudspeaker at the centre of the scene: in multichannel it takes the central area there; prefer 7.1.4 — Dolby Atmos / ITU-R BS.2051 J.

Atmos not produced by a browser

The Dolby Atmos presets of the loudspeaker table place the loudspeakers at the positions of an Atmos room: they are layouts, not an Atmos stream.

EarMotion cannot produce an Atmos stream, and no web page can: it takes a licensed encoder, or the spatial layer of the operating system. Nobody will receive Atmos from a browser, whatever the map.

What is possible, and useful: work as you would for an Atmos render — objects that move freely rather than a mix fixed to channels — then listen to the result in binaural on headphones, or on whatever setup you have. The placement carries over; the encoding does not.

Worth knowing too: Atmos defines no fading law. Distance, the Doppler effect and occlusion remain the job of the tool that produces the sound — here, EarMotion. What you hear of distance therefore comes from your writing, not from the format.

Depending on the device

Steps to follow, if you must deliver on several outputs

  1. Write on headphones, in stereo: that is where zones and fades are set the fastest.
  2. Switch to binaural and listen again: separate what has merged.
  3. For a room, first choose Envelope or direct multichannel — this is not an output setting, it is a writing choice. Envelope keeps your zones; direct multichannel replaces them with the geometry of the setup.
  4. Place the loudspeakers (Audio output › Speaker table), tune them if needed (Output › Speaker calibration), then listen again while walking: it is the only check that counts.
  5. Do not promise Atmos. Promise binaural, which can be heard anywhere.

See also section 4 of the full guide (listening and spatialization) for the detailed settings of each mode.