Listen: where is the sound?

Two B-format recordings by John Leonard:

A couple of Chinook helicopters
Fireworks
Sketch the sound’s location and any movement.
Compare your sketch with a neighbor.
Compare the sustained helicopter sound with the brief fireworks events.

Which directions felt clear? Where were you uncertain?

Timing: 0–6 minutes elapsed (6 minutes). Before class: download the excerpt and audition it through a B-format decoder configured for the room’s actual seven main loudspeaker positions and output routing. Confirm the source convention against the decoder. The four source channels are not speaker feeds. Use the room’s established bass management for the subwoofer; do not map a raw ambisonic component directly to LFE. A separate, filtered LFE effect is covered in the ATK setup lesson. Prepare playback before students arrive, without turning this opening into a routing tutorial. A horizontal 7.1 array does not reproduce the recording’s full height information. The official ATK catalogue identifies this as a B-format recording. The current WAV URL returned HTTP 200 on September 23, 2026; the old jlc2.amb link returned 404. Room playback still needs an audition. Prepare both excerpts through the same room decoder. The Fireworks WAV URL also returned HTTP 200 on September 23, 2026. Audition both before class and set a comfortable level for the fireworks transients. Spend one minute setting the question, two minutes playing both short excerpts and making individual sketches, two minutes comparing sketches, and one minute collecting observations. Ask whether brief events or sustained sounds were easier to locate and what listeners heard between events. Differences between seats are useful observations, not wrong answers. Do not promise a particular trajectory before auditioning. Then introduce the imagined bird in front and traffic behind as a simpler scene we can reason about throughout the lesson. These opening excerpts motivate the theory. Leave microphone comparisons to the recording lesson and file/decoder practice to the ATK lesson; choose other ATK excerpts there. Credit and catalogue: https://www.ambisonictoolkit.net/download/recordings/

What you should be able to explain

What ambisonic channels represent.
Where encoding, transformation, and decoding belong.
Why two four-channel files may need a format conversion.
What changes when we change the playback system.

The production signal path

Capture or encode a sound field.
Transform the field, such as rotating it.
Decode it for headphones or a speaker layout.

Keep an ambisonic master so you can make different playback versions.

Why separate the field from the speakers?

Developed in the 1970s by Gerzon, Craven, Fellgett, and others.
Recordings store sound-field components.
Decoders turn those components into playback channels.
One ambisonic master can serve different listening systems.

Two ways into the field

An ambisonic microphone captures a scene, including its ambience.
An encoder places a mono source at a chosen direction in the field.

We could record the whole environment or combine an ambience recording with a separately encoded bird call.

A-format: capsule signals

Four raw signals from a tetrahedral microphone.

First-order B-format: field components

W: omni. X: front/back. Y: left/right. Z: up/down.

Predict the components

Imagine one source directly in front of the microphone, at ear height.

Moving front to back: which component reverses polarity?
Moving above to below: which component changes?
Compare predictions with a partner.

A-format and B-format

	A-format	First-order B-format
Signals represent	Microphone capsules	Field components
Channels	Four for this array	Four for full-sphere FOA
Next step	Mic-specific conversion	Transform or decode

Both have four channels. Their meanings differ.

FuMa and AmbiX

Both can represent first-order B-format.

FuMa: W, X, Y, Z channel order, with FuMa scaling.
AmbiX: ACN order and SN3D scaling. At first order: W, Y, Z, X.

The file and processor must agree on order and scaling.

Rotating our scene

Rotate the whole field by a quarter turn.

Where do the bird and traffic move?
Does their separation change?
Could this move only the bird?

Decoding for the listener

A binaural decoder produces headphone signals.
A loudspeaker decoder produces feeds for a specified array.

What happens to height reproduction if all the speakers are at ear level?

Higher order: more spatial detail

Full-sphere channels: first order 4, second order 9, third order 16.

Ambience and individual sounds

A recorded field can provide surrounding ambience.
A separate bird source can move independently.

Which approach would you choose for a player-controlled vehicle, and why?

Find the problems

This workflow contains deliberate mistakes. What would you change?

Combine a raw mono bird call with four-channel AmbiX ambience.
Decode the result to binaural stereo.
Put a FuMa field-rotation processor after the decoder.
Save only the binaural stereo file, then try to use it to create a 7.1 mix.

With a partner, identify what needs changing and explain why.

Check the signal path

Available sources and processors:

Mono bird call and AmbiX ambience.
FuMa field-rotation processor.
Binaural decoder expecting FuMa.

Sketch the combined path for headphones. Mark the format conversion and where to save a master for speaker playback.

Next lessons

Ambisonics recording : arrays, recorder choices, and listening examples.
ATK in REAPER : auditioning files, routing, and rendering.