How to Work with Ambisonics: Workflow from Recording to Delivery

johannes.schuett@zhdk.ch (Johannes Schuett) — Fri, 15 May 2026 00:00:00 +0000

How to Work with Ambisonics

Workflow Guide · Recording · Production · Delivery · Live · ICST / ZHdK

Advanced REAPER + ICST Plugins Max/MSP optional Csound optional

🎛️ Prerequisite: Best Practices 📐 Ambisonics Formats 🔧 ICST Plugin Documentation

This page describes the complete Ambisonics workflow — from microphone placement to finished delivery or live system. It is not a beginner tutorial, but a reference for practitioners: compact, decision-oriented, and aligned directly with the ICST ecosystem.

🎙 Recording A-Format → B-Format

🎛 Production REAPER · Encoding · Mixing

📦 Delivery Export · Formats · Render

🔊 Live System Design · Real-time

Recording

Microphone SelectionA-to-B ConversionField RecordingQuality Control

Microphone Selection

Ambisonics microphones record in A-Format — the raw capsule signals, not yet a sound-field format. Conversion to B-Format (ambiX) happens afterwards, either internally in the microphone or via software.

Microphone	Order	Strengths	Typical Use
Zoom H3-VR	1st / 4 ch	Compact, affordable, internal B-Format output	Field recording, education
Sennheiser Ambeo VR	1st / 4 ch	Robust, widely used, good sound	Field recording, studio
Rode NT-SF1	1st / 4 ch	Affordable entry, solid quality	Field recording, production
DPA d:mension	1st / 4 ch	Very linear response, low coloration	Music recording, studio
Zylia ZM-1	3rd / 19 ch	Higher resolution, good directionality	Research, HOA production
EigenMike em32	4th / 32 ch	Reference HOA, maximum resolution	Studio, scientific recording

Rule of thumb: 1st-order microphones are sufficient for FOA productions and teaching. For HOA productions from 3rd order up, a Zylia ZM-1 or EigenMike is appropriate.

A-to-B Conversion

⚠ A-Format is not Ambisonics

A-Format files sound spatially incoherent when loaded directly as B-Format. Conversion must occur before any further processing.

The conversion matrix compensates for capsule spacing, frequency response, and phase errors of the specific microphone. Manufacturer-specific matrices matter — not all A-to-B tools work with all microphones.

Import A-Format raw file into REAPER (4 / 19 / 32 channels depending on microphone)
Load manufacturer-specific A-to-B plugin: Sennheiser Ambeo Orbiter, SoundField Ambisonic Toolkit, or ICST JS Plugin
Verify channel assignment: W (pressure) on channel 1, then X, Y, Z following ACN ordering
Audition a short test segment binaurally — confirm stable directionality and coherent depth
Export result as ambiX WAV and document metadata (microphone, location, take number)

Field Recording Workflow

Phase	Action
Before recording	Acoustic survey: reflections, noise sources, walkable listening zones. Level check with test tones.
Placement	Microphone height approx. 1.5 m for natural listening perspective. Avoid nearby reflective surfaces unless intentional.
Levels	Target level approx. −18 dBFS, 12 dB headroom. Monitor all raw channels separately.
Monitoring	Binaural over headphones during recording. Check for spatial coherence and wind noise.
Documentation	Immediately after each take: location, mic position, weather conditions, time of day, notable details.

Common Problems and Solutions

Problem	Cause	Solution
Channel swap, unstable directionality	Incorrect ACN ordering	Test W/X/Y/Z assignment on site with an impulse
Comb filtering, spatial artifacts	Wrong or missing A-to-B conversion	Check spectrum, use manufacturer-specific matrix
Clipping on individual capsule	Single channel overloaded	Meter all raw channels separately
Wind noise, LF rumble	Insufficient wind protection	Double windshield, LF roll-off below 80 Hz

Production & Mixing

Session SetupRoutingEncodingSpatial MixingMonitoring

Session Architecture in REAPER

A clean Ambisonics session has a clear hierarchy:

Source Tracks (Mono / Stereo / A-Format)
└→ ICST AmbiEncoder (Encoding + Panning)
└→ HOA Bus (B-Format, N channels)
├→ Decoder Track → Speaker Output
└→ B-Format Master → Export

Core rule: never bypass the HOA bus

No direct source-to-master path. Every source goes through the AmbiEncoder — otherwise the binaural render will sound different from the speaker setup.

Routing Step by Step

Create HOA Bus: channel count matching desired order (4 / 9 / 16 / 25 / 36 channels)
Insert ICST AmbiEncoder on every source track: set azimuth, elevation, spread
Route encoder output to HOA Bus — never directly to master
Insert ICST AmbiDecoder at the end of the chain: load speaker layout preset
Build a separate binaural branch: AmbiHeadphone on its own monitoring track
Prepare B-Format Master as a separate render track (without decoder)

Spatial Mixing — Parameters

Parameter	Tool	Purpose
Azimuth / Elevation	AmbiEncoder GUI or automation	Base position and movement
Spread / Width	AmbiEncoder → Width	Point source vs. diffuse cloud
Depth / Distance	Level + pre-fader reverb in HOA domain	Near/far illusion
Rotation	AmbiTransformer or OSC	Scene rotation, compass alignment
FX in HOA domain	FX plugins after encoder, before decoder	Room sound that colours the entire scene

Reverb and FX

Place reverb and room effects after the encoder and before the decoder — not on individual source tracks. This keeps room sound format-independent and ensures it is correctly exported with the B-Format master.

Typical chain for an HOA reverb:

Source Track → AmbiEncoder → [HOA FX: SN3D-compliant Reverb] → HOA Bus → Decoder

Monitoring

⚠ Never run binaural and speaker decoder simultaneously

Running a binaural decoder and a speaker decoder in parallel causes phase cancellation. Build monitoring branches as exclusive sends.

Binaural: AmbiHeadphone on a dedicated monitor track, solo routing.
Speakers: ICST AmbiDecoder with the preset for the current room.
Preset switching: Only swap the decoder — the HOA bus remains unchanged.

Choosing HOA Order

Order	Channels	Recommendation
1st	4	Binaural production, education, FOA recordings
3rd	16	Standard for HOA productions and mid-sized arrays
5th	36	Large arrays, high directional resolution
7th	64	Scientific reference, maximum resolution

Delivery & Export

B-Format MasterFormatsBinaural RenderPlatforms

Golden Rule: Always Render from the B-Format Master

Never render from the decoder output

The decoder output is speaker-specific — it only works for one particular array. The B-Format master is format- and room-independent and remains usable with any future decoder.

Export Workflow

Solo the B-Format master track and disable all other outputs
Render a short test file (approx. 10 s) and re-import it
Audition the test file through the binaural path — confirm spatial coherence
Start the full render: WAV / RF64 for large multichannel files
Document metadata in the project notes

Recommended meta-text for REAPER Project Notes:

Render: B-Format Master | Format: ambiX (ACN/SN3D) | SR: 48000 Hz | Bit: 32-bit float | Channels: 16 | HOA: 3rd | Monitoring: Binaural ✓ / Array ✓

Format Decisions

Situation	Format	Rationale
New archive master	ambiX (ACN/SN3D), multichannel WAV/RF64	Standard convention, maximum compatibility with all HOA tools
Binaural delivery (streaming, preview)	2-channel WAV, 48 kHz / 24-bit	Universally playable; document HRTF choice
Speaker stems for performance	N-channel WAV, channel count = array size	Deliver decoder preset and speaker layout alongside
Legacy system requiring FuMa	FuMa (W/X/Y/Z, MaxN normalisation)	Only when a tool explicitly requires FuMa
YouTube 360 / VR platform	Binaural stereo + spatial metadata	Check Spatial Media Metadata Tool (Google) per platform requirements

Channel Count by HOA Order

Order	Channels	Formula
1st (FOA)	4	`(1+1)²`
2nd	9	`(2+1)²`
3rd	16	`(3+1)²`
5th	36	`(5+1)²`
7th	64	`(7+1)²`

Export Checklist

B-Format master track confirmed as source
Channel count matches HOA order
Sample rate: 48 kHz
Bit depth: 32-bit float (master) or 24-bit (delivery)
Format: ambiX (ACN/SN3D) documented
Filename includes order and format
Test render before full export
Re-import and binaural check passed
Project notes updated
B-Format archive copy backed up

Live Performance & Installation

System DesignReal-time DecodingMax/MSPOSCRobustness

Core Question: Where Does Decoding Happen?

The most important decision in live HOA systems is decoder placement:

Setup	Advantages	Risks
Laptop on stage (performer decodes)	Direct control, flexible spatial responses	Single point of failure, limited latency budget
FOH system decodes	Separation of performance and control room, proven infrastructure	Communication critical, B-Format streaming required
Dedicated render machine	Most stable solution, performer independent	Synchronisation, network setup required

System Design by Context

Context	Priorities	Recommended Tools
Concert / Stage	Latency · robustness · FOH compatibility	REAPER + ICST Decoder · backup binaural mix
Club / Electronic	Real-time panning · beat sync · interaction	Max/MSP + OSC + Ableton Link
Gallery / Installation	Continuous operation · sensors · multiple zones	Max/MSP + binaural listening stations
Hybrid Studio/Live	Combine production and performance	REAPER as recorder · Max as spatialiser

Max/MSP for Live HOA

Load externals: ambiencode~, ambidecode~, ambipanning~ from the ICST Max tools
Generative panning: drunk, noise~, or cycle~ feeding azimuth and elevation inputs
OSC routing: control spatial parameters over the network (tablet, sensor, external software)
Load speaker preset into ambidecode~ and send a test tone through all positions
Stress-test for continuous operation: patch must run stably for hours without memory leaks or crashes
Prepare a backup patch: switch to binaural stereo if the array fails

OSC Control

OSC messages to the ICST AmbiEncoder enable real-time positioning from external sources — tablet, sensor, another piece of software, or a second device.

Key preparation:

Document OSC port and message namespace in the REAPER template.
Test latency: OSC over UDP is unconfirmed delivery; for critical messages, consider confirmed transmission.
Fallback: REAPER automation lanes as backup if OSC fails.

→ Full OSC syntax reference: OSC Syntax for the ICST AmbiEncoder

Irregular Arrays and AllRADecoder

For irregular speaker layouts (non-spherical, asymmetric), the AllRADecoder from the IEM Plugin Suite offers more flexibility than the ICST decoder. Workflow:

Enter speaker positions into AllRADecoder and compute the decoder matrix.
Export the matrix as a preset and load it into the live patch.
Always measure with the real setup — do not rely on simulation alone.

Quick Reference: Decision Guide

Situation	Decision
New production — which format?	ambiX (ACN/SN3D) · order determined by target format and microphone
Opening an unknown file?	Check channel count → formula (N+1)² → read ordering tag in file header
Preparing an export?	Solo B-Format master track · test render · re-import · binaural check
Reusing a FuMa file?	Convert to ambiX first, then integrate into session
Live setup — array fails?	Switch to binaural stereo backup — set it up beforehand
Monitoring sounds wrong?	Check: binaural and speaker decoder running in parallel?

Further Resources

Best PracticesSetup rules, routing, monitoring, export Ambisonics FormatsA-Format · B-Format · ambiX · FuMa · ACN/SN3D Render B-FormatExport guide for REAPER Advanced Workshop4-hour workshop: audit · recording · composition ICST Plugins DocsPlugin documentation for REAPER and Max/MSP IEM Plugin SuiteAllRADecoder · BinauralDecoder · further HOA tools Cabbage AudioCsound as VST3/AU for algorithmic HOA composition OpenAIR / SOFA-HRTFHRTF files for binaural renderers

Mixing on ICST Ambisonics