Workshop

1
Workshop  From 5.1 to Sound Field
Synthesis..." AES 120th Convention, Paris
2006Higher Order Ambisonics promises and
reality

• Jérôme Daniel, France Telecom RD

2
Traditional 1st order Ambisonics B-Format
encoding

• Panoramic sound recording
• Coincident omni (W) and bidirectional (X,Y)
  microphones
• Front-back, Left-Right separation
• Directional information
• amplitude relationships
• Description of wave propagation
• direction speed ? localization
• Independent of any loudspeaker layout

3
Reproduction over loudspeakers spatial decoding

• Simulate any coincident mic setup
• Recombine B-Format directivity patterns
• Decoding operation matrix signals W,X,Y
• One virtual microphone per loudspeaker
• ... as many as wanted, but...
• sound image blur remains the same

4
Reproduction over loudspeakers spatial decoding

• Simulate any coincident mic setup
• Recombine B-Format directivity patterns
• Decoding operation matrix signals W,X,Y
• One virtual microphone per loudspeaker
• ... as many as wanted, but...
• sound image blur remains the same
• Optimized decoding for localization

HF (gt600- 700 Hz) Concentrate energy
contributions in the expected direction (? less
altered ILD, ITD)
(LF lt 600-700z) Reproduce true wave propagation
at the listener scale (? good ITD)
minimise opposite contributions
Compromise for large area Malham
Optimize localization at the sweet spot Gerzon
5
"Traditional" 1st order Ambisonics pros cons

• Pros
• Compact multichannel format (no redundancy)
• Spatial homogeneity
• Acoustic fidelity (regarding propagation
  properties)
• Easily extended to 3D (additional Z)
• Flexibility sound field transformation
  reproduction setups
• Commercialized B-Format microphones (eg
  SoundField)
• Cons
• Blurred / unstable sound images ("tiny" sweet
  spot)
• Not well adapted to irregular/unbalanced
  loudspeaker arrangements (esp. ITU setup)
• Limitations due to low directivity of usual
  mikes, esp. at LF
• ... thats why non-coincident microphone
  approaches might be preferred

6
Introducing Higher Order Ambisonics (HOA)

• Increase angular discrimination in spatial
  encoding
• add directivities with "faster" angular variation

2nd order
3rd order
4th order
1st order
7
Introducing Higher Order Ambisonics (HOA)

• Increase angular discrimination in spatial
  encoding
• add directivities with "faster" angular variation
• Increase angular selectivity of loudspeakers
  contributions
• selective virtual microphone directivities
• better use of narrowed loudspeakers

8
Introducing Higher Order Ambisonics (HOA)

• Increase angular discrimination in spatial
  encoding
• add directivities with "faster" angular variation
• Increase angular selectivity of loudspeakers
  contributions
• selective virtual microphone directivities
• better use of narrowed loudspeakers

3rd order
4th order
1st order
2nd order
9
Rendering properties of higher spatial resolution

• Acoustic reconstruction
• Enlarged sweet area ? "Holophony" Nicol, Daniel
• Enhanced distance encoding ? control of the wave
  curvature

spherical wave (R1m) (gaussian pulse)
monochromatic plane wave (f600Hz)
1st order
5th order
10th order
2nd order

• Quality of sound images localization clues for a
  centred listener

Order M 1 2 3 4
flim 700 Hz 1300 Hz 1900 Hz 2500 Hz
aE 45 30 22.5 18
good reconstruction (?good ITD) up to flim
blur angle due to HF clues alteration
(ILDITD) above flim
10
Compatibility with irregular/unbalanced
arrangements

• Synthesize directivities adapted to ITU
  inter-loudspeaker angles
• From 4th order ambisonics Craven, 2003
• Using 5th order resolution Laborie et al
  better front channels separation
• Possible decoding criterion (among others)
  imitate pair-wise pan-pot

11
Compatibility with irregular/unbalanced
arrangements

• Synthesize directivities adapted to ITU
  inter-loudspeaker angles
• From 4th order ambisonics Craven, 2003
• Using 5th order resolution Laborie et al
  better front channels separation
• Possible decoding criterion (among others)
  imitate pair-wise pan-pot
• 4th order decoding over enriched ITU setup
  (521)
• C (0), LR(-30), SLSR(-120) LR(-70)
  B (180)
• Demonstration on a 8-loudspeaker setup (kindly
  provided by Cabasse)

? "energy vector" ( target, ie ideal sound
image)
12
Extension to 3D encoding and reproduction
Reproduction over a 3D rig
Encoding into 3D HOA Format
Spatial decoding (similar to 2D)
Sound Field Rotation

• 3D encoding and decoding
• Dynamic binaural reproduction
• Virtual loudspeakers ? doesnt sound so good
• Enhanced method better efficiency (CPU)
  rendering
• Sound field rotation driven by head-tracker
• Demo Poster session P31, Tuesday, 1400 - 1530

13
First conclusion on Higher Order Ambisonics

• Pros
• Scalable multichannel format
• Spatial homogeneity
• Acoustic fidelity "high spatial
  definition"?Wave field reconstruction
• Easily extended to 3D Efficient binaural
  spatialisation
• Even more flexibility sound field
  transformation
• reproduction setups, including irregular
  arrangements like ITU
• Cons
• nothing? ?
• What do we need in practice?
• HOA (or  high spatial resolution ) microphone
  systems
• Spatial processing tools

14
Higher Order Ambisonics Microphone Systems

• Synthesis of Spherical Harmonics
• Extension of differential microphones
• Pressure gradient and higher order derivatives
• using non-coincident acoustic sensors!
• Non concentric sensor distribution (Trinnov)
• Distribution over a rigid sphere (FT)
• Meyer, Elko, Kubli Rafaely Ward,
  Abhayapala
• Trade-off on the size of the array
• bigger is better to have spatial resolution at LF
• smaller is better to reduce spatial aliasing (at
  HF)
• A few words on FT prototype
• Designed for "proof of concept" (homogeneous 3D)
• 32 sensors ? 4th order 3D (and even 5th order 2D)
• Objective measurements validation Moreau et
  al
• ? Poster session P31, Tuesday, 1400 - 1530

15
Tools and applications

• Recording and mixing tools
• Prototypes of HOA mic (FT, Trinnov)
• Suite of VST plug-ins ? demo
• Use in common audio edition tools, or interactive
  audio progr.
• Applications
• Music, documentary, fictions
• Sharing of events/ambiances (eg familial use),
  teleconferences
• Interactive audio and multi-media
• A flexible multi-channel 3D audio format
• Games, Virtual/Mixt Reality
• News nodes for virtual scene description in MPEG4
  (AudioBIFSV3)
• label a multi-channel stream as a HOA content
  (AudioChannelConfig)
• a new kind of sound object that describes a
  Surrounding Sound Field (SurroundingSound)

16
Demonstrations

• Loudspeaker reproduction
• Reproduction of 4th order 3D recordings over
  enriched ITU setup (5 to 8 ldspk)
• Acknowledment
• Many thanks to Cabasse
• and RD manager Yvon Kernéis
• Head-tracked binaural reproduction
• Moreau et al Poster session P31, Tuesday, 1400
  - 1530
• Could also be shown after this workshop