Underwater Sound Spatialization Model

1
Underwater Sound Spatialization Model

• Mario Michan
• University of British Columbia

2
Outline

• Sound Physics and Psychoacoustics
• Spatial Audio Systems and HRTF
• Underwater Sound and Spatialization
• Bone Conduction and STF
• Model and Results

3
Physics of Sound

• Sound are pressure waves (propagate)
• Longitudinal and Transverse (shear stresses)
• Can be modeled using the Helmholtz equation for
  sound pressure in aqueous mediums and as a
  harmonic stress in solid mediums
• We can model the propagation by coupling the
  equations
• We can assume the mediums are linear and apply
  Fourier Analysis and the Convolution Theorem

4
PsychoacousticsSpatialization of Sound

• Auditory system ability to localize sound
• Leads to Echolocation applications
• Main cues
• ITD
• ILD
• Pinna Notch
• Reverberation and echo
• HRIR HRTF
• Different for each individual (need PCA or
  something similar)
• We use the LISTEN database of 51 subjects

5
Underwater Sound

• Very poor ability to localize sound underwater
• Impedance match with head (ILD)
• Faster speed of sound 5X air speed (ITD)
• Attenuation of Higher Freq and lower threshold of
  hearing (Pinna notch)
• Middle ear blockage by air water interface
• No natural echolocation possible since HRTF
  doesnt work

6
Bone Conduction

• Sound propagate through the bones directly into
  the inner ear (Cochlea)
• Efficient way to transmit sound underwater
• We could convey sound spatialization information
  this way using the HRTF
• But sound get modified through the bones
• So we need the Skull Transfer Function
• How do we calculate the STF and ISTF?
• We model it using the equations of sound and a
  differential equation numerical solver (Finite
  Element Analysis)

7
Model

• Model the sound propagation through the skull
  using a CAD model of the skull
• Couple the stress model with the acoustic model
• Measure the response of the system at the Cochlea

8
CAD Model
Water Sphere
Simplify
Sound Source
Sound Source
9
(No Transcript)
10
Solid Medium (Deformation)
11
Acoustic Medium (SPL)
12
STF and Modified TF
13
Experiments

• Two experiments
• Above water air
• Underwater
• Results STF showed marginal improvement over
  simple HRTF
• Causes very simple model need more computational
  power
• However, very good correlation of above and
  underwater experiments good tool
• Also it is possible to spatialize sound
  underwater!!!