hrtfs can be calculated

35
HRTFs can be calculated 0 P n 2 2 0 P kP ound-hard boundaries: ound-soft boundaries: 0 P pedance boundary conditions: P i P g n merfeld radiation condition or infinite domains): lim 0 r P r ikP r elmholtz equation: Boundary conditions: 2 2 2 2 2 2 2 2 2 2 2 ' ' ' ' ' p p p p c c p t x y z Wave equation: Fourier Transform from Time to Frequency Domain '(, ,,) (,,;) i t p xyzt Pxyz e d

Upload: damian

Post on 08-Jan-2016

26 views

Category:

Documents


0 download

DESCRIPTION

HRTFs can be calculated. Wave equation:. Fourier Transform from Time to Frequency Domain. Helmholtz equation:. Boundary conditions:. Sound-hard boundaries:. Sound-soft boundaries:. Impedance boundary conditions:. Sommerfeld radiation condition (for infinite domains):. - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: HRTFs can be calculated

HRTFs can be calculated

0P

n

2 2 0P k P

Sound-hard boundaries:

Sound-soft boundaries: 0P

Impedance boundary conditions:P

i P gn

Sommerfeld radiation condition (for infinite domains): lim 0

r

Pr ikP

r

Helmholtz equation:

Boundary conditions:

2 2 2 22 2 2

2 2 2 2

' ' ' ''

p p p pc c p

t x y z

Wave equation:

Fourier Transform from Time to Frequency Domain

'( , , , ) ( , , ; ) i tp x y z t P x y z e d

Page 2: HRTFs can be calculated

HRTFs can be computed• Boundary Element Method• Obtain a mesh• Using Green’s function G

• Convert equation and b.c.s to an integral equation

• Need accurate surface meshes of individuals• Obtain these via computer vision

, ;, ;

y

yy y

p y G x y kC x p x G x y k p y d

n n

,4

ikeG

x y

x yx y

Page 3: HRTFs can be calculated

Current work: Develop Meshes

Original Kemar surface points from Dr. Yuvi Kahana,ISVR, Southampton, UK

Page 4: HRTFs can be calculated

New quadric metric for simplifying meshes with appearance attributes

Hugues HoppeMicrosoft Research

Presented by Zhihui Tang

Page 5: HRTFs can be calculated

Introduction• Several techniques have been developed for

geometrically simplify them. Relatively few techniques account for appearance attributes during simplification.

• Metric introduced by Garland and Hecbert is fast and reasonably accurate. They can deal with appearance attribute.

• In this paper, developed an improved quadric error metric for simplifying meshes with attributes.

Page 6: HRTFs can be calculated

Advantage of the new metric:

• intuitively measures error by geometric correspondence

• less storage (linear on no. of attributes)

• evaluate fast (sparse quadric matrix)

• more accurate simplifications(experiments)

Page 7: HRTFs can be calculated

What is Triangle Meshes

Vertex 1 xVertex 1 x11 y y11 z z11 Face 1 1 Face 1 1 2 32 3

Vertex 2 xVertex 2 x22 y y22 z z22 Face 2 1 Face 2 1 2 4 2 4

Vertex 3 xVertex 3 x33 y y33 z z33 Face 3 2 Face 3 2 4 54 5

………… …………

• Geometry Geometry pp R R33

• attributesattributes normalsnormals, , colorscolors, , texture coordstexture coords, ..., ...

Page 8: HRTFs can be calculated

Notation

• A triangle mesh M is described by:

V , F.

• Each vertex v in V has a geometric position pv in R3 and A set of m attribute scalars sv in Rm. That is v is in Rm+3.

Page 9: HRTFs can be calculated

Previous Quadratic Error Metrics

• Minimize sum of squared distances to planes(illustration in 2D)(illustration in 2D)

Page 10: HRTFs can be calculated

Mesh simplification

Page 11: HRTFs can be calculated

Simplification of Geometry

Qv(v) = Qv1(v)+Qv2(v)

Qf(v=(p))=(ntv+d)2=vt(nnt)v+2dntv+d2

=(A,b,c)=((nnt),(dn),d2)

Qf is stored using 10 coefficients.

Vertex position vmin minimizing Qv(v) is the solution

of Av = -b

Page 12: HRTFs can be calculated

Simplification of Geometry and Attributes

• This approach is to generalize the distances-to-plane metric in R3 to a distance-to- hyperplane in R3+m.

• Qf(v)=||v-v’||2 =||p-p’||2+||s-s’||2

• Storage requires (4+m)(5+m)/2 coefficients

Page 13: HRTFs can be calculated

New Quadric Error Metric

Page 14: HRTFs can be calculated

New Quadric Error Metric

• Qf(v)=||p-p’||2+||s-s’||2

Page 15: HRTFs can be calculated

(A,b,c) =

Page 16: HRTFs can be calculated
Page 17: HRTFs can be calculated

Storage Comparison

Page 18: HRTFs can be calculated

Experiment

Page 19: HRTFs can be calculated

Attribute DiscontinuitiesExample: a crease ,intensities.

Modeling such discontinuities needs store multiple sets of attribute values per vertex.

Wedges are very useful in this context.

Page 20: HRTFs can be calculated

Wedge

Page 21: HRTFs can be calculated

Wedge(II)

Page 22: HRTFs can be calculated

Wedge unification

Page 23: HRTFs can be calculated

Simplification Enhancements

• Memoryless simplification

• Volume preservation

Page 24: HRTFs can be calculated

Memoryless simplification

Page 25: HRTFs can be calculated

Volume preservation(I)

Page 26: HRTFs can be calculated

Volume preservation(II)

Page 27: HRTFs can be calculated

Results(I)

• Distance between two meshes M1 and M2 is obtained by sampling a collection of points from M1and measuring the distances to the closest points on M2 plus the distances of the same number of points from M2 to M1

• Statistics are reported using L2 norm and L-infinity norm

• For meshes with attributes, we also sample attributes at the same points and measure the divisions from the values linearly interpolated at the closest point on the other mesh.

Page 28: HRTFs can be calculated

Results(II)

Page 29: HRTFs can be calculated

Mesh with color

Page 30: HRTFs can be calculated

Results(IV)

Page 31: HRTFs can be calculated

Results (V)

Page 32: HRTFs can be calculated

Mesh with normals

Page 33: HRTFs can be calculated

Wedge Attributes

Page 34: HRTFs can be calculated

Radiosity solution

Page 35: HRTFs can be calculated

Results (VI)