A digitally controlled two dimensional loudspeaker array
Paolo Martignon (paolo.martignon@inwind.it) Fons Adriaensen (LAE Group) Daniele Torelli Angelo Farina (LAE Group)
10/09/2008 AES 32th Conference DSP for loudspeakers
A digitally controlled two dimensional loudspeaker array Paolo - - PowerPoint PPT Presentation
A digitally controlled two dimensional loudspeaker array Paolo Martignon (paolo.martignon@inwind.it) Fons Adriaensen (LAE Group) Daniele Torelli Angelo Farina (LAE Group) AES 32th Conference 10/09/2008 DSP for loudspeakers The full staff: LAE
Paolo Martignon (paolo.martignon@inwind.it) Fons Adriaensen (LAE Group) Daniele Torelli Angelo Farina (LAE Group)
10/09/2008 AES 32th Conference DSP for loudspeakers
Project management, geometry and filter design SW development on Linux PC Consultance on SW, HW and algorithms Wiring design and realization Place, instrumentation, hardware help and consultance Chandelier iron structure realization
10/09/2008
1 2 3 10/09/2008
10/09/2008
Array Sensible parameters: The front curvature is obtained by means of a gain‐delay set ..
A λ
convex focus concave
A λ
x λ Δ
Set the beam width.. Spatial aliasing, secondary lobes.. Gains Delays Common filter r
10/09/2008
max
2 2
x
x sen c f xsen λ λ ϕ ϕ = > ⋅Δ ⇒ ≤ Δ
xf c f sen
em
Δ ≤ 2 ) ( ϑ
The spatial Nyquist theorem:
Sampling condition: Reconstruction condition: antialiasing sampling filter
10/09/2008
Omnidirectional Rigid pistons
xf c f sen
em
Δ ≤ 2 ) ( ϑ
10/09/2008
disc, though it is not
plane
testing
10/09/2008
10/09/2008
10/09/2008
1 channel
Special 32 Ohm model by Ciare, Italy
228 loudspeakers 64 channels Frounhofer distance for a group of speaker at 10 kHz is about 4 meters
10/09/2008
Absorbtion 1 kHz 2 kHz 500 Hz
A λ
convex main focus concave
10/09/2008
64 output channels
10/09/2008
Common WFS EQ and speaker EQ (mag and phase) Antialiasing sampling filter Gains and delays Antialiasing reconstruction filter Single speaker relative EQ Input FIR Array shape Dynamic delays‐gains matrix , Output FIRs Speaker directivity Output FIRs
A λ
convex main focus concave 256 taps 4096 taps
10/09/2008
Inverse filtering..
Output FIRs Inverse filtering (Ole Kirkeby)
10/09/2008
10/09/2008
“WFS focus” SW _ _ _ WFS focus SW _ _ _ Brutefir Input FIR Array shape Dynamic delays+gains Speaker directivity Output FIRs
WFS focus: By Daniele Torelli and Fons Adriaensen input FIR dynamic gains and delays
“WFS focus” BRUTEFIR By Anders Torger MADI OUT “WFS focus” “WFS focus” ….. S1 S2 … S5
64 Ch 64 Ch 64 Ch 64 Ch 64 Ch
10/09/2008
Sample rate doubling, 96 kHz Input FIR convolution, SR = 48 kHz DYNAMIC FRACTIONAL DELAY and GAINS Output FIRs 64 Ch Effect of a fractional delay with linear interpolation. Spatial information
10/09/2008
“WFS focus” “WFS focus” “WFS focus” …..
64 Ch 64 Ch 64 Ch ….. MADI inputs (external sources) Audio application, I.E. “ARDOUR” OSC MIDI controller MIDI ARDOUR “WFS focus” plugin OSC AUDIO signal SPACE control Spherical coordinates centered on main focus fractional delay and gains
10/09/2008
Mono signals
center and a peripheral point 5 m distant (center of
with people beneath the array
10/09/2008
10/09/2008