40 th AES Conference
Tokyo, 8-10 October 2010
A Spherical Microphone Array For
Synthesizing Virtual Directive Microphones In Live Broadcasting And In Post Production
Angelo Farina, Andrea Capra, Lorenzo Chiesi, Leonardo Scopece
Strategie Tecnologiche
Centro Ricerche e Innovazione Tecnologica Academic Spinoff University of Parma
RAI RAI - - state of art state of art
• The Holophone H2 Pro is a microphone system equipped with 8 capsules placed on a egg‐shaped framework. The audio signals are delivered directly in G‐format or using an audio mixer.
• The directivity of each
Holophone capsule was
measured in the Anechoic
Room of “Università’ di
Ferrara” (Italy)
250 Hz
‐30
‐25
‐20
‐15
‐10
‐5 0 1
2 3 4 5 6 7
8 9 10
11 12
13 14
15 16
17 18 19 20 21 22 23 24 25 26 27 28 29 3130 3332 3534 36 37 3938 4140 4342 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58
59 60
61 62
63 64
6566676869 7071 72
C CS L R LS RS Top
500 Hz
‐30
‐25
‐20
‐15
‐10
‐5 0 1
2 3 4 5 6 7
89 10
11 12
13 14
15 16
17 18 19 20 21 22 23 24 25 26 27 28 29 3130 3332 3534 37 36 3938 4140 4342 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59
60 61
62 63
646566676869 7071 72
C CS L R LS RS Top
1000 Hz
‐30
‐25
‐20
‐15
‐10
‐5 0
1 2 3 4 5 6
7 8 9
10 11
12 13
14 15
16 17 18 19 20 21 22 23 24 25 26 27 28 29 3130 3332 3534 37 36 3938 4140 4342 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59
60 61
62 63
646566676869 7071 72
C CS L R LS RS Top
2000 Hz
‐30
‐25
‐20
‐15
‐10
‐5 0 1
2 3 4 5 6 78
9 10
11 12
13 14
15 16
17 18 19 20 21 22 23 24 25 26 27 28 29 30 3231 3433 3635 37 3938 4140 4342 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58
59 60
61 62
63 64
6566676869 7071 72
C CS L R LS RS Top
4000 Hz
‐30
‐25
‐20
‐15
‐10
‐5 0
1 2 3 4 5
6 7 8 9
10 11
12 13
14 15
16 17 18 19 20 21 22 23 24 25 26 27 28 29 3130 3332 3534 37 36 3938 4140 4342 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59
60 61
62 63
646566676869 7071 72
C CS L R LS RS Top
8000 Hz
‐30
‐25
‐20
‐15
‐10
‐5 0 1
2 3 4 5 6 7
89 10
11 12
13 14
15 16
17 18 19 20 21 22 23 24 25 26 27 28 29 3130 3332 3534 37 36 3938 4140 4342 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59
60 61
62 63
646566676869 7071 72
C CS L R LS RS Top
Holophone
Holophone polar patterns polar patterns
• Directivity and angles between single capsules are not changeable in post‐processing.
• There isn’t enough separation between sources because of the low directivity of the capsules.
For this reason the probe should be placed very close to the scene that is object of recordings
• Surround imaging is in any case inaccurate, albeit the recording sound spacious and with good frequency response (thanks to the DPA omnis)
HOLOPHONE PROBLEMS:
HOLOPHONE PROBLEMS:
The The EIGENMIKE EIGENMIKE TM TM
Array with 32 ½” capsules of good quality, frequency response up to 20 kHz
Preamplifiers and A/D converters inside the sphere, with ethernet interface
Processing on the PC thanks to a VST plugin (no GUI)
The The EIGENMIKE EIGENMIKE TM TM
The Eigenmike is designed for High Order Ambisonics processing
His software computes internally the spherical harmonics up to third order
Up to 8 virtual microphones can be synthesized, with one of 7 pre-defined patterns
The The EIGENMIKE EIGENMIKE TM TM
The RAI project The RAI project
• “Virtual” microphones with high directivity, controlled by mouse/joystick in order to follow in realtime actors on the stage. They should be capable to modify their directivity in a sort of
“acoustical zoom”.
• Surround recordings with microphones that can be modified (directivity, angle, gain, ecc..) in post‐production.
GOALS:
GOALS GOALS
We want to synthesize virtual
microphones highly directive,
steerable, and with variable
directivity pattern
MICROPHONE ARRAYS:
MICROPHONE ARRAYS: TYPES AND PROCESSING TYPES AND PROCESSING
Linear Array Planar Array Spherical Array
Processing Algorithm
SIGNAL PROCESSOR x
1(t)
x
2(t)
x
M(t) M inputs
y
1(t) y
2(t)
y
V(t)
... ... V outputs
.. . x
1(t)
x
2(t)
x
M(t)
h
1,1(t) h
2,1(t)
h
M,1(t)
y
1(t)
General Approach General Approach
• Whatever theory or method is chosen, we always start with M microphones, providing M signals x m , and we derive from them V signals y v
• And, in any case, each of these V outputs can be expressed by:
processor M inputs
V outputs
M
m
v m m
v t x t h t
y
1
, ( )
) ( )
(
Microphone
Microphone arrays: target, processing arrays: target, processing
Is the time‐domain sampled waveform attempting to emulate that of a virtual microphone characterized by a certain polar pattern and aiming (our target)
Pre‐calculated filters
x 1 (t) x 2 (t) x 3 (t) x 4 (t)
y v (t)
M
m
v m m
v t x t h t
y
1
, ( ) )
( )
(
Traditional approaches Traditional approaches
• The processing filters h m,v are usually
computed following one of several, complex mathematical theories, based on the solution of the wave equation (often under certaing simplifications), and assuming that the
microphones are ideal and identical
• In some implementations, the signal of each
microphone is processed through a digital
filter for compensating its deviation, at the
expense of heavier computational load
Traditional Spherical Harmonics approach Traditional Spherical Harmonics approach
Spherical Harmonics (H.O.Ambisonics)
Virtual microphones
A fixed number of “intermediate” virtual microphones is computed,
then the dynamically-positioned virtual microphones are obtained
by linear combination of these intermediate signals.
Novel approach Novel approach
• No theory is assumed: the set of h m,v filters are derived directly from a set of impulse response measurements, designed according to a least‐squares principle.
• In practice, a matrix of impulse responses is measured, and the matrix has to be numerically inverted (usually employing some regularization technique).
• This way, the outputs of the microphone array are maximally close to the ideal responses prescribed
• This method also inherently corrects for transducer deviations and acoustical artifacts (shielding,
diffractions, reflections, etc.)
The microphone array impulse responses c m,d , are measured for a number of D incoming directions.
We get a matrix C of measured impulse responses for a large number P of directions
m=1…M mikes d=1…D sources
c ki
Novel approach Novel approach
D , M d
, M 2
, M 1
, M
D , m d
, m 2
, m 1
, m
D , 2 d
, 2 2
, 2 1
, 2
D , 1 d
, 1 2
, 1 1
, 1
c ...
c ...
c c
...
...
...
...
...
...
c ...
c ...
c c
...
...
...
...
...
...
c ...
c ...
c c
c ...
c ...
c c
C
The virtual microphone which we want to synthesize must be specified in the same D directions where the impulse responses had been measured. Let’s choose a high‐order cardioid of order n as our target virtual
microphone.
This is just a direction‐dependent gain.
The theoretical impulse response
coming from each of the D directions is:
Target Directivity Target Directivity
n
Q n , 0 . 5 0 . 5 cos( ) cos( )
d ,
d Q
p
Novel approach Novel approach
m = 1…M microphones
d = 1…D directions
Applying the filter matrix H to the measured impulse responses C, the system should behave as a virtual microphone with wanted directivity
h
1(t) h
2(t)
h
M(t)
p d (t)
Target function
c 1,d (t)
AM,v
δ(t)
A1,v δ(t)
A2,v δ(t)
M
m
d m
d
m h p d D
c
1
, 1 ..
But in practice the result of the filtering will never be exactly equal to the prescribed functions p d …..
c 2,d (t)
c M,d (t)
Novel approach Novel approach
We go now to frequency domain, where convolution becomes simple multiplication at every frequency k, by taking an N-point FFT of all those impulse responses:
We now try to invert this linear equation system at every frequency k, and for every virtual microphone v:
0 .. / 2
..
1
1
, ,
, k N
D P d
H C
M
m
d k
m k
d m
k DxV DxM
k MxV
C H P
This over-determined system doesn't admit an exact solution, but it is possible
to find an approximated solution with the Least Squares method
Least
Least - - squares solution squares solution
We compare the results of the numerical inversion with the theoretical response of our target microphones for all the D directions, properly delayed, and sum the
squared deviations for defining a total error:
The inversion of this matrix system is now performed adding a regularization parameter , in such a way to minimize the total error (Nelson/Kirkeby
approach):
It revealed to be advantageous to employ a frequency-dependent regularization parameter k .
P
k MxD k DxM k MxM
k j MxD DxV
k k MxV
I C
C
e Q
H C
*
*
Spectral shape of the regularization Spectral shape of the regularization
parameter parameter k k
• At very low and very high frequencies it is advisable to increase the value of .
H
L
It is possible to compute just once the following term:
Then, whenever a new set filters is required, this is generated simply applying to R the gains Q of the target microphone:
FIR filters realtime synthesis algorithm:
Thanks to Hermitian symmetry properties, a real-FFT algorithm can be employed
Real Real - - time synthesis of the filters h time synthesis of the filters h
k MxD
k DxM k
MxMk j k MxD
k MxD
I C
C
e R C
*
*