Processing method and system for panning audio objects
The current invention related to methods and systems for panning audio objects on multichannel loudspeaker setups. The invention relates to a method of processing an audio object along an axis, said audio object comprising an audio object abscissa and an audio object spread, for spatialized restitution thereof over a plurality of sound transducers, N in number, aligned along said axis; each of said sound transducers comprising a transducer abscissa; N being at least equal to two; said method comprising a plurality of steps.
1. A method of processing an audio object along an axis, said audio object ( 151 ) comprising an audio object abscissa and an audio object spread, for spatialized restitution thereof over a plurality of sound transducers, N in number, aligned along said axis; each of said sound transducers comprising a transducer abscissa ( 152 ); N being at least equal to two; said method comprising the steps of:
executing a first process ( 110 ) comprising a mapping of the transducer abscissa ( 152 ) of each of said plurality of sound transducers and of the audio object abscissa ( 151 ) on a circle quadrant, yielding N transducer angles ( 154 ) for said plurality of transducers and one audio object angle ( 153 ), θ s , for said audio object;
executing a third process ( 130 ) comprising the substeps of:
( 132 ) computing an effective number of transducers ( 159 ), β i for each transducer i of the plurality of transducers via
β
i
=
∑
j
=
1
N
1
(
2
)
u
(
1
+
cos
(
θ
i
-
θ
j
)
)
u
u
∈
[
0
,
∞
]
,
i
∈
[
1
…
N
]
,
wherein θ i is the transducer angle for the transducer i, θ j is the transducer angle for the transducer j, and u denotes spread and
( 133 ) computing a transducer gain P i ( 160 ) for each of said plurality of transducers, i∈[1 . . . N], via
P
i
(
θ
is
)
=
1
(
2
)
u
(
1
+
cos
(
θ
is
)
)
u
,
u
∈
[
0
,
∞
]
,
i
∈
[
1
…
N
]
,
where θ is is the angle between the audio object and the transducer i; and
executing a fourth process ( 140 ) comprising the substeps of:
( 142 ) computing an initial gain value G i ( 163 ) for each of said plurality of transducers, N in number, by dividing said transducer gain ( 162 ) by said effective number of transducers ( 159 )
G
i
(
θ
s
)
=
P
i
(
θ
is
)
β
i
,
i
∈
[
1
…
N
]
,
and
( 143 ) ensuring power conservation by computing a total emitted power, P e , via P e (θ s )=Σ i=1 N (G i (θ s )) 2 and computing, for each of said plurality of transducers, N in number, a corrected gain ( 164 ), A i , via
A
i
=
G
i
P
e
,
i
∈
[
1
…
N
]
;
wherein:
said method further comprises executing a second process ( 120 ) comprising the substeps of:
( 122 ) identifying, from the plurality of transducers, a first transducer α ( 155 ) and a second transducer β ( 156 ) that are closest to the audio object, and
( 123 ) computing gains Q α ( 157 ) and Q β ( 158 ) according to a stereo panning law over said first transducer α ( 155 ) and said second transducer β ( 156 );
said third process ( 130 ) further comprises:
an additional substep of ( 131 ) creating a virtual transducer comprising a virtual transducer angle equal to said audio object angle ( 153 ) and adding said virtual transducer angle to a list of transducers angles ( 154 ), N in number, thereby creating an expanded list of transducer angles, N+1 in number; and
a gain substep ( 133 ) of computing said transducer gain, said gain substep ( 133 ) further comprising computing a virtual transducer gain P N+1 ( 161 ) corresponding to said virtual transducer angle, via:
P
i
(
θ
is
)
=
1
(
2
)
u
(
1
+
cos
(
θ
is
)
)
u
,
u
∈
[
0
,
∞
]
,
i
=
N
+
1
,
P
i
(
θ
is
)
=
1
(
2
)
u
(
1
+
cos
(
θ
is
)
)
u
,
u
∈
[
0
,
∞
]
,
i
∈
[
1
…
N
+
1
]
,
where θ N+1,s is the angle between the audio object and the virtual transducer; and
said fourth process ( 140 ) further comprises:
an additional substep of ( 141 ) redistributing said virtual transducer gain P N+1 ( 161 ) over said first transducer α ( 155 ) and said second transducer β ( 156 ) by using said gains Q α ( 157 ) and Q β ( 158 ) computed in the second process ( 120 ), yielding a modified gain P′ α ( 162 ) for said first transducer α ( 155 ) and a modified gain P′ β ( 162 ) for said second transducer β ( 156 ) according to
P
i
′
=
P
i
2
+
u
1
+
u
·
Q
i
2
,
where i=α or i=β;
wherein said computing of said initial gain value G i ( 163 ) is done with said modified gain P′ α ( 162 ) instead of said gain P α for said first transducer α ( 155 ) and said modified gain P′ β ( 162 ) instead of said gain P β for said second transducer β ( 156 ).
2. Method according to claim 1 , wherein said stereo panning law is any or any combination of the following: tangent panning law, sin-cos panning law.
3. A method of processing an audio object, for spatialized restitution thereof over a plurality of sound transducers, N in number, positioned on an inner surface of a parallelepipedic room comprising a ceiling, a front wall and a lateral wall; N being at least equal to two, said sound transducers positioned according to an XYZ orthonormal frame comprising an X axis, a Y axis and a Z axis, whereby said Z axis extends toward and is orthogonal to said ceiling, the Y axis extends toward and is orthogonal to said front wall and the X axis extends toward and is orthogonal to said lateral wall, wherein each of said transducers and said audio object comprise Cartesian coordinates ( 200 ) with respect to said XYZ orthonormal frame for an abscissa; wherein said audio object comprises a spread value with respect to said XYZ orthonormal frame, wherein said method comprises the steps:
in a first step ( 201 ), obtaining a Z-gain ( 207 ) for each of said plurality of transducers, using only the Z abscissae of said plurality of transducers and the Z spread value,
in a second step ( 202 ), determining a unique Z coordinates list for a transducer arrangement, effectively constructing Z-layers,
in a third step ( 203 ), obtaining Y-gains ( 208 ) for each of said plurality of transducers and for each of said Z-layers, using only said Z-layer's transducers' Y abscissae and the Y spread value,
in a fourth step ( 204 ), determining, for each said Z-layer, unique Y coordinates list, effectively constructing Y rows,
in a fifth step ( 205 ), obtaining X-gains ( 209 ) for each of said plurality of transducers, for each Z layer and for each Y row, using only the rows' transducers' X abscissae and the X spread value, and
in a sixth step ( 206 ), multiplying said X-gains ( 209 ), Y-gains ( 208 ) and Z-gains ( 207 ) element-wise, and applying 2-norm normalization to obtain final transducer gains ( 210 ) for the whole transducer arrangement,
wherein:
said determining of said Z-gain ( 207 ) in the first step ( 201 ) is performed with the method according to claim 1 along the Z-axis,
said determining of said Y-gain ( 207 ) in the third step ( 203 ) is performed with the method according to claim 1 along the Y-axis, and
said determining of said X-gain ( 207 ) in the fifth step ( 205 ) is performed with the method according to claim 1 along the X-axis.
4. A method of processing an audio object, for spatialized restitution thereof over a plurality of transducers, N in number, positioned on an inner surface of a sphere, N being at least equal to two; said audio object comprising an audio object position and an audio object spread; said method comprising the steps of:
executing a first process ( 301 ) comprising the substeps of:
(pre)computing the effective number of transducers β i based on the plurality of transducers, where i represents the number of transducers, the audio object position and the audio object spread, and
modifying β i by an affine function between 1 and the original value of β i , yielding modified effective number of transducers ( 313 ); and
executing a second process, for given object coordinates, comprising
a first step ( 302 ) that computes Vector-Based Amplitude Panning (VBAP) gains for each facet in the mesh and finds enclosing facet for which each of the transducer gains Q i are positive, and discards the other gains, yielding three VBAP gains ( 314 ),
a second step ( 303 ) that creates a virtual transducer in the transducer arrangement, positioned at the object position ( 311 ), so that the modified arrangement comprises N+1 transducers,
a third step ( 304 ) that computes original Speaker Placement Correction Amplitude Panning (SPCAP gains) ( 315 ) for the N+1 transducers,
a fourth step ( 305 ) that redistributes the computed gain for the virtual (N+1)-th transducer by using the three VBAP gains Q i ( 312 ) computed above in the above first step ( 302 ) and the original SPCAP gains ( 315 ), yielding N modified SPCAP gains ( 316 ),
a fifth step ( 306 ) that computes the initial gain values G i ( 317 ) by dividing the original SPCAP gains ( 316 ) by the modified effective number of transducers ( 313 ) as precomputed by the first system above
G
i
(
θ
s
)
=
P
i
(
θ
is
)
β
i
,
i
∈
[
1
…
N
]
,
wherein θ i is the transducer angle for the transducer i, where θ is is the angle between the audio object and the transducer i, and
a sixth step ( 307 ) that ensures power conservation by computing the total emitted power, P e , via P e (θ s )=Σ i=1 N (G i (θ s )) 2 and by dividing the initial gains values ( 317 ) to yield the corrected gains ( 318 ), A i , for each transducer:
A
i
=
G
i
P
e
,
i
∈
[
1
…
N
]
,
wherein:
the computation of said effective number of transducers ( 313 ) uses the following formula:
β
i
=
∑
j
=
1
N
1
(
2
+
1
u
)
u
(
1
+
1
u
+
cos
(
θ
i
-
θ
j
)
)
u
,
u
∈
[
0
,
∞
]
,
i
∈
[
1
…
N
]
β
i
=
1
(
2
+
1
u
)
u
(
1
+
1
u
+
cos
(
θ
i
-
θ
j
)
)
u
,
u
∈
[
0
,
∞
]
,
i
∈
[
1
…
N
]
;
the third step ( 304 ) of the second process uses the following formula:
P
i
(
θ
is
)
=
1
(
2
+
1
u
)
u
(
1
+
1
u
+
cos
(
θ
is
)
)
u
,
u
∈
[
0
,
∞
]
,
i
∈
[
1
…
N
+
1
]
wherein θ is is the angle between the source and the transducer and u denotes spread; and
the fourth step ( 305 ) of the second process uses the following formula:
P
i
=
P
i
2
+
u
1
+
u
·
Q
i
2
,
i such that speaker i belongs to the active VBAP facet.