Determination of the adequate measurement window for sound source localization in echoic environments
A method for sound source localization comprises the steps of computing an envelope of a signal; performing a non-linear, onset preserving smoothing of the envelope; and evaluating cues for sound source localization, using the points where the smoothing changes from its rising phase to its falling phase as measurement points.
1. A method for sound source localization, comprising the steps of:
receiving an aural signal by at least one microphone;
computing an envelope of said aural signal;
performing a non-linear, onset preserving smoothing of said envelopes to generate a smooth envelope, the smooth envelope set to a value of the envelope
except where the envelope changes from a rising phase to a falling phase, the smooth envelope is set to the value of the envelope multiplied by an inhibition factor, and
except where a previous value of the smooth envelope is greater than the envelope, the smooth envelope is set to decay from the previous value based on a time constant;
evaluating cues for sound source localization, using measurement points from the smooth envelope, where measurement points are where said smoothing envelope changes from a rising phase to a falling phase; and
localizing a point of origin of the aural signal based upon said cues.
2. The method of claim 1 , wherein said smoothing of the envelope is performed by a function defined by the following term:
x
s
(
k
)
=
{
0
k
=
0
x
(
k
)
x
s
(
k
-
1
)
≤
x
(
k
)
⋀
k
>
0
x
(
k
)
·
ϑ
x
s
(
k
-
1
)
>
x
(
k
)
⋀
x
s
(
k
-
1
)
≤
x
(
k
-
1
)
⋀
k
>
0
(
1
-
1
/
τ
)
·
x
s
(
k
-
1
)
+
1
/
τ
·
x
(
k
)
x
s
(
k
-
1
)
>
x
(
k
)
⋀
x
s
(
k
-
1
)
>
x
(
k
-
1
)
⋀
k
>
0
wherein τ represents the time constant and θ represents the inhibition factor.
3. The method of claim 1 where said time constant (τ) and said inhibition factor (θ) are determined adaptively based on measurements of room characteristics.
4. The method of claim 1 , wherein said smoothing is calculated in the log domain.
5. The method of claim 1 , wherein said cues for sound source localization include at least one of a Time Delay of Arrival (TDOA), an Interaural Intensity Difference (IID) or an Interaural Envelope Difference (IED).
6. The method of claim 1 , further comprising the step of forming a measurement window based on the measurement points.
7. The method of claim 6 , wherein a length of said measurement window is fixed.
8. The method of claim 6 , wherein a length of said measurement window is determined by said smooth envelope.
9. The method of claim 1 , wherein said aural signal is a band-pass signal.
10. The method of claim 1 , wherein said envelope is the envelope of the envelope modulations of unresolved harmonics.
11. The method of claim 1 , wherein said smoothing is performed by a first order infinite impulse response (IIR) filter.
12. A computer software program product, embodied on a non transitory computer readable medium for performing the method of claim 1 when run on a computing device.
13. A computer program stored in a non transitory computer readable medium for performing the method of claim 1 when run on a computing device.