Acoustic-environment mismatch and proximity detection with a novel set of acoustic relative features and adaptive filtering
A method of performing distance estimation between a first recording device at a first location and a second recording device at a second location includes: estimating acoustic relative transfer function (RTF) between the first recording device and the second recording device for a sound signal, e.g., by applying an improved proportionate normalized least mean square (IPNLMS) filter; and estimating the distance between the first recording device and the second recording device based on the RTF. The at least one acoustic feature extracted from the RTF estimated between the first recording device and the second recording device includes at least one of clarity index, direct-to-reverberant ratio (DRR), and reverberation time. A distributed-gradient-boosting algorithm with regression trees is used in combination with signal-to-reverberation ratio (SRR) and the at least one acoustic feature extracted from the RTF to estimate the distance between the first recording device and the second recording device.
1. A method of performing distance estimation between a first recording device at a first location and a second recording device at a second location, comprising:
estimating an acoustic relative transfer function (RTF) between the first recording device and the second recording device for a sound signal; and
estimating a distance between the first recording device and the second recording device based on the acoustic RTF.
2. The method according to claim 1 , wherein:
the first recording device is a fixed recording device at the first location, and the second recording device is a movable recording device.
3. The method according to claim 2 , wherein:
an improved proportionate normalized least mean square (IPNLMS) filter is applied for an estimation of the acoustic RTF; and
the distance between the first recording device and the second recording device is estimated based on at least one acoustic feature extracted from the acoustic RTF estimated between the first recording device and the second recording device.
4. The method according to claim 3 , wherein:
the first recording device is part of a smart home speaker, and the second recording device is part of one of a wearable device or a mobile device.
5. The method according to claim 3 , wherein:
the at least one acoustic feature extracted from the acoustic RTF estimated between the first recording device and the second recording device includes at least one of clarity index, direct-to-reverberant ratio (DRR), sparseness, and reverberation time.
6. The method according to claim 5 , wherein:
a distributed-gradient-boosting algorithm with regression trees is used in combination with signal-to-reverberation ratio (SRR) and the at least one acoustic feature extracted from the acoustic RTF to estimate the distance between the first recording device and the second recording device.
7. The method according to claim 2 , wherein:
an improved proportionate normalized least mean square (IPNLMS) filter is applied for the estimation of the acoustic RTF, and
the distance between the first recording device and the second recording device is estimated based on a set of acoustic features extracted from the acoustic RTF estimated between the first recording device and the second recording device, the set of acoustic features including clarity index, direct-to-reverberant ratio (DRR), and reverberation time.
8. The method according to claim 7 , wherein:
the first recording device is part of a smart home speaker, and the second recording device is part of one of a wearable device or a mobile device.
9. The method according to claim 8 , wherein:
a distributed-gradient-boosting algorithm with regression trees is used in combination with signal-to-reverberation ratio (SRR) and the set of acoustic features extracted from the acoustic RTF to estimate the distance between the first recording device and the second recording device.
10. The method according to claim 9 , wherein the sound signal is one of i) white shaped noise, and ii) speech signal.
11. A system for performing distance estimation between a first recording device at a first location and a second recording device at a second location, comprising:
an estimation module for estimation of an acoustic relative transfer function (RTF) between the first recording device and the second recording device for a sound signal; and
a distance estimation module configured to estimate a distance between the first recording device and the second recording device based on the acoustic RTF.
12. The system according to claim 11 , wherein:
the first recording device is a fixed recording device at the first location, and the second recording device is a movable recording device.
13. The system according to claim 12 , wherein:
an improved proportionate normalized least mean square (IPNLMS) filter is applied by the estimation module for the estimation of the acoustic RTF; and
the distance between the first recording device and the second recording device is estimated based on at least one acoustic feature extracted from the acoustic RTF estimated between the first recording device and the second recording device.
14. The system according to claim 13 , wherein:
the first recording device is part of a smart home speaker, and the second recording device is part of one of a wearable device or a mobile device.
15. The system according to claim 14 , wherein:
the at least one acoustic feature extracted from the acoustic RTF estimated between the first recording device and the second recording device includes at least one of clarity index, direct-to-reverberant ratio (DRR), sparseness, and reverberation time.
16. The system according to claim 15 , wherein:
a distributed-gradient-boosting algorithm with regression trees is used in combination with signal-to-reverberation ratio (SRR) and the at least one acoustic feature extracted from the acoustic RTF to estimate the distance between the first recording device and the second recording device.
17. The system according to claim 12 , wherein:
an improved proportionate normalized least mean square (IPNLMS) filter is applied by the estimation module for the estimation of the acoustic RTF; and
the distance between the first recording device and the second recording device is estimated based on a set of acoustic features extracted from the acoustic RTF estimated between the first recording device and the second recording device, the set of acoustic features including clarity index, direct-to-reverberant ratio (DRR), and reverberation time.
18. The system according to claim 17 , wherein:
the first recording device is part of a smart home speaker, and the second recording device is part of one of a wearable device or a mobile device.
19. The system according to claim 18 , wherein:
a distributed-gradient-boosting algorithm with regression trees is used in combination with signal-to-reverberation ratio (SRR) and the set of acoustic features extracted from the acoustic RTF to estimate the distance between the first recording device and the second recording device.
20. The system according to claim 19 , wherein the sound signal is one of i) white shaped noise, and ii) speech signal.