Double-talk detection
Within audio telecommunications for audio-visual conferencing a means of reliable double-talk detection is required for several reasons, including control of residual echo, and control of adaptive algorithms used to estimate the echo. Existing techniques tend to either fail when echo return loss is very low or require highly complex algorithms to increase the range of detecting whether double-talk is occurring. The invention is based upon a new method of estimating the power envelope of the double-talk signal. This estimation is then employed in conjunction with a slightly modified version of an echo return loss enhancement (ERLE) based statistic, to form a double-talk decision based on a combination of the two statistics. The resulting decision rule works for both low-ERLE and high-ERLE situations, which is useful when there can be large variations in the magnitude of the echo path seen by the algorithms, or where near-end noise prevents the achievement of high ERLE.
1. A method comprising:
providing a first signal from a local transducer;
providing a second other signal received from a remote end of a communication path;
determining based on the first signal and the second signal a first indication of doubletalk based on echo return loss;
determining based on the first signal and the second signal a second indication of doubletalk based on other than echo return loss; and,
using a combination of the first indication and the second indication to provide an output indication of doubletalk.
2. A method according to claim 1 wherein the first indication and the second indication are formed by comparing a determined value to a threshold value.
3. A method according to claim 2 wherein the threshold values are adjusted in dependence upon a statistical analysis of the echo return loss.
4. A method according to claim 3 wherein the threshold values are adjusted such that when echo return loss is lower, dependence upon the second indication is increased.
5. A method according to claim 3 wherein the threshold values are adjusted such that when echo return loss is higher, dependence upon the first indication is increased.
6. A method according to claim 3 wherein the threshold values are adjusted such that when echo return loss is higher, dependence upon the second indication is decreased.
7. A method according to claim 3 wherein the threshold values are adjusted such that when echo return loss is lower, dependence upon the first indication is decreased.
8. A method according to claim 1 wherein:
doubletalk is determined when E 2 <E 2 Thresh , wherein E 2 =E 1 /ERLE and wherein E 1 =S IN /S OUT .
9. A method according to claim 8 wherein:
doubletalk is determined based on a value DT 1 wherein DT 1 =S IN −Est(Echo IN ) and wherein Est(ECHO IN )=R OUT /(ERL*ERLE).
10. A method according to claim 9 wherein DT 1 is scaled by its variance.
11. A method according to claim 9 wherein double talk is determined based on the combined function such that ƒ(DTx, Ex)>CombThresh.
12. A method according to claim 11 wherein the combined function is of the form (α*DTx)/(NoiseVar*β*Ex)>CombThresh.
13. A method according to claim 12 wherein the combined function is ((DTx>DTx Thresh ) (Ex<Ex GatingThresh ))∥((Ex<Ex Thresh ) (DTx>DTx GatingThresh )).
14. A method according to claim 9 wherein DT 1 is scaled in accordance with DT 2 =(S IN −Est(ECHO IN ))/N effpwr .
15. A system for doubletalk detection comprising:
a receiver for receiving a first signal from a local transducer;
a receiver for receiving a second other signal from a remote transducer, and having already been processed;
a processor for determining based on the first signal and the second signal a first indication of doubletalk based on echo return loss and a second indication of doubletalk based on other than echo return loss and for using a combination of the first indication and the second indication to provide an output indication of doubletalk.