Estimating bulk delay in a telephone system
A bulk delay estimating circuit matches time intervals representing a signal with time intervals representing an echo of the signal to identify the echo and estimate bulk delay. Bulk delay is estimated by computing (1, 2, . . . n) intervals representing the signal, computing (1, 2, . . . n) intervals representing an echo of the signal, computing absolute differences between corresponding intervals to produce n absolute differences, summing the n absolute differences, and providing an output indicating whether or not the sum is less than a predetermined amount. The intervals are determined by defining a plurality of numbered frames, comparing the energy of a signal during each frame with at least one threshold, storing the numbers of the frames in which the threshold is exceeded, and defining an interval as the period from one frame in which the threshold is exceeded to the next frame in which the threshold is exceeded. Bulk delay is estimated from the frame numbers of the signal, its echo, and the duration of a frame.
1. In a telephone including an adaptive echo canceller having a finite impulse response filter and capable of cancelling echoes longer than filter length, the improvement comprising:
said finite impulse response filter having less than 1024 taps; and
a bulk delay estimating circuit that matches time intervals representing a signal with time intervals representing an echo of the signal for programming said finite impulse response filter;
wherein the bulk delay estimating circuit includes:
a first input for an undelayed signal;
a first circuit for computing intervals in the undelayed signal;
a second input for receiving a delayed signal;
a second circuit for computing an equal number of intervals in the delayed signal;
a subtracting circuit coupled to said first circuit and to said second circuit, said subtracting circuit subtracting corresponding intervals from each circuit to produce a plurality of differences; and
a summation circuit coupled to said difference circuit, said summation circuit accumulating a net difference and providing an output indicating whether or not the net difference is less than a predetermined amount.
2. The telephone as set forth in claim 1 wherein sixteen intervals are computed for each signal.
3. A method for estimating bulk delay of a signal, said method comprising the steps of:
computing intervals representing the signal;
computing an equal number of intervals representing an echo of the signal;
subtracting corresponding intervals to produce a plurality of differences;
adding the differences; and
providing an output indicating whether or not the sum of the differences is less than a predetermined amount.
4. The method as set forth in claim 3 wherein each computing step includes the steps of:
defining a plurality of numbered frames;
comparing the energy of a signal during each frame with at least one threshold;
storing the numbers of the frames in which the threshold is exceeded; and
defining an interval as the period from one frame in which the threshold is exceeded to the next frame in which the threshold is exceeded.
5. The method as set forth in claim 4 wherein said comparing step includes the steps of:
comparing the energy of a signal with a first threshold;
comparing the energy of a signal with a second threshold, wherein the second threshold is greater than the first threshold.
6. The method as set forth in claim 5 wherein said step of storing includes the step of identifying the frame in which the energy of a signal is between the first threshold and the second threshold, thereby substantially eliminating errors due to signals that are not echoes of the signal.
7. The method as set forth in claim 3 wherein the first computing step includes the steps of:
defining a plurality of numbered frames;
comparing the energy of the signal during each frame with a first threshold;
storing the numbers of the frames in which the first threshold is exceeded;
wherein an interval is the period from one frame in which the first threshold is exceeded to the next frame in which the first threshold is exceeded.
8. The method as set forth in claim 7 wherein the second computing step includes the steps of:
defining a second threshold as a function of the first threshold;
defining a plurality of numbered frames;
comparing the energy of an echo of the signal during each frame with the second threshold;
storing the numbers of the frames in which the second threshold is exceeded;
wherein an interval is the period from one frame in which the second threshold is exceeded to the next frame in which the second threshold is exceeded.
9. The method as set forth in claim 3 wherein the second computing step is delayed by a predetermined amount from the first computing step, thereby substantially eliminating errors due to signals that are not echoes of the signal.
10. The method as set forth in claim 3 wherein the computing steps are preceded by the step of looking for a quiet period in the signal.
11. A method for estimating bulk delay of a signal, said method comprising the steps of:
defining a plurality of numbered frames;
looking for a quiet period by
comparing the energy of the signal during each frame with a first threshold, and
finding a quiet period as a minimum number of consecutive frames in which the first threshold is not exceeded;
computing intervals representing the signal by
comparing the energy of the signal during each frame with a second threshold,
storing the numbers of the frames in which the second threshold is exceeded, and
defining an interval as the period from one frame in which the second threshold is exceeded to the next frame in which the second threshold is exceeded;
computing an equal number of intervals representing an echo of the signal by
setting a third threshold in proportion to the energy of the signal in one of the frames in which the second threshold was exceeded,
comparing the energy of the echo during each frame with the third threshold,
storing the numbers of the frames in which the third threshold is exceeded, and
defining an interval as the period from one frame in which the third threshold is exceeded to the next frame in which the third threshold is exceeded;
subtracting corresponding intervals to produce a plurality of differences; adding the differences; and
providing an output indicating whether or not the sum of the differences is less than a predetermined amount.
12. The method as set forth in claim 11 wherein said second computing step includes the steps of:
setting a third threshold in proportion to the minimum energy of the signal in one of the frames in which the second threshold was exceeded; and
setting a fourth threshold in proportion to the maximum energy of the signal in one of the frames in which the second threshold was exceeded.
13. The method as set forth in claim 12 wherein said second computing step includes the steps of:
comparing the energy of the echo with the third threshold;
comparing the energy of the echo with the fourth threshold.
14. The method as set forth in claim 13 wherein said second computing step includes the step of storing the numbers of the frames in which the energy of the echo is between the third threshold and the fourth threshold, thereby substantially eliminating errors due to signals that are not echoes of the signal.
15. The method as set forth in claim 11 wherein the second computing step is delayed by a predetermined amount from the first computing step, thereby substantially eliminating errors due to signals that are not echoes of the signal.
16. The method as set forth in claim 11 including the step of:
estimating bulk delay from the frame number of the first frame in which the energy of the signal exceeded the second threshold, the frame number of the first frame in which the energy of the echo exceeded the third threshold, and the duration of a frame.