Method for weighted overlap-add
A method for generating an output sequence of samples in response to a first and a second subsequence of samples, the method comprising —applying a weighted overlap-add procedure to the first and second subsequences so as to generate the output sequence of samples, —optimizing a weighting function involved in the weighted overlap-add procedure in response to a measure of matching between the output sequence of samples and one or more target sequences of samples.
1. A method for generating an output sequence of samples in response to a first and a second subsequence of samples corresponding to a digitized audio signal received over a network, the method comprising:
applying a weighted overlap-add procedure to the first and second subsequences to generate the output sequence of samples; and
optimizing a weighting function involved in the weighted overlap-add procedure in response to a measure of matching between the output sequence of samples and one or more target sequences of samples, the weighted overlap-add procedure comprising multiplying the first subsequence sample-by-sample with a first window and multiplying the second subsequence sample-by-sample with a second window, and a shape of at least one of the first and second windows is optimized in response to the measure of matching, and the optimizing includes adjusting at least one of the weighting function involved in the weighted overlap-add procedure or an additional weighting function involved in the weighted overlap-add procedure, the weighting function involved in the weighted overlap-add procedure represented by a set of parameters that are based on a decomposition of a temporal shape of the weighting function into a linear combination of basic window shapes.
2. A method according to claim 1 , wherein the weighting function is optimized in response to two or three target sequences of samples.
3. A method according to claim 1 , wherein the optimizing includes adjusting at least the weighting function involved in the weighted overlap-add procedure and the additional weighting function involved in the weighted overlap-add procedure.
4. A method according to claim 1 , wherein the set of parameters are based on a decomposition of a temporal shape of the weighting function into a linear combination of two or three basic window shapes.
5. A method according to claim 1 , wherein the set of parameters is generated by applying at least one tapped delay line to at least one of the first and second subsequences of samples.
6. A method according to claim 1 , wherein the set of parameters is generated by applying two tapped delay lines to the first and second subsequences of samples.
7. A method according to claim 1 , wherein the set of parameters is generated by applying three tapped delay lines.
8. A method according to claim 1 , further comprising modifying the weighting function involved in the weighted overlap-add procedure in response to a measure of distance between the output sequence of samples and the at least one target sequence of samples.
9. A method according to claim 8 , wherein the measure of distance is based on a summed squared error.
10. A method according to claim 9 , wherein the measure of distance is based on a weighted summed squared error.
11. A method according to claim 8 , wherein the measure of distance is minimized according to a least-squares procedure.
12. A method according to claim 11 , wherein the measure of distance is minimized according to a weighted least-squares procedure.
13. A method according to claim 1 , wherein samples of the first and second subsequences represent digitized audio samples.
14. A method according to claim 1 , wherein one of the first and second subsequences of samples is a sequence of concealment samples.
15. An arrangement for receiving the digitized audio signal over the network, the arrangement including:
memory means for storing frames representative of a received digitized audio signal, and
processor means for performing the method as claimed in claim 1 .
16. A computing device comprising:
at least a memory and a processor configured to implement operations to generate an output sequence of samples in response to a first and a second subsequence of samples corresponding to a digitized audio signal received over a network, the operations comprising:
applying a weighted overlap-add procedure to the first and second subsequences to generate the output sequence of samples; and
optimizing a weighting function involved in the weighted overlap-add procedure in response to a measure of matching between the output sequence of samples and one or more target sequences of samples, the weighted overlap-add procedure comprising multiplying the first subsequence sample-by-sample with a first window and multiplying the second subsequence sample-by-sample with a second window, and a shape of at least one of the first and second windows is optimized in response to the measure of matching, and the optimizing includes adjusting at least one of the weighting function involved in the weighted overlap-add procedure or an additional weighting function involved in the weighted overlap-add procedure, the weighting function involved in the weighted overlap-add procedure represented by a set of parameters that are based on a decomposition of a temporal shape of the weighting function into a linear combination of basic window shapes.