Hi everyone, I am opening this PR as the first part of some gammatone filterbank optimizations, proposed with #115.
Here I am only replacing Armadillo with Eigen as the linear algebra backend. This change by itself is not providing any runtime improvements but it allows for a faster implementation of the gammatone filterbank if we use the Eigen matrix operations directly. I added more info on the proposed optimization in #115.
The main purpose of this PR is to split the first part of #115, to facilitate the review process.
Hi everyone, I am opening this PR as the first part of some gammatone filterbank optimizations, proposed with #115.
Here I am only replacing Armadillo with Eigen as the linear algebra backend. This change by itself is not providing any runtime improvements but it allows for a faster implementation of the gammatone filterbank if we use the Eigen matrix operations directly. I added more info on the proposed optimization in #115.
The main purpose of this PR is to split the first part of #115, to facilitate the review process.