Let's x be the input signals batch, dimensions: [batch size, signals length]
No problem producing cochleagrams batch; resulting cochleagrams dimension: [batch size, bands number, cochleagrams length]
However when trying to resample using apply_envelope_downsample, it seems to not handle batches and to expect the input array to be a single cochleagram.
Also the error that may result from using a batch does not refer to this limitation, instead it gets ahead and considers the second dimension (number of bands in the case of a batch) to be the time dimension and fails when the number of sampling points is not integer.
Let's x be the input signals batch, dimensions: [batch size, signals length]
No problem producing cochleagrams batch; resulting cochleagrams dimension: [batch size, bands number, cochleagrams length]
However when trying to resample using apply_envelope_downsample, it seems to not handle batches and to expect the input array to be a single cochleagram.
Also the error that may result from using a batch does not refer to this limitation, instead it gets ahead and considers the second dimension (number of bands in the case of a batch) to be the time dimension and fails when the number of sampling points is not integer.