Parameter generation¶
Numpy implementation of speech parameter generation algorithm.
build_win_mats(windows, T) |
Builds a window matrix of a given size for each window in a collection. |
mlpg(mean_frames, variance_frames, windows) |
Maximum Parameter Likelihood Generation (MLPG) |
mlpg_grad(mean_frames, variance_frames, …) |
MLPG gradient computation |
unit_variance_mlpg_matrix(windows, T) |
Compute MLPG matrix assuming input is normalized to have unit-variances. |
reshape_means(means, static_dim) |
Reshape means (T x D) to (T*num_windows x static_dim). |