STFT#

class src.stft.StftManager.StftManager(dataset_tensor: Tensor, dataset_dir: str, dataset_type: str, data_load: str, tensor_coefficients_dir: str, tensor_transform_info_dir: str, stft_reconstruct_dir: str)#

Bases: object

Class that implements the STFT manager.

compute_stft(verbose: bool = True)#

Method that handles STFT computation.

Parameters:

verbose (bool) -- Boolean that indicates weather to print specific output.

save_coefficients(coefficients: Tensor, nb_saved_tensor_coefficients: int, verbose: bool = False)#

Method that saves STFT data.

Parameters:

  • coefficients (torch.Tensor) -- STFT data tensor.

  • nb_saved_tensor_coefficients (int) -- Number of saved data files.

  • verbose (bool) -- Boolean that indicates weather to print specific output.

save_transform_info(verbose: bool = False)#

Method that saves Transforms metadata.

Parameters:

verbose (bool) -- Boolean that indicates weather to print specific output.

update_transform_info(coefficients: Tensor, max_qv: float = 0.99)#

Method that handles Transforms metadata computation.

Parameters:

  • coefficients (torch.Tensor) -- STFT data tensor.

  • max_qv (float) -- A scalar parameter in [0, 1].

class src.stft.StftManager.Stft(n_fft: int, win_length: int, hop_length: int, power: float | None, normalized: bool, center: bool, batch_size: int, stft_reconstruct_dir: str)#

Bases: object

Class that implements the STFT computation.

property mini_batch#
property mini_batch_dir#
stft_fn(batch_index: int, verbose: bool = False)#

STFT computation method.