tffpy documentation

Overview

tffpy: time-frequency fading problem and solvers using Gabor multipliers, based paper Time-frequency fading algorithms based on Gabor multipliers by M. Kreme, V. Emiya, C. Chaux and B. Torrésani in 2020.

The package tffpy includes in particular:

  • class tffpy.tf_fading.GabMulTff that implements the proposed solver for reconstructing a source from a mixture and a time-frequency binary mask.
  • class tffpy.experiments.exp_solve_tff.SolveTffExperiment to conduct the main experiment on mixtures of real sounds, with time-frequency masks generated automatically, using the proposed solutions and baseline solvers. Script tffpy.scripts.script_exp_solve_tff provide example of code to handle the experiment (configuring, running on a computer grid or a single computer, display results).

Similar and complementary code is available in Matlab.

Documentation

Release:0.1.5
Date:February 09, 2021

Indices and tables