Skip to content

Latest commit

 

History

History
7 lines (5 loc) · 758 Bytes

File metadata and controls

7 lines (5 loc) · 758 Bytes

Statistically-optimal-estimation-of-signals-in-modulation-spaces-using-Gabor-frames

Real data simulation

Two mono WAV files, a piano recording and a blackbird recording from xenocanto both sampled at 44.1 kHz. White gaussian noise is then added to obtain various signal-to-noise ratios, -10, 1, 10, 30 and 100. Denoising was then applied using the universal threshold. We present results for the denoising with gabor frames using a gauss function as well as level 4 B-splines as window function. The results can be compared to wavelet shrinkage methods using the biorthogonal B-splines of level 4 ('bior4.4'). The parameters are chosen according to our publication.

Python code

The python code used for the simulation is also available. /Simulation