My latest implementation of the Flexible Audio Source Separation Toolbox (FASST – http://bass-db.gforge.inria.fr/fasst/) in Python is not available. It can be found on my github account: http://github.com/wslihgt/pyfasst. At the moment of writing, the EPFL git repository for pyFASST mirrors the same version. One can clone it with the following command (in a command line terminal, […]
When I bought my mac, during my PhD preparation, I decided not to install Matlab anymore, for various reasons: first, there are free and open source softwares that advantageously replace Matlab, and second, because I wanted to check whether these alternatives were viable in academic research. In this article, I want to explore (and share) my experience in replacing a precise point in Matlab, which is the MEX-file mechanism. How do I create extensions for Python/NumPy or Octave, to make the computations faster? If I can write a C code of my algorithm, can I use it under Python? Should I really do that?!! Well, without being an expert in programming, I believe my experience can enlighten these questions, if not addressing them. This article also serves as a container for personal notes and reminders.
I just stumbled on this video, on universcience.tv:
We propose a semester project for EPFL students on “Visual Cues for Computer Aided Pronunciation Training (CAPT)”, a collaboration between the LTS5 and SpeedLingua. See here for details.