Timo Baumann - Software

This page contains information about some recent software which I developed in order to do my research. As I use quite a bit of very decent scientific (and non-scientific) open-source software, I want to give back the little I can.

The packages below should be more or less self-containing, but you will sometimes have to resolve dependencies to other projects yourself. Do not expect too much documentation-wise. You will be unlikely to get up-and-running quickly, if you are not good at Java (or Perl) programming and don't have any knowledge of speech-recognition to start with. But if you know the basics, the code should be fairly easy to read.

Pitch Tracking* for Sphinx-4

A pitch tracker that works as a Sphinx-4 frontend processor.
README, complete package

* And signal power. But that's easy.

A more up-to-date and more complete version of this code is available as part of InproTK, our toolkit for incremental spoken dialogue systems.

Filtering of incremental output for Sphinx-4

An implementation of the methods described in our paper at NAACL-HLT 2009 for the Sphinx-4 speech recognition.
README, example config, complete package

A more up-to-date and more complete version of this code is available as part of InproTK, our toolkit for incremental spoken dialogue systems.

(IN)TELIDA

TELIDA has its own page.

InproTK

InproTK has its own page.