_____________________________________________________________
  _                                         _           _   
 (_)_ __  _ __  _ __ ___    _ __  _ __ ___ (_) ___  ___| |_ 
 | | '_ \| '_ \| '__/ _ \  | '_ \| '__/ _ \| |/ _ \/ __| __|
 | | | | | |_) | | | (_) | | |_) | | | (_) | |  __/ (__| |_ 
 |_|_| |_| .__/|_|  \___/  | .__/|_|  \___// |\___|\___|\__|
         |_|               |_|           |__/               
_____________________________________________________________

(c) 2007, 2008, 2009 inpro project, Timo Baumann, Gabriel Skantze
funded by DFG Emmy-Noether programme

please feel free to contact timo@ling.uni-potsdam.de 
for any questions and comments. 
see bottom of this page for license and citation information.

_________________________________________
‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾
Pitch-Tracking with the Sphinx-4 FrontEnd
_________________________________________
‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾
This package contains pitch tracking code that can be 
integrated into the Sphinx-4 [1] FrontEnd or used as a 
stand-alone pitch tracker.

The algorithm implemented closely follows the first three 
steps of the YIN algorithm [2], but differs in the fourth
step: YIN implements complete DP, while we use a configurable
look-ahead allowing incremental processing at the cost of
imperfect dynamic programming results. Our testing shows
that a look-ahead of only a few frames (< 20) allows for
similar results as complete optimization.

Also, PitchTracker.java contains several unused functions
that can be used to vary the candidate extraction and 
selection. We believe the current setting should give the
best possible results.

The current implementation is hard-wired to 16000 Hz audio.
Extensions to other frequencies should be straight-forward 
and code contributions regarding this (as well as to all  
other aspects of the code) are very welcome.

____________________________________
‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾
Using as a Stand-Alone Pitch-Tracker
____________________________________
‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾
With the distribution, and sphinx4.jar in the current 
directory just run

java -jar PitchTracker.jar summkurz.wav

If your sphinx4.jar is located somewhere else, you will
have to alter the path in PitchTracker.jar!META_INF/MANIFEST.MF

Of course, you may use any other 16kHz audio file
other than the provided summkurz.wav

Pitch is output in Hz for every 10 ms of audio.

_______________________________________
‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾
Pitch-Tracking in the Sphinx-4 FrontEnd
_______________________________________
‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾
Add the pitch tracker as a DataProcessor somewhere near 
the beginning of the FrontEnd pipeline (when Data objects 
are already chunked to 10ms and are still audio samples).
See config.xml for an example.

Pitch and voicing information will be output as part of
PitchedDoubleData (see PitchedDoubleData.java). This 
PitchedDoubleData could either be used in a later 
processor to calculate non-standard ASR features (which
I have not tried), or passed on to some other component. 

This can be accomplished by implementing the Listener
interface in 
org.cocolab.inpro.pitch.notifer.SignalFeatureListener,
plugging in the SignalFeatureExtractor into the pipeline
below the PitchTracker and registering your listener 
with the SignalFeatureExtractor. 

This unfortunate procedure is necessary, as there is 
no way of passing PitchedDoubleData completely through
the recognition process as preceding Data objects are 
discarded by later frontend modules.

_______________________________________
‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾
License and Citation in Scientific Work
_______________________________________
‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾
This is free and open source software, released under the
General Public License 3. Please see the file COPYING for details.

The license sets the requirements for using this software.
The requirements in this section are not legally binding; 
however, they represent things that the author would appreciate: 
If you write a paper based on work that uses this software, 
please include a citation to one of our papers, and send a copy 
of your paper. If you create a derivative work based on a this 
software, tell me about it and let me know how I can get a 
copy, and include this request if it’s appropriate.

See the file CITING for references to our papers that describe
this software. Feel free to cite the paper that seems more 
appropriate.

_______________________________________
‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾
[1]: 
cmusphinx.sourceforge.net/sphinx4/
[2]: 
de Cheveigné and Kawahara (2002).
YIN, a fundamental frequency estimator for speech and music.
The Journal of the Acoustical Society of America, volume 111