This application will detail my interests in the CLAM project and
explain what I can offer the project as a participant in the summer of
code. My main interest is in speech processing and recognition. This
gives me a partial overlap in interests in the audio processing
capabilities of CLAM, as well as other interests that could add to
CLAM's capabilities and user base. Specifically, I am familiar with
many of the DSP techniques used in CLAM, but with more of an emphasis
on speech than music. This slight difference in focus applies both at
the DSP level, as well as the user interface level, since linguists
and speech scientists have different needs from a user interface as
well as the DSP behind it: at the DSP level, there are warpings to the
spectrum to simulate the auditory system, and the LPC model is based
on an articulatory model, for example, and at the user interface
level, there are certain displays that linguistics prefer, like black
and white spectrograms and multi-tiered transcriptions, to name a few.
Given my background in linguistics (BA), computer science (MS, working
on PhD), and EE/signal processing (2 speech processing courses), the
project idea that I am most interested in is the vowel synthesizer .
Below in the detailed description I will explain my past experience
building a LPC/source-filter based vowel synthesizer in Matlab for my
speech processing class. I'll also explain other ideas I have, both
to further develop this project, and other ideas I have for
incorporating speech recognition abilities into CLAM, such as singing
synthesis, . I will also describe facilities available to me at my
university that may be helpful.