KlattWorks is a new interface to the Klatt (1980) Synthesizer that allows for greater flexibility and control of your parameters.
It is specifically designed for the speech perception researcher, but others may find it useful as well.
KlattWorks is not a synthesizer itself (it actually uses an external application to perform the synthesis), rather it
simply manages and creates sets of Klatt Parameters that describe synthetic utterances.
KlattWorks has several features that extend existing implementations of the Klatt Synthesizer:
- Algorithmic Description of Parameters: KlattWorks allows the user to describe the values of
a single parameter (e.f. F0) as mathematical functions of time. This allows the user to efficiently
manipulate parameters (for experimental manipulations) by simply changing the parameters of that
function. A wide range of these functions is included (linear, sine, logistic, etc...). Functions
may also be layered together to form a complete description and they can also be scripted.
- Non-linear descriptions of parameters: KlattWorks introduces the logistic function for use
in speech synthesis. This function describes two steady states with a nonlinear shift in between them.
For example, F1 may start at a steady state of 300hz and shift to another steady state of 600hz at 30ms.
This non-linear function is an excellent way to approximate transitions between vowels (consontants).
- Scripting: Every function and operation in KlattWorks can be built into a script for generating
the complete word. This provides an easy way to duplicate stimulus items and making small changes. It also
provides a compact record of your stimulus.
- Naturalness:KlattWorks provides an easy to use interface to the popular Praat speech analysis
program. This allows you to automatically read formant and pitch frequencies of natural speech and use them
as the starting point for Klatt Synthesis.
KlattWorks was concieved in 2000 and has been in continual development since then. It has been used experimentally
in a series of speech perception experiments described in McMurray, Tanenhaus and Aslin (2002), and McMurray,
Tanenhaus, Aslin and Spivey (in press).
KlattWorks is freely available to the public. The author asks that any paper that uses it cites a paper currently
being prepared for publication:
McMurray, B. (in preparation) KlattWorks: A [somewhat] new systematic approach to formant-based speech synthesis for empirical research.
At this time, a compact installation program has been created, and can be obtained here.
KlattWorks is only available for PC's running Microsoft Window 98 or higher and Microsoft Access 2000. A run-time version of MS Access may soon be available that
would permit users without Access to run it. We're having trouble running it under 64-bit versions of windows, but it seems to runs stably using virtual machines to simulate a 32-bit environment.
An "in progress" version of the manual in PDF format is available here.
Klatt, D. (1980) Software for a Cascade/Parallel Synthesizer. Journal of the Acoustical Society of America, 67, 971-995.
McMurray, B., Tanenhaus, M., and Aslin, R. (2002) Gradient effects of within-category phonetic variation on lexical access. Cognition, 86(2),, B33-B42.
McMurray, B., Tanenhaus, M., and Aslin, R. and Spivey (in press) Probabilistic constraint satisfaction at the lexical/phonetic interface: Evidence for gradient effects of within-category VOT on lexical access.
Journal of Psycholinguistic Research.
Page last updated on 12/3/02
by Bob McMurray