Shortlist B & Merge B

The programs used for the simulations reported in: Norris, D. & McQueen, J. (2008) Shortlist B: A Bayesian model of continuous speech recognition, are available here: ShortlistB.zip, ShortlistB.tar.gz.

The zip files contain the source files, the lexicon used in the simulations, the diphone confusion matrix, and instructions for use. The programs should compile with any C++ compiler, but they have only been tested with GCC.

Remember, this isn't a piece of commercial user-friendly software. It was developed specifically for the purposes of running the simulations in the Shortlist B paper. If you want to use this for serious research purposes we strongly suggest contacting us for advice first. The documentation is pretty minimal, and probably isn't detailed enough to guarantee that the program is going to behave exactly as you think it will.

Nevertheless, the program might be useful to give you an idea of what the program output looks like - especially if you speak Dutch and are familiar with the CELEX DISC notation used for the input and output.

Because Shortlist B depends on the database of Dutch diphone confusions [Smits, R., Warner, N., McQueen, J. M., & Cutler, A. (2003). Unfolding of phonetic information over time: a database of Dutch diphone perception. Journal of the Acoustical Society of America, 113(1), 563-574.], it is highly unlikely that there will ever be a version that works for English.


Dennis Norris & James McQueen