-- Technical Slides --

Pay no attention to the slides behind this curtain...

Presentation Setup - 1 software requirements

this presentation tarball and web browser
MP3 player
festival
Edinburgh Speech Tools (the speech-tools package in Debian, /usr/libexec/speech-tools/ in the festival package in Fedora Core)
LLG festvox WAV file ZIP archive (akwavs.zip)
GNU prolog (for la nelci bebna)
NU prolog and some variant of lex (both for nitcion's semantic analyzer)
Perl (for local execution of the random sentence generator)

Presentation Setup - 2 jbo_us1 festvox

# ln -s english /usr/share/festival/voices/lojban
# ln -s kal_diphone /usr/share/festival/voices/lojban/kal_diphone_jbo_us1
# cp local/festvox/kal_diphone_jbo_us1/festvox/kal_diphone_jbo_us1.scm /usr/share/festival/voices/lojban/kal_diphone_jbo_us1/festvox/

Presentation Setup - 3 LLG festvox

Note: Fedora Core users must first execute ln -s . /usr/libexec/speech-tools/bin to compensate for a non-standard ESTDIR layout on that distribution.

Install (decompress) the wavs you downloaded into festvox/ljb_diphone/wav/
make sure your ESTDIR (Edinburgh Speech Tools Directory) is set properly (for Debian it should be set to /usr) and run make. This should build the timings from the Praat files and generate the pitchmark files from the wavs.
Edit ljb_diphone/festvox/lojban_diphone.scm so it reflects your installed path. XXX this should go away

Presentation Setup - 4 LLG parser

$ gunzip -c local/parser.shar.gz > parser.shar
$ unshar parser.shar
$ cp mill/parser/*.c . # poor man's patch
$ make `for i in *.c; do echo "$i" | sed 's/\.c$/.o/'; done`
$ cc -l c -o node *.o

Presentation Setup - 5 analyzer

$ lex -l -oanalyser.c analyser.lex
$ gcc -l c -l fl -o analex analyser.c
$ cat mill/genesis-excerpt.txt | node -dl | analex > cooked.txt
$ $YOURPROLOG


  | ?- ['mill/analyser.pl'].

  | ?- parse('cooked.txt').





Pre-Presentation Test

  S5
  web connectivity
  MP3 player
  GNU Prolog
  jbo_us1 festvox
  LLG festvox
  LLG parser
  nitcion's analyzer
  Perl