You can use this procedure if you have trouble recording to one of the Lab machines:
Record all prompts with another
recording program, such as rec ("rec -r16000 --file=<foo.wav>", e.g.) on the lab machines, Praat or Audacity on another machine (e.g., your
laptop), as .wav files. Each prompt should be a different .wav file,
with names identical to those in e.g. tts/time/wav. Record at a rate
of 16KHz mono if possible -- at least 16kHz is necessary.
Copy the waveforms into e.g. tts/time/wav.
Run ./bin/get_wavs wav/*.wav to convert the files to 16KHz mono which is required for the rest of the system to work.