LAB 15: Speech Synthesis Using a Formant Synthesizer

Labs: 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17

The main purpose of this laboratory is to introduce you to formant synthesis through synthesis of a monosyllabic English word or a phrase.

Here are the resources you will need:

Lab Handout (PDF)

Supplement to Speech Synthesis Using a Formant Synthesizer (PDF)

Utterances

• New recordings to be made during the lab

If you do not have access to the tools for recording the new utterances, you can use one of the utterances from Lab 5, part B.

The Tools

• record - used to make recordings (VAX)
• lspecto - used to make spectrograms, F0 contours, and formant tracks (UNIX®/Linux®)
• xkl - used for spectrum and waveform analysis and formant synthesis (UNIX®/Linux®)

Reference Readings

Klatt, D. H. "Software for a Cascade/Parallel Formant Synthesizer." J Acoust Soc Am 67 (1980): 971-995.

Klatt, D. H., and L. C. Klatt. "Analysis, synthesis, and perception of voice quality variations among female and male talkers." J Acoust Soc Am 87 (1990): 820-857. (Optional)

Klatt, D. H. "Description of the cascade/parallel formant synthesizer." Chapter 3 of a book in preparation.