Hi All, <br><br>I just wanted to post some first baby steps I made when playing around with the vowel synthesizer that David made (experimentalVowelSynth.clamnetwork in the example data of the NetworkEditor). The description is below, but since it is longish I wanted to ask first about the wiki page I made for keeping project notes:
<a href="http://iua-share.upf.edu/wikis/clam/index.php/SoC_Wiki_Pages">http://iua-share.upf.edu/wikis/clam/index.php/SoC_Wiki_Pages</a> ... Is this an appropriate place to keep notes? If there's a better place let me know.
<br><br>About David's vowel synthesizer, I had the idea of using natural voice source (the vocal fold vibration) as an input to the vowel synthesizer instead of the harmonic peak generator. The natural voice source data comes from a EGG (electroglottogram) at the USC phonetics lab, which measures the opening and closing of the vocal folds.
<br><br>It seemed like this would be hard to do, but after meeting with Xavier last week, I realized it wasn't. The main thing was learning how to use jack to get the sound into clam, and the rest was rearranging a couple of the arcs in the network.
<br><br>The vowel synthesizer already has an audio input, which is used to compare the user's formants (spectral envelope) with the synthetic one (as in the case of using the synthesizer to practice the vowels of a foreign language). I just unplugged the harmonic peak generator and plugged the input's spectral peaks (from the SMSAnalysisCore0's third output) into the vowel resonator. Then I played the EGG data (which is in wav format) with xmms and routed it into the vowel synthesiser. The result sounded a bit more natural than the original, but this may be because of the natural pitch and duration modulation rather than the actual EGG input (the EGG input is a measure of the opening and closing, which has the same pitch as the acoustic source but is not the same thing). It also sounds a bit funny because the data is supposed to come from sentences said with a given emotion.
<br><br>I put the data at <a href="http://sail.usc.edu/~kazemzad/EGG/only_source/">http://sail.usc.edu/~kazemzad/EGG/only_source/</a> and the network at <a href="http://sail.usc.edu/~kazemzad/EGG/clam/experimentalVowelSynth_from_EGG.clamnetwork">
http://sail.usc.edu/~kazemzad/EGG/clam/experimentalVowelSynth_from_EGG.clamnetwork</a> . Well, let me know if you have any feedback about this or the wiki location.<br><br>Thanks,<br>Abe<br><br><br><br><br>