Click the 
to play the audio,
and click the 
to stop playback.
Click the 
to view the true spectral envelopes and vocal tract shapes before and after optimization.
The second audio player in each cell presents a vowel that was synthesized with the modal glottal configuration, which is derived from averaging glottal parameters across all optimized vowels.
The value in each cell represents the mean squared error (MSE, in dB) between the synthesized vowel and the original vowel’s true spectral envelope. For each vowel, the lowest MSE among the algorithms is highlighted.