The vocal folds make the source: a set of harmonics spaced by the pitch (F0). The vocal tract is the filter: its resonances (formants) boost some harmonics. Change one, and watch the other stay put.
Try it: keep the vowel and slide the pitch (the formant peaks stay put). Then keep the pitch and switch vowels (the harmonics stay put).
Vowel presets are average formant frequencies of adult male speakers from Peterson & Barney (1952), J. Acoust. Soc. Am. 24:175–184. Formant bandwidths are set to typical values (80, 100 and 150 Hz). The source is drawn with harmonic amplitudes falling as 1/n (about −6 dB per octave, the usual combined effect of the vocal-fold source and radiation from the lips). The sound you hear is built from exactly the harmonics in the bottom panel, so the picture and the sound match.