I have to assume you’re using an expression pedal for this.  There are other ways, where some of this would still apply.  Q-Wah – Intensity cranked – Vocal Wah – or Bass Vocal Wah, for extended range.  Manual Shape, with Manual Mod mapped to the expression pedal.  Neutralizing all of the secondary modulation controls, to start off.

I think one of the keys to this (and it will still be just an approximation) is to cross over the Start Vowel & End Vowels.  Also map expression from BEAT to BIRD in Start Vowel.  And in the opposite direction, from BIRD to BEAT for End Vowel.  Lots of fine-tuned variations by narrowing down each Vowel range on the two pedal mappings.  Intensity can get dialed down for more.

Foot position varies the crossover points.  I never hurts to feed it some extra harmonics (distortion).  More complexity available by using two Q-Wahs in Parallel, at slightly different settings.  When you’re happy with the results, you can put it over-the-top by folding in a Flanger, also controlled by expression.  Lots of feedback, and ‘crossing over’ ultra-short delays in a similar way.