If you get involved in phonetic vocal trapezoid and how vovels are generated in the mouth the principle of the Grendel Formant Filter becomes clear:
The Size parameter influences the size of the virtual mouth by shifting all filter cutoff frequencies parallelly. The parameters Articulator X and Y refer to the international phonetic vovel diagram (see links) and select the desired vovel. All three parameters can be voltage controlled and are optimized to -/+5V voltages (but can handle up to -/+10V), yet the articulators work best either in the range of 0...5V or -2.5V...+2.5V
Frequency-rich signals like sawtooth and noise is recommended. Great for close simulation of human vocals are asymetrical triangles with leading edge shorter than the trailing edge. This can be achieved with function generators like Maths and Function or the RS-95e VCO.
audio: input and output
CV inputs: Size, Articulator X and Articulator Y
3U Eurorack module, 12HP wide, 30mm deep
power consumption: each 65mA @ positive and negative rail