I’d like to use Edge Impulse to create a model that takes speech and outputs articulation parameters used to control Pink Trombone, an articulatory speech synthesizer. I’ve done some examples before using simpler models (I’m not that good at ML stuff), and I figured Edge Impulse would be a better solution.
I’ve used Edge Impulse before to create a classifier with our smart insoles, as well as a simple regression model using a Wii Balance board. Is it possible to create a regression model that provides multiple outputs (index and diameter of multiple vocal tract constrictions, 6 total)?