Four minutes and fifty-nine seconds into my second LABO video, I hit on a popular question that many viewers seemed to have:
Would it be possible to create a Waveform Card that approximates human speech?
To begin with, this felt like a solid "no" to me - if you've ever watched human speech on an oscilloscope, you know that the waveform produced is far from periodic.
However, I noticed that the five vowel sounds ("Ah", "Ee", "Eh", "Ooh", and "Oh") did produce roughly periodic waveforms when spoken alone.
I made 5 more cards using samples from FL Studio's built-in speech synthesizer, and the results (even when digitized by the Switch's low-resolution infrared camera) were fairly impressive.
In a fervent, last-ditch effort to make speech synthesis on Nintendo LABO ToyCon Piano happen, I spent a while layering component vowel sounds in an attempt to approximate human speech.
The results weren't perfect, but with subtitles, I think you can hear what I was going for.
("I Like You" breaks down to "Ah-Ee Ah-Ee-(staccato "Eh" to make a "K" sound) "Eh"-(that blends into)-"Ooh")