On my first waveform card video, a viewer asked if the 3D-printed cards explicitly had to be white to be recognized by the Switch's infrared camera. To me, this seemed like a definite "no" - the camera was just detecting the card's shape, and it couldn't even see light in the visible spectrum.
To answer their question (and to spice things up a little bit - those white cards were starting to look a little drab), I decided to print the cards for this second video using red and yellow filament.
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.
When I made the files for this project available on hunterirving.com, I had no idea so many people would find them useful.
I guessed that most who downloaded them would use the provided STL files to make 3D-printed cards as I had, but it seems that the SVGs I included turned out to be much more popular:
Reddit user compacta_d used a Kongsberg cutting table to cut these cards from E flute corrugated.
Reddit user AchillesPDX used a hobby laser cutter to cut these cards from cardboard.
Bigtime YouTube channel Melody Geeks even used a few of them in some of their LABO cover videos.
Knowing that other people were able to gain something from my project is a pretty great feeling.
Look forward to more open-source projects from me in the future!