Experiments with LPC (Linear Predictive Coding) for speech synthesis purposes.
Old school speech synthesis from TI chips (Speak&Spell, and even old car warning systems) sounds super cool.
Ultimately, the purpose of this study is to recreate a similar effect based on typical methods of the 80ies, even though the intent is not to make an exact replica of TI speech synthesis chips.
This will also be an opportunity to dig a little further into topics like analysis/synthesis with overlap, a bit of FFT, fine spectrum interpretation, voice/noise discrimination etc.