AI Speech Generator OR We're not all born with a big radio voice

Asher - K0AU

With all the AI progress in natural-sounding speech generators, how about a Flex feature to

a. Auto-generate a radio or contest voice from Mic or DAX audio in real-time

b. Text-to-speech using the same parameters

Speech generator is AI inference which reportedly DSPs are good at. In most cases the mic audio path does not go through a PC, so putting it in the radio DSP path is more general and would work with Maestro. V1 can support just a few languages and probably not target tonal languages unless Flex has a big market there we don't see.

Huge AI potential, especially for contesting. It's the biggest technical opportunity IMO since direct A/D sampling. The radio/edge processing should all be AI inference and I can imagine cloud processing using recorded audio and contest logs for training for other, unimagined AI features.

AI (Artificial Intelligence) Sound Processing for Ham Radio (youtube.com)