P.A.R.T.y🦄)
the speech engine is getting flipping ready to go here. sharing here 4 head scripts I've been developping since last week.
the idea/rule: we speak trough our phones, sending voice messages in our mother languages. there's a beauty node js i made just pulling out my incoming audio messages. then a #python script using a very smaill glitchy #vosk-model (perhaps i'll get a fancy one later ;))...once the english transcript is done, I'm just rendering that into one of the three voices, if it was italian the virtual
@giovanna_sevi will poetically appear into the PA system, if french
@camilleschneegans and if spanish
@franuberto ... that transcript in a separate channel is prompting some early gpt trahsy LLM model (intentional, even we can change model as we go...), AND that response "creative" or not, will get render again with one of the performers voice... latency is flirting in the threshold here... part of the experience, adding to the neutrality, and the mistery of the script some spicy "waiting time"...
and of course, if you play the video game, and eat some pills (the video game will be part of the reality in P.A.R.T.y🦄) you'll be changing the humor/content of the prompted system for instance #gpt french performer has for now this options to go: default, recursive, lucid_break and noise_overload... long story short... next 17th of june, at MFS
@HfMT part of my #artilacs frame? >>>
@blurrededges2026 ...
more soon, lav'u alll...
take care, have fun!