OK, so this strays a little from the junto brief, but this is where things led!

My field recording ended up being a noodling around on a SW radio dial, avoiding actual voices as much as possible ... The  resulting recording went through (iZotope RX) speech-to-text, which created the crazy text copied below.  That text then became the basis for the (Google) AI 'vocal', which inspired the final track ... a kind of bendy AI rap ... The SW radio recording is still in there somewhere as well ...

Here's the transcription of what the speech-to-text 'heard' - follow along! The track goes through the complete text twice:

ctured 

is 

that yougoing on 

sayin 

itn't 

hbs 

and i thought it in 

no one d'artagnan offer you there said is 

isled on the he said to missus ellivoice i followed with england is now indeed one of masks thou some otherwise if i think employed having his 

eyes on women soon none of the orange do past for battle but it white 

allwith a black beyond oil the touch of the touch of mitya felt a thought he 

you're just a little fairy and osborne 

ilie 

i'm still 

you'll 

have spaster nothing of saint eliza lady's 

a o'er 

my 


More on this 608th weekly Disquiet Junto project, Nature-to-Text (The Assignment: Using a speech-to-text tool, turn a field recording into instructions for a composition), at: https://disquiet.com/0608/

About the Disquiet Junto: https://disquiet.com/junto/

Subscribe to project announcements: https://tinyletter.com/disquiet-junto/

Project discussion takes place on llllllll.co: https://llllllll.co/t/disquiet-junto-project-0608-nature-to-text/

more halF unusuaL

halF texT [disquiet0608]

Paper view by NotebookLM for paper https://arxiv.org/abs/2410.00037 (2024) Moshi: a speech-text foundation model for real-time dialogue.

It is the 2nd version supposed to be more technical then previous one

Paper review - Moshi_ a speech-text foundation model for real-time dialogue V2

Welcome to halF unusuaL, an evolving space inspired by play in the spirit of ‘hai’ (meaning “fun” or “unusual” in the word haiku). Here, improvisation connects the unexpected, mining the boundary between music and non-music to shape sound worlds that feel oddly familiar in their oddity. Listen freely, explore deeply, and for the sounds to live for your eyes as well as your ears, please stop by the YouTube channel. Thank you for joining this journey! Your support helps keep this afloat ... and is HUGELY appreciated 🥰

I'm machine learning enthusiast with applied math background and a lot experience as full stack developer.
I'm from Ukraine but now I'm working remotely in Montenegro.

New tracks tagged #speech-to-text