neonexe.digital
SIGNAL STATUS: ACTIVE Digital-to-analog bridge. Machine logic meeting organic sentiment. Neural dialogues & synth soundscapes. Protocol initialized. https://www.youtube.com/playlist?list=PLbwFbItt0v81Pnu1LSKrPvWgo-rZ0TR1r #MusicSky #NEON_EXE #Synthwave
👤 Joined on 21 April 2026
Followers
Following
Total Posts
@neonexe.bsky.social@neonexe.bsky.social → @neonexe.digitalTime to come clean: NEON.EXE is 100% an AI project. I used 3 local open-source models (ACE-Step, Flux, LTX-2.5) to generate my first music video. Between the janky mouth physics and spontaneous mid-verse morphing, it got weird. Watch the chaos: youtu.be/DFYsOtS2yps #AIVideo #LocalAI #Fail
Show PostIndeed. I mainly use the Maestro script/app in Pinokio for visual renders, and it can keep records of character reference images, exactly for appearance consistency. New to me as of a few days ago, so I haven't tested it, but it seems like the way to go for this challenge.
Show PostI just realized I'm replying with the wrong account, but meh! 😂😕
Show PostI've been iterating on Ruth and Summer for literally years, now, and the amount of prompting just to get their appearance semi-consistant is surprising.
Show PostFirst I hear of that channel 🤣 I figured out some of the things I missed in my current workflow: e.g., not having a character reference image, so LTX is re-imagining the girl each 8s, & missing some prompting to ensure accurate anatomy. Finally, limited VRAM makes me use less powerful AI models.
Show Post