fofrAI Shares Lyria 3.5 Club Chatter Audio
Research engineer highlights realistic yet indistinct crowd sounds from generative model prompt.
TLDR
fofrAI, identified as a research engineer working on generative image and video models, posted about an audio output from Lyria 3.5. He called it one of his favorite results and praised the real but indiscernible chatter of many people in a club. The prompt he used was "just a conversation between many people in a club, no actual singing." The attached video showed a static blurred photograph. He added that he could not help feeling like he was missing out.
Combined views
2.1M
7 Sources, first seen 26d ago