Show HN: A singing synthesizer for the browser with algorithmic 3-part harmony https://ift.tt/Bpa7N2C

Show HN: A singing synthesizer for the browser with algorithmic 3-part harmony Hello HN! This is a demo I built during my batch at the Recurse Center. It's built using WebAudio, WebAssembly (via Rust), and the Canvas API. The source code for this demo, along with my other vocal synthesis related RC demos[0], can be found at a monorepo[1] where I've been dumping all my RC work and logs. The sound is generated using a physical model of the human singing voice, using a work-in-progress project called VoxBox [2]. The harmonization is done using something that kind of resembles a markov chain. Only, instead of using weighted probabilities and randomness, I use a selection heuristic that chooses a chord based on how often it has been used and how much voice movement is required. Thanks for reading! Happy to answer any other questions. 0: https://ift.tt/ZCP1XLO 1: https://ift.tt/9d0Utxg... 2: https://ift.tt/Bfd3c06 https://ift.tt/SLAd5UK August 4, 2024 at 05:05PM

Comments

Popular posts from this blog

Show HN: Agent File (.af) – An open file format for agents https://ift.tt/fzI5HcG

Show HN: Sort lines semantically using llm-sort https://ift.tt/7vEeHKP

Show HN: AgentKit – JavaScript Alternative to OpenAI Agents SDK with Native MCP https://ift.tt/H4Kz5Vi