Blob Opera is a Google Arts and Culture experiment created by developer David Li where four animated blobs sing operatic harmonies you control in real time. Each blob represents a different vocal range (soprano, mezzo-soprano, tenor, bass) and was trained on recordings from professional opera singers. Drag a blob up or down to change pitch; left or right to shift vowel sounds. All four voices harmonize simultaneously based on where you place them. The output is genuinely musical: the AI has learned the harmonic relationships of Western choral singing and applies them in real time. You can record and share your performances. It demonstrates how machine learning can model the relational structure of music in a way that's immediately playable by anyone, no musical training required. Blob Opera went viral in late 2020 and resurfaces regularly in 'best AI demos' threads. It's a frequent recommendation in educational settings as a friendly entry point to ML concepts. Typical first reaction from new users: they play with it much longer than they expected. Based on community discussions from Reddit and design communities.

What the community says

One of the most celebrated AI experiments ever released. Reddit and design communities consistently recommend it as an elegant demo that anyone can enjoy. The most common feedback: people expect to spend 30 seconds with it and end up staying for ten minutes. No notable criticism beyond wishing it had more features. Based on community discussions from Reddit and design communities.