Home▸AI & Machine Learning▸Speech Recognizer — HMM Viterbi Decoding Live

🎙️ Speech Recognizer — HMM Viterbi Decoding Live

Watch a hidden Markov model decode a synthetic acoustic feature sequence into the most likely phoneme path using the real Viterbi algorithm, with a live trellis visualization.

AI & Machine Learning3DAdvanced60 FPS

⚙ Under the hood

Watch a hidden Markov model decode a synthetic acoustic feature sequence into the most likely phoneme path using the real Viterbi algorithm, with a live trellis visualization.

Three.jsWebGLAIMachine Learning

3D · Three.js / WebGL renderer · 60 FPS target · runs fully client-side, no install