AI Glossary

Audio tokenization

What does “Audio tokenization” mean?

Audio tokenization: Representing audio as discrete tokens that a model can process or generate.

Audio tokenization visual guide: Representing audio as discrete tokens that a model can process or generate.
Visual guide Open full size (new tab)

What is an example of “Audio tokenization”?

A speech model predicts codec tokens that are converted back into a waveform.

Keep learning

See also

Speech-to-textTermText-to-speechTermSpeaker diarizationTermSpeech translationTermAudio source separationTerm

Same topic

Neural audio codecTermCFG scaleTermControlNetTermAI avatarTermDenoising diffusion implicit modelTerm