Audio, Music, and AI Lab (AMAAI)

The Audio, Music, and AI lab at Singapore University of Technology and Design focuses on cutting-edge innovations in multimodal AI, more specifically: Audio and Music AI.

More info and publications here.

Popular software:

• SonicMaster: all-in-one music restoration and mastering: code - examples - live demo
• Jam 0.5: text-to-song: code - examples- Dataset in collaboration with Declare lab
• SonicVerse: time-aware music captioning: code - live demo
• Music2Emo: emotion detection from music: code - live demo
• Mustango: text-to-music generation: code - live demo
• Video2Music: video-to-music generation: code
• Text2midi: text-to-midi generation: code
• nnAudio: on-the-fly spectrogram extraction: code

Popular Datasets:

• JamendoMaxCaps: text captions with instrumental music audio
• MusicBench: text captions with music audio
• MidiCaps: text captions with music midi (large-scale)
• SonicMaster: music with mastered / enhanced version and enhancement caption