MuBu

MuBu (multi-buffer) is a Max toolbox for multimodal analysis of sound and motion, sound synthesis and interactive machine learning. It allows to create interactive gesture-based sonic systems, and it is also the base for the CataRT system for Corpus-based Concatenative Synthesis.

Download and Tutorials

• Mubu is available in the Max Package Manager and from the Ircam Forum (opens new window)
• MuBu-related video tutorials (opens new window)
• Video tutorials patches @IrcamForum (opens new window)

MuBu can be used in different typical use cases:

Recording, Playing, Analyzing and Visualizing Multimodal Data

• audio and audio descriptors
• sensors data and motion descriptors
• MIDI
• temporal markers

Real-time Processing of Audio and Sensor Data

• filtering
• segmentation
• computing descriptors (pitch, timbre, FFT, MFCC, wavelets, statistic)

Interactive Machine Learning

• KNN (k-nearest neighbours search)
• PCA (principal component analysis)
• GMM (Gaussian mixture model recognition), GMR (regression)
• HMM (hidden Markov model recognition), XMM (regression);
• DTW (dynamic time warping)
• Gesture following (GF) and Gesture Variation Following

Interactive Sound Synthesis

• Granular synthesis
• Concatenative synthesis
• Additive synthese

License: Forum (Toolbox distributed freely, proprietary code)