Link Search Menu Expand Document

Analysis processes

Real-time music information retrieval (MIR) processes are available, thanks to the Gist library. They allow to easily build audio-reactive visuals, by extracting meaningful parameters from a live or pre-recorded audio signal.

Here is a video example of their usage:

Envelope

Envelope

These processes allow to extract the volume of an audio signal.

Onset detection

Onset

These processes allow to detect hits and impacts and are useful for beat detection. The first outlet outputs the detection signal. The second outlet outputs an impulse when an onset is detected.

Pitch detection

Pitch

These processes allow to estimate the pitch of a monophonic signal (between 150 and 1500 Hz).

Spectral parameters

Spectral parameters

These processes allow to extract parameters related to the spectrum and the timbre of an audio signal.

Spectrum extraction

Spectrum extraction

These processes convert the input audio signal in an “audio” signal containing the spectra, and are mainly useful for further processing in custom processes.

Beat and tempo following

Beat Tracker combines onset analysis with a tempo estimate and a beat clock. It also accepts external beat events. Tempo, Speed, beat/bar phase, timing pulses and confidence outputs support live following; monitor confidence rather than treating every audio onset as a reliable beat.

Point tracking

Point Tracker associates incoming 2D or 3D detections with stable identities, smooths their motion and reports entry, confirmation, disappearance and revival. It consumes detections rather than detecting objects itself. Its records can directly drive Entity To MIDI.

Pixel analysis and result selection

Pixel Utilities extract image lightness or color components as arrays. For a list of scores or distances, Array Best Match selects an extremum and produces probabilities. Multi-choice instead smooths several inputs and requires a winner to exceed both a threshold and a margin.