Music context system audio track structure and method of real-time synchronization of musical content

A digital signal processor “DSP” for separating a transient part of a digital representation of a signal from a relatively stationary part of the digital representation of the signal and to find a boundary position of the transient part, the DSP arranged to:

• produce a time domain representation of the digital signal;
• generate an amplitude envelope of the transient part;
• set a power threshold and detect localised peaks in the amplitude envelope that exceed the power threshold;
• and
• with respect to each localised peak, working backwards in time from that localized peak to find a sampling point at which a characteristic of the digital signal before that point in time is maximally different from that same characteristic of the digital signal after that point in time;
• and
• select, as the boundary position of the onset for the localized peak, the sampling point that is maximally different.

A system comprising a digital signal processor arranged, for a digital audio file having transient parts and relatively stationary parts, to:

• separate the transient parts from the relatively stationary parts to produce a time domain representation, wherein said relatively stationary parts are frequencies that appear across consecutive sampling windows and transient parts are frequencies that demonstrate significant changes in energy in contiguous sampling windows where each sampling window has a duration of between about 10 milliseconds and about 50 milliseconds;
• generate an amplitude envelope of the transient parts;
• set a power threshold and detect a localized peak in the amplitude envelope that exceed the power threshold;
• and
• from the localized peak, to iterate backwards in time on a sample-by-sample basis over the duration of the sampling window to find a sample point at which a measurable characteristic of the digital audio signal, measured during a first period having a first duration extending backwards in time before the sampling point, is maximally different from the measurable characteristic of the digital audio signal measured during a second period also having the first duration, the second period extending forwards in time from the sampling point;
• and
• select, as the onset, the sample that is maximally different.