I know some of the folks involved in this effort, I think some additional insight might help here.
Very broadly, the work and algorithms behind WolframTones attempt to transpose the mathematical qualities of patterns into an auditory medium. A pattern (generated algorithmically) is analyzed and the derivative of the pattern determined; this yields a shorthand way of describing how the pattern changes over time. This change information is then mapped onto a musical scale as a way of perceiving the data audibly, much as a graph is used to perceive data visually. MIDI is the conduit used to convert mathematical values into notes.
Once the method of mapping patterns to notes is established, several possibilities arise. WolframTones allows you to generate new patterns that produce different kinds of audible sequences - yes primitive sounding now, but interesting in that it's essentially music based on mathematical noise mapped to scales. Change the scale from a western to a non-western variety, or change the mapping transfer function, and the style of music can be changed. Could be rather impressive down the road once people refine the process and learn more of what it can do. Early synth work was very primitive also.
But perhaps most interesting, once a library of pattern-to-music transformations is created they can be used as a tool to analyze unknown patterns. Feed an incoming data stream into the system and listen to what comes out; tweak the algorithmic properties to focus. Another auditory method of perceiving patterns and behavior may someday become very useful in some fields. Data analysis by the blind, perhaps?
A few folks felt this was a concept worth mucking around with; you can view this product as the first result.
Slashdot Mirror
I know some of the folks involved in this effort, I think some additional insight might help here.
Very broadly, the work and algorithms behind WolframTones attempt to transpose the mathematical qualities of patterns into an auditory medium. A pattern (generated algorithmically) is analyzed and the derivative of the pattern determined; this yields a shorthand way of describing how the pattern changes over time. This change information is then mapped onto a musical scale as a way of perceiving the data audibly, much as a graph is used to perceive data visually. MIDI is the conduit used to convert mathematical values into notes.
Once the method of mapping patterns to notes is established, several possibilities arise. WolframTones allows you to generate new patterns that produce different kinds of audible sequences - yes primitive sounding now, but interesting in that it's essentially music based on mathematical noise mapped to scales. Change the scale from a western to a non-western variety, or change the mapping transfer function, and the style of music can be changed. Could be rather impressive down the road once people refine the process and learn more of what it can do. Early synth work was very primitive also.
But perhaps most interesting, once a library of pattern-to-music transformations is created they can be used as a tool to analyze unknown patterns. Feed an incoming data stream into the system and listen to what comes out; tweak the algorithmic properties to focus. Another auditory method of perceiving patterns and behavior may someday become very useful in some fields. Data analysis by the blind, perhaps?
A few folks felt this was a concept worth mucking around with; you can view this product as the first result.