I don't quite like the concept, myself being - or trying desperatly to be - in the music biz;
saying that this "machine" does math and heavy statistical comparisons on bpm, pitch, instruments, lenghts, accoustic factors, etc, to recognize some patterns that were succesfull, it must do heave datamining on large databases of Known Success and their proprietary attributes.
What I want to say is that it compares the "new song" over many fields to already known hits of the industry. A new song with particularities alike those of old hits will probably score a good mark since it is has a potential of success according to the stats; but this system, I believe, doesn't yield much room for any new and creative original artists that are doing very new stuff, yet unknown to the public.
For instance, go wayyyy back in the days, lets say a good 30years back; use this system to compare [almost] any Pink Floyd's song or album and predict the success.... wonder how 'Meddle' would have got...
Don't get me wrong, I'm an higly scientific mind, I'm in electrical engineering and I love music and would like to do projects on it; and I find this one very interesting; but to use it as a contract-signing decisional 'tool', I must say I'm afraid for the future of the industry.
Slashdot Mirror
I don't quite like the concept, myself being - or trying desperatly to be - in the music biz; saying that this "machine" does math and heavy statistical comparisons on bpm, pitch, instruments, lenghts, accoustic factors, etc, to recognize some patterns that were succesfull, it must do heave datamining on large databases of Known Success and their proprietary attributes. What I want to say is that it compares the "new song" over many fields to already known hits of the industry. A new song with particularities alike those of old hits will probably score a good mark since it is has a potential of success according to the stats; but this system, I believe, doesn't yield much room for any new and creative original artists that are doing very new stuff, yet unknown to the public. For instance, go wayyyy back in the days, lets say a good 30years back; use this system to compare [almost] any Pink Floyd's song or album and predict the success.... wonder how 'Meddle' would have got... Don't get me wrong, I'm an higly scientific mind, I'm in electrical engineering and I love music and would like to do projects on it; and I find this one very interesting; but to use it as a contract-signing decisional 'tool', I must say I'm afraid for the future of the industry.