It seems to me that algorithms for identifying similar sounds would be likely to find an lot of pretty generic noises similar to a sample of wind -- other wind, distant traffic, etc etc.
As such, BFM giving such a sample to Youtube's automated infringement detector strikes me as copyright trolling -- it's inevitably gonna get lots of false positives, and they get to make a bit of money off each before the YT appeals process works itself out.
As such, BFM giving such a sample to Youtube's automated infringement detector strikes me as copyright trolling -- it's inevitably gonna get lots of false positives, and they get to make a bit of money off each before the YT appeals process works itself out.