Completely rewrote the tabbing algorithm. Now uses librosa's onset detection with tweaked parameters to perform onset detection, rather than the custom-build algorithm used previously. The new algorithm works much better at correctly detecting notes and not getting false positives (even with my crappy laptop mic!). Now takes lick tuning into account when generating tabs. The webapp code has been completely refactored to be much more sensible and readable. Concerns are much more clearly separated between files. Types are more robust than before.
Completely rewrote the tabbing algorithm. Now uses librosa's onset detection with tweaked parameters to perform onset detection, rather than the custom-build algorithm used previously. The new algorithm works much better at correctly detecting notes and not getting false positives (even with my crappy laptop mic!). Now takes lick tuning into account when generating tabs. The webapp code has been completely refactored to be much more sensible and readable. Concerns are much more clearly separated between files. Types are more robust than before.
Closes #57 and #30.
Relates to #58.