Firstly thank you to all the authors for making this available. Your work is still very much appreciated!
In working with the dataset we identified two alignments where all the notes appeared to be aligned to the second note onset, instead of the first note onset. This introduces a constant delay in the annotations rendering them inaccurate. The files are:
04_BN3-154-E_comp (off by + 0.409s)
04_Jazz1-200-B_comp (off by + 0.309s)
While it's possible to fix the annotations by subtracting those constants from all the timings, it would make fair comparison with current published results more challenging. Perhaps you could just leave a note somewhere on the Zenodo page? People might want to exclude those files from their training data for example.
Firstly thank you to all the authors for making this available. Your work is still very much appreciated!
In working with the dataset we identified two alignments where all the notes appeared to be aligned to the second note onset, instead of the first note onset. This introduces a constant delay in the annotations rendering them inaccurate. The files are:
04_BN3-154-E_comp
(off by + 0.409s)04_Jazz1-200-B_comp
(off by + 0.309s)While it's possible to fix the annotations by subtracting those constants from all the timings, it would make fair comparison with current published results more challenging. Perhaps you could just leave a note somewhere on the Zenodo page? People might want to exclude those files from their training data for example.