(b) Solo: instrument dataset.
(d) Backing: instrument dataset.
Alg1 obtained a slightly higher IPS score for the backing tracks with the instrument dataset.
For both algorithms, scores obtained for the instrument dataset are higher than the ones obtained with the voice dataset.
However, it should also be noted that larger confidence intervals are also obtained for the instrument dataset, suggesting that some instruments are better handled than others.
During the S-WPIA processing, when the time tag of a particle-event data exceeds the end time of a wave dataset, it then shifts to the processing of the next particle instrument dataset.
Figure 6a,b shows the scores for the solo tracks for the voice and instrument datasets, respectively.
Similarly, Figure 6c,d shows the scores for the backing tracks for the voice and instrument datasets, respectively.
Their modular and flexible nature allows easy adaptation to other instruments and datasets.
The IUGONET metadata (MD) are written in XML format, and the style (or format) is defined for individual keywords (Observatory, Instrument, Person, Dataset, and Granule to reach each specific data file), in accordance with the SPASE ontology.
