Evaluation of signal representations for predicting orchestral effects

The second subtask aims to evaluate computational models of timbre representations, applied to multi-track audio renderings of orchestral music.