Digital studios trace a great amount of processes and artifacts. The important flow of these traces calls for a system to support their interpretation and understanding. We present our design and development of such a system in the digital music production context, within the gamelan research project. This trace-based system is structured in three main layers: track production process, interpret collected traces according to a dedicated domain ontology, help querying and visualizing to foster production understanding. We conclude by discussing some hypotheses about trace-based knowledge engineering and digital music production understanding.