Semantic context detection based on hierarchical audio models