Hi All. I'm new to ResearchGate, please let me know if this style of question isn't appropriate. 

I'm interested in whether or not a gulf exists in the capacities of algorithms and biological neural nets to compute sound localisation in 'realistic' environments e.g. multiple sound sources present, surfaces are present off which sound can reflect. 

How well do our best algorithms perform in these scenarios? How do artificial neural nets perform in comparison? If so, is there something inherent to  neural nets that allows them to perform computations like these more efficiently than algorithms?