You are hereProjects/ GAPR - General Audio Pattern Recognizer

Projects/ GAPR - General Audio Pattern Recognizer


GAPR set-up Dolphins cannot make human sounds, and humans cannot make dolphin sounds.  This is mainly because of the ultrasonic frequencies present in dolphin sounds, which are far beyond our upper hearing capability (see Introduction to dolphin acoustics).  In order to establish some form of two-way communication between humans and dolphins, we need to find a way to bridge this gap.

The “General Audio Pattern Recognizer”, or GAPR (pronounced “Gapper”, as in someone bridging a gap), is a piece of software developed by Preben Wik.  It will allow humans to make and recognize dolphin sounds – that is, once we find appropriate dolphin sounds that can be grouped in distinct classes.  This will then allow dolphins to easily perceive, distinguish and copy these sounds (something which would be very hard to do with, for example, English words) – after all, the arbitrary sounds they are asked to reproduce are already part of their own vocal repertoire.

In general, GAPR is a tool for classifying and translating arbitrary audio patterns (be it dolphin sounds or any other sounds, for that matter) into symbolic representations, and for linking combinations of these symbols to words.  GAPR can be used both for setting up an artificial language structure, and to run a real-time translation for dialog or interactive games. You can read more about the GAPR here.