We know that the inner ear of a human listener converts a time domain acoustic signal into a frequency-based representation before it encodes it into neural impulses. However, the inner ear’s frequency representation is different to that obtained from a Fourier transform. We also know that how it handles amplitude of the audio signal is also nonlinear. This nonlinearity can make the intelligibility of speech suffer when the loudness of the speech and other noises is high.

The talk should be accessible to people from all the different areas of endeavor within the ISCVE. It will first examine the operation of the ear, including its dynamic non-linear behaviour. It will then examine the difference between the Fourier Transform and the human auditory system and highlight how they trade off time and frequency resolution differently. From this we will be able to understand some of the issues in measuring sound intelligibility and why it gets worse at high signal levels. This has implications for how we test and commission voice alarm systems.

Finally, we shall look at the unique ways the human auditory system allows us to hear the incredible complexity of the audio signal and how that might affect what we do in the future to make it even better.