Neural networks (NN) are a combination of simple computational elements trained to perform specific mapping tasks between input and output data. In this case, input data were segments of audio waveforms (music and speech signals) and the required output was the auditory noise-masking threshold created by these data. Such a threshold determines the relative level of noise and distortion components which will be subjectively masked by the audio signal. It was found that an NN could perform such a task with a small error margin, but lengthy training was necessary with segments of such data.
Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member and would like to subscribe to the E-Library then Join the AES!
This paper costs $33 for non-members and is free for AES members and E-Library subscribers.