Masking plays an important role in the measurement of the quality of audio devices. Masking is the result of smearing in the time-frequency domain and compression in the amplitude domain. This paper determines the optimal psychoacoustic representation by matching the time-frequency smearing and amplitude compression to databases that contain judgments on the quality of audio devices. The optimal time-frequency smearing that is found is in line with psychoacoustic data for music codecs but in contradiction to psychoacoustic data for speech codecs.
This paper costs $33 for non-members and is free for AES members and E-Library subscribers.