Demonstration 38. Masking Level Differences (2:03)

A stepwise decreasing 500Hz tone is applied to the left ear. Count the number of steps you can hear. Staircases are presented twice

500 hz in left

Now the signal is maked with noise

Signal Left noise mask left

Next the same masking noise is applied to both ears

Signal left Noise mask both

Both signal and noise appear in both ears

Signal in both Noise in both

Finally the signal and noise appear in both ears but the signal phase is reversed in one of the ears.

Signals out of phase

Besides enabling us to localize sound sources, binaural hearing helps us to receive auditory signals in noisy environments. Licklider (1948) discovered the phenomen called "masking level difference" for speech; Hirsh (1948) demonstrated it for sinusoidal signals, as in this demonstration.

The first example in the demonstration begins by playing a 500-Hz signal of 100 ms duration in the left ear. Then this signal is masked by noise. The staircase procedure is used, so that you can count the level at which the signal becomes inaudible. The staircase contains 10 steps; the first step is -10 dB, and the remaining 9 steps are -3 dB each.

In the third example, noise is added to the other ear. The noise now appears have a different spatial location (albeit inside the head) from the signal, and the signal is much easier to hear. In the fourth example, the signal is again made difficult to hear by adding a signal to the noise in the right ear, so that the auditory images of signal and noise again coincide.

The fifth example demonstrates that inverting the signal at one of the ears places the noise and signal in different auditory locations. In this configuration, the signal easy to hear.

The signal, which is similar to No. 11 in the "Harvard tapes" can be summarized as follows:

1) S[L]; 2) S[L]N[L]; 3) S[L]N[LR]; 4) S[LR]N[LR]; 5) S[L]S2[r]N[LR], where S = signal, S2 = signal of reversed phase, N = noise, L = left, R = right. The signal is me easily heard in examples 1, 3, and 5. In the original Harvard tapes, the demonstration is repeated at 2000 Hz, where the effect has faded, thus demonstrating the strong frequency dependence of masking level difference (Hirsh, 1948).