Temporal masking describes the effect of your ears missing quiet sounds after hearing a loud sound. This effect is probably known by most of you (temporal post-masking), but did you know that this also happens before you hear a loud sound (temporal pre-masking)? How can your brain know when hearing a quiet sound that there will be a loud one coming and dismiss it beforehand? Magic? Can our brains peek into the future?
Premasking happens around 50ms before the loud sound, so you won’t hear anything up to 50ms (or maybe even more?) before a loud sound appears. But how can the brain know when it hears a quiet sound that there will be a louder one and ignore the quiet one? I learned this fact first in Prof. Möller’s lecture “Speech Communication” last week, but he couldn’t explain how it is possible. He suggested that it might be the time the brain needs to process the sounds. But the effect also happens for quiet sounds right before (<1ms) the loud sound, so this is not an explanation.
Well, today, I was in the lecture “Communication Acoustics” and we discussed this issue. As it was going nowhere again, I though of a model that could explain this effect.
First, you need to know that your ear works just like an ADU (analog-digital-unit), which sends out a signal to your brain whenever it receives a sound with a specific frequency. Louder sounds makes the ear send out more signals. Because signals are infinitely short, so to let you hear a continous sound, your brain counts the amount of signals over 200ms and then tells you what how loud the sound was for the last 200ms. This works just fine at normal sound volume. Imagine a child counting the signals and telling you every 200ms how much signals he has counted. Now what if there is a really loud sound? Well, the loud sound will cause a lot of signals need to be counted, let’s say 1000. But what if the child counting the signals could only count to 100? He will tell you as soon as he has already counted to 100, before the usual 200ms are over (buffer overflow)! Now you instantly react to the loud voice without much delay.
But what if there was a quiet sound 50ms (let’s say 5 signals) before the loud sound? The child counted the 5 signals and waits for 200ms more seconds to tell you, but after 50ms, there is the loud sound (1000 signals) and he counts 95 from it, so he reached 100 and tells you that 100 have been reached. Now you know the loud sound happened (right when the child told you) but the information of the 5 signals from the quiet sound are lost.
This is how I think premasking works. As far as I know there is no scientific explanation for this effect yet, but feel free to prove me wrong.