Continuity illusion examples
Try listening to this sentence; what do you hear?
Demo 1
You probably heard a sentence with a cough part-way through it. What you may not have realised is that part of the sentence was actually missing, and had been replaced by the cough. Have a listen again; can you tell which sound was replaced?
When we listen in noisy situations it is quite common for portions of a sound to be briefly masked by other sounds. The brain has to “fill in” the missing information as best it can. When it does so we are often not aware that this has happened, and “hear” the portion of the sound that has been masked. The example you have just heard demonstrates this fact; instead of part of the sentence being masked, it is completely replaced by the cough. (Incidentally, the missing bit is the “s” in "legislature”).
This “continuity illusion” doesn’t just happen with speech, as illustrated in sound “a)” in the graph below. The graph shows a tone that wobbles up and down in frequency, with the middle portion replaced by a noise (grey rectangle). Have a listen:
Demo 2
There are several interesting things about this illusion:
(i) You don’t just hear the tone continue, but it continues to wobble. There is nothing in the noise burst that “contains” a wobbling tone, any more than it “contains” a steady tone.
(ii) In the sound you have just heard, the wobble continues in the same way after the noise as it “would have” if it had really continued. But if we flip the wobble after the noise (so the tone goes up when it should really have gone down (sound b), and vice versa) we still hear the illusion, and don’t notice that something funny has happened:
Demo 3
This means that you’re brain has filled in the wobble, but has thrown away information about the up-and-down frequency changes that make the wobble!
The illusion doesn’t just depend on the part of the tone that occurs before the noise, but also on what happens afterwards. In other words, the continuity illusion acts “backwards in time”, as illustrated in the sound illustrated in the next graph:
We call this sound “elephant noise”, because the frequency-versus-time graph of the sound looks a bit like the body and the trunk of an elephant. (Also, we don’t get to go to the zoo very often). The body sounds like a white noise and the trunk sounds like a whistle. We’ve found that listeners hear the whistle to start earlier, and to have a longer duration, than when a similar trunk is presented later on, without the “body”. This is quite striking, because there is nothing in the sound to suggest that the whistle has started, until the noise has ended. Have a listen to this sound, in which a whole “elephant noise” is followed by an isolated “trunk”, and decide which whistling sound lasts longer:
Demo 4
Illusions like the ones you’ve just listened to can tell us a lot about how we hear. The brain does not just process sound “one bit at a time”, but encodes higher-order features such as the wobble, and can even act “backwards in time”, so that listeners hear the trunk in the elephant noise start before it really does.
If you have not already done so, why not click here to find out how the brain determines which sounds belong with one another?
This material was brought to you by Alex Billig and Bob Carlyon of the MRC Cognition and Brain Sciences Unit, Cambridge, UK.
Try listening to this sentence; what do you hear?
Demo 1
You probably heard a sentence with a cough part-way through it. What you may not have realised is that part of the sentence was actually missing, and had been replaced by the cough. Have a listen again; can you tell which sound was replaced?
When we listen in noisy situations it is quite common for portions of a sound to be briefly masked by other sounds. The brain has to “fill in” the missing information as best it can. When it does so we are often not aware that this has happened, and “hear” the portion of the sound that has been masked. The example you have just heard demonstrates this fact; instead of part of the sentence being masked, it is completely replaced by the cough. (Incidentally, the missing bit is the “s” in "legislature”).
This “continuity illusion” doesn’t just happen with speech, as illustrated in sound “a)” in the graph below. The graph shows a tone that wobbles up and down in frequency, with the middle portion replaced by a noise (grey rectangle). Have a listen:
Demo 2
There are several interesting things about this illusion:
(i) You don’t just hear the tone continue, but it continues to wobble. There is nothing in the noise burst that “contains” a wobbling tone, any more than it “contains” a steady tone.
(ii) In the sound you have just heard, the wobble continues in the same way after the noise as it “would have” if it had really continued. But if we flip the wobble after the noise (so the tone goes up when it should really have gone down (sound b), and vice versa) we still hear the illusion, and don’t notice that something funny has happened:
Demo 3
This means that you’re brain has filled in the wobble, but has thrown away information about the up-and-down frequency changes that make the wobble!
The illusion doesn’t just depend on the part of the tone that occurs before the noise, but also on what happens afterwards. In other words, the continuity illusion acts “backwards in time”, as illustrated in the sound illustrated in the next graph:
We call this sound “elephant noise”, because the frequency-versus-time graph of the sound looks a bit like the body and the trunk of an elephant. (Also, we don’t get to go to the zoo very often). The body sounds like a white noise and the trunk sounds like a whistle. We’ve found that listeners hear the whistle to start earlier, and to have a longer duration, than when a similar trunk is presented later on, without the “body”. This is quite striking, because there is nothing in the sound to suggest that the whistle has started, until the noise has ended. Have a listen to this sound, in which a whole “elephant noise” is followed by an isolated “trunk”, and decide which whistling sound lasts longer:
Demo 4
Illusions like the ones you’ve just listened to can tell us a lot about how we hear. The brain does not just process sound “one bit at a time”, but encodes higher-order features such as the wobble, and can even act “backwards in time”, so that listeners hear the trunk in the elephant noise start before it really does.
If you have not already done so, why not click here to find out how the brain determines which sounds belong with one another?
This material was brought to you by Alex Billig and Bob Carlyon of the MRC Cognition and Brain Sciences Unit, Cambridge, UK.