What is the maximum frequency jump between two increasing tones that could be perceived as continuous?
by Jeremy Ma
Insipired by the shepard tones, I was intrigued about the continuity of tones that increase in frequency.
In their paper, Warren et. al (1982) showed that perceptual filling in occurs with a discontinuous tone if the gap is masked with noise. Although it is also unclear whether perceptual filling in occurs in a tone that constantly and consistently increases in frequency, there should still be some degree of continuity between two of such tones. This study aims to investigate this the limits of this continuity by finding the maximum frequency jump that could be perceived as continuous, or in other words, the max gap in which these two tones would be grouped as one tone.
Keep in mind the question isn't whether 'if one can detect if the two tones are consistently increasing', as that depends on one's individual ability to distinguish tones; but rather 'if one perceives the two tones as one'.
The illusion consists of two logirthimic chirps that both each grow at a steady rate, the first chirp growing from 200hz to 1600hz over 4 seconds, but only the first 2 seconds was used. The beginning of the second chirp is set to be N steps above where the first chirp should be at the same time, and the growth rate of the second chirp is the same as the first.
In order to induce this sense of continuity, we can mimic the shepard tones by using an amplitude envelope to create a 'gap' in the recording. We can do this by linearly decaying chirp1 to the 2 second mark, adding a silence of 125ms, then increasing the amplitude of chirp2.
R. Warren, Obusek, and Ackroff. (1982). "Auditory Induction: Perceptual Synthesis of Absent Sounds"
Smooth original
smooth with fade
1 step
1.5 step
2 steps
3 steps
6 steps
Comments
Jeremy Ma
c) I would expect that a >1.5 step difference would be around the max gap that our brain can fill in, so yes it does answer the question proposed.
d) I actually originally had this illusion with noise in the middle, but I found that perceptual filling doesn't really occur when there is a noise, possibly because of the continuously changing nature of the tone vs the static noise. I found it easier to fill in the gap with gradually amplitude decay. In the future, I could decrease the gap of 0 amplitude to help connect the two tones.
0
Cesar Duran
Interesting Experiment
While I agree with Griffin in that it's difficult to connect the two increasing tones, it was easier to convince myself that the tones could be perceived as one if I mentally fill in the gap to match the rising pattern. With this in mind, I found it possible to believe that the two tones were part of a continuous singular tone until there was a >1.5 step difference. With that big of a gap, it was difficult to mentally imagine a tone that would fit the gap.
This illusion roughly addresses the question presented, however it may be more beneficial to use Gaussian white noise versus silence to allow for easier filling in.
0
Griffin Leonard
a) I perceived every clip as two different increasing tones separated by a gap, except for the original (which I perceived as a single, continuous tone).
b) This illusion demonstrates that simply having an increasing tone (without noise in the background) is not enough to cause perceptual filling-in of a gap in the tone. This seems to answer the question "can two increasing tones separated by silence be perceived as continuous?" (to which the answer seems to be no,) as opposed to "what is the maximum frequency jump between two increasing tones that could be perceived as continuous?". Perhaps if the gaps of silence were removed or if noise was added, the original question could be addressed.
