Repeated sequences of tones can be heard as belonging to a single perceptual organisation (fusion) or to two groupings (fission). The frequency difference between successive tones plays a key role in determining which organisation prevails, as first demonstrated by Miller & Heise (1950) using alternating pairs of tones with increasing frequency separation. The point at which the sequence 'breaks' into two is termed the 'trill threshold'.
Van Noorden (1975) investigated the role of other factors such as presentation rate in streaming. In one condition, he used repeated HLH_ stimuli, where H and L represent high and low frequency tones, and _ indicates a silent interval of the same duration as the H and L tones. This stimulus uses rhythmic change as an indicator of fission or fusion. If the low and high tones fuse, listeners should hear a galloping (HLH_) rhythm. However, if the L and H tones segregate, two isochronous rhythms should appear. One, formed by the H tones, is twice as fast as the other, formed by the L tones.
Van Noorden discovered that some conditions produced an 'ambiguous' percept: i.e. listeners could choose to hear fusion or fission, and switch between these percepts.
Launch the demonstration with the command 'streamer'. Clicking anywhere in region 1 results in the delivery of a stimulus with the tone repetition time and high-low frequency separation specified by the position in the grid. The number of cycles delivered is governed by a menu option (4). Ten or more cycles may be required to hear one or other organisation.
After some practice at hearing the two rhythmic possibilities, you may wish to record your responses. After each stimulus presentation, pressing 's' or 'f' will result in a red circle (2) or green cross (3) appearing at the click location. After a number of such identifications, compare your responses with those summarising van Noorden's subjects by checking the checkbox (5).
- Miller & Heise (1950). JASA, 22, 637-638.
- Van Noorden (1975). PhD Thesis, Eindhoven University of Technology.
- Bregman's book (Auditory Scene Analysis, MIT Press, 1990) contains an extensive discussion of auditory stream segregation.
- A number of computational models which attempt to explain fusion/fission have appeared lately:
- Beauvois & Meddis (1995). JASA
- Brown & Cooke (1998). Ch 7 in Computational Auditory Scene Analysis (ed: Okuno & Rosenthal, LEA).
- McCabe & Denham (1997). JASA, 101(3), 1611-1621.
- Todd (1996). Network: Computation in Neural Systems, 7, 349-356.
- Various theories have been put forward to explain these and other streaming effects. Some of these are discussed in:
- Rogers & Bregman (1993). Perc. & Psych., 53(2), 179-189.
- Hartmann & Johnson (1991). Music Perception, 9(2), 155-184.
Produced by: Martin Cooke
Release date: June 22 1998
Permissions: This demonstration may be used and modified freely by anyone. It may be distributed in unmodified form.