Fast alternation between two pitches splits into two independent melodic lines

What it is

Play the same short tones alternating between two pitches — high, low, high, low — slowly, and it sounds like one up-and-down melody. Speed the alternation up past a certain point, and the same two pitches suddenly "split": listeners now hear two separate, internally flat lines, one made only of the high tones and one only of the low tones, and the sense of alternation disappears. This is stream fission, the reverse of the fusion described in the sibling entry.

It is easy to mistake for a change in timbre or loudness. In fact neither pitch's physical parameters change at all throughout — the only variable is playback speed, yet the percept changes qualitatively.

Why it happens

The auditory system's ability to track a single stream has an implicit "bandwidth" limit: at a given presentation rate, successive notes can only continue to be heard as part of the same stream if the frequency jump between them stays under some bound. Slowing playback down leaves more time to "bridge" between two notes even with the same frequency gap, loosening that bound, so the two pitches can still be heard as ups and downs within one stream. Speeding playback up shortens the available bridging time, so the same frequency gap now exceeds the bound; the system finds no basis for linking them into one stream, and instead pulls out the notes that are close to each other in frequency — all the highs together, all the lows together — as separate streams, since the smaller jump is naturally the easier one to keep tracking.

This limit is usually described as a fission-fusion boundary: for a given frequency separation there is a critical rate below which fusion holds and above which fission occurs; for a given rate there is a critical frequency separation with the same effect. The two are interdependent — neither variable can be read in isolation.

Studying it

The classic paradigm systematically varies the combination of frequency separation and presentation rate, having listeners report one stream or two, tracing out the boundary curve between the fusion and fission regions. An objective measure is often used alongside subjective report: a rhythm-judgment task, where once fission has occurred, timing relations across the two pitches (e.g., whether a note is delayed) become hard to detect — a drop in accuracy signals that fission has taken place.

The main value of this paradigm for interface work is parameter calibration: it tells a designer at what pitch spacing and tempo an alternating sequence of information will, or will not, be unintentionally torn apart by the listener's own perception.

Where it stops holding

• The fission-fusion boundary is not a single fixed number. It shifts with the absolute frequency region of the pitches, whether timbre is matched across them, and total sequence duration (fission becomes more likely the longer the sequence continues) — no single rate or frequency-gap value should be treated as a universal threshold.
• The boundary describes an average tendency across listeners; individual differences are substantial, and musical training makes people more likely to pick out the two split lines.
• The effect holds specifically for strict alternation. If the two pitches are not strictly alternating but randomly interspersed or mixed with other pitches, the shape of the boundary curve changes and cannot be applied directly.

Related

• Same group: A3.15.1 Frequency-close, temporally continuous sounds fuse into one auditory stream · A3.15.3 Stream formation depends on temporal regularity; irregular gaps hinder it · A3.15.4 Simultaneous sources need onset-time differences to be separated
• Nearby: A3.13 Temporal resolution and rhythm perception (the resolution limit of presentation rate itself) · A3.06 The cocktail party effect
• Search terms: stream fission · galloping rhythm · fission-fusion boundary · van Noorden