applies the sine and then the inverse sine1. The two are inverse operations, so one might expect back, but the range of is confined to , and whatever falls outside is folded back, producing a triangle wave.
| Interval | Formula | Slope |
|---|---|---|
On the first interval and cancel exactly. On the next the slope turns to , and the pattern repeats every .
Both and are odd, so their composition is odd too and the graph has point symmetry about the origin.
The graph is a polyline joining segments of slope and alternately. The vertices sit at , where the value is . There the slopes from the two sides are and , so the function is not differentiable.
Continuous yet not smooth at those points, it is a different creature from the smooth wave of . The zeros are at .
Computing the derivative produces a square wave.
So the derivative of a triangle wave is a square wave, and conversely integrating a square wave that alternates between and produces a triangle wave. Integrating a function with jumps yields a function with corners, and that correspondence is visible here.
Only odd multiples of the frequency appear, and the coefficients decay as .
| Waveform | Smoothness | Decay of the coefficients |
|---|---|---|
| Square wave | has jumps | |
| Triangle wave | has corners |
The fall is quicker than for a square wave, so fewer terms give a good approximation. The general rule that integrating speeds up the decay of the coefficients by one step shows through here.
The triangle wave is one of the basic waveforms of a synthesizer, softer than a square wave and brighter than a sine, because its higher harmonics fall off as . In electronics it is also compared against a constant voltage to generate a pulse-width-modulated signal.