The scenario is a small platformer with a single player character, three jump types — standard, double, and wall — and a need for the jump to read differently depending on the character’s weight and the scene’s density. The project has no recorded foley, no sound designer, and one afternoon for audio. The goal is not five unrelated sounds. It is one base sound and four derived variations that share enough structure to read as the same character, and enough difference to read as different events.
The case study below follows the derivation in order. It starts with the base sound, then shows the specific parameter change that produces each variation, and then explains what has to stay constant across all five. The parameters are not a specification; they are the decisions that were made in this project, with the reasoning that made them.
The Base: 400 → 800 Hz, 80 ms Sweep, 120 ms Total, 50% Square
The base jump is the documented retro recipe: a square wave with a pitch sweep from 400 Hz to 800 Hz over 80 ms, a total duration of 120 ms, and a 50% duty cycle. That is one octave of upward sweep, which is the range that reads as a jump rather than as a blip or a power-up, and it sits in the band where a laptop or phone speaker still reproduces the sweep.
The square wave and pitch sweep guide covers the oscillator and envelope in more detail. For this case study, the base is fixed: every variation below starts from these parameters and changes exactly one axis.
Variation 1: Heavy — Same Contour, Lower Start, Longer Body
The first variation is for a character that should feel heavier than the base. The mistake would be to lower the pitch without changing anything else, which produces a slower-sounding jump but not a heavier one. The change that actually shifts the perceived weight is to lower the starting frequency while keeping the sweep span roughly the same, and to extend the body of the sound slightly.
The parameters for this variation are: start 300 Hz, end 600 Hz, sweep duration 90 ms, total duration 160 ms, and a low-mid boost around 250–400 Hz. The sweep is still one octave, so the contour still reads as a jump, but the lower register and the longer body make it read as a heavier character. The heavy character jump guide covers the full rationale; the case-study point is that only three parameters changed.
The risk with this variation is that lowering the start too far pushes it below the band a phone speaker reproduces, and the jump loses its sweep on small devices. The check is the same one the base uses: high-pass at 400 Hz and confirm the upward motion is still audible. If it is not, the start has gone too low.
Variation 2: Quick — Higher Start, Shorter Body, Thinner Duty Cycle
The second variation is for a character or a scene where the jump needs to read faster. The base is already short, so the change is not simply to shorten it further; the change is to shift the register up and to thin the waveform so the sound cuts rather than lands.
The parameters are: start 500 Hz, end 1000 Hz, sweep duration 60 ms, total duration 90 ms, and a 25% duty cycle instead of 50%. The higher register reads as lighter and quicker, and the narrower duty cycle removes the fuller body that the 50% square has. The result is a jump that is recognisably the same event as the base but reads as a faster, smaller character.
The risk here is the opposite of the heavy variant: raising the register too far turns the jump into a laser or a UI blip. The pitch variation guide covers the identity threshold; the case-study point is that the sweep span, not the absolute pitch, is what keeps the sound a jump. One octave at 500–1000 Hz is still one octave, and it still reads as upward motion.
Variation 3: Double Jump — Same Start, Two-Stage Contour, Shorter Total
The third variation is a double jump. The mistake would be to reuse the base jump with a higher pitch. The change that actually communicates a second resource is a two-stage contour: the pitch rises, plateaus briefly, and rises again, or rises and then falls slightly at the end. The base is a single upward sweep; the double jump is a compound shape.
The parameters are: start 600 Hz, rise to 900 Hz over 50 ms, plateau for 20 ms, rise to 1200 Hz over 30 ms, then a very short fall back to 1000 Hz in the final 15 ms. Total duration is 115 ms, slightly shorter than the base. The added layer is a short high shimmer at −8 dB that follows the second rise, which gives the sound a “second push” character without making it a different event.
The double jump guide covers the contour design in more detail; the case-study point is that the double jump shares the base’s starting register and sweep span but changes the shape of the contour, which is the one axis that distinguishes a second resource from a louder first one.
Variation 4: Wall Jump — Base Minus Low End, Plus Impact and Whoosh
The fourth variation is a wall jump, and it is the only one in this set that changes the structure rather than a parameter. A wall jump is not a single sweep; it is an impact against a surface followed by a launch away from it, and the sound has to carry both.
The derivation from the base is to take the base jump and remove its low-mid body, then add two new elements. The first is a short filtered noise burst at the onset — the impact against the wall — with a duration of 40–60 ms and a low-pass around 500 Hz. The second is a panned whoosh that follows the impact by 10–20 ms, with a spectral sweep from roughly 800 Hz to 2 kHz over 100 ms, panned from the wall normal. The base’s upward sweep is kept as a thin layer underneath, but it is no longer the identity of the sound.
The wall jump guide covers the direction and effort cues in detail; the case-study point is that the wall jump is the variation that justifies its difference by changing the event’s structure, not just its parameters.
What Stays Constant Across All Five
The five variations differ along four axes — weight, speed, contour shape, and event structure — but they share three properties that make them read as the same character’s jump family.
The trigger placement. All five fire on the jump’s execution frame, not on the input frame. A wall jump that fires on input and a standard jump that fires on execution will feel like two different characters, because the timing relationship is different. The Unity 2D jump setup guide covers the execution-versus-input distinction; the case-study point is that the shared timing is what makes the family legible.
The runtime pitch variation. All five use the same per-trigger pitch offset of roughly ±1 to ±2 semitones, applied at playback. The offset is what keeps any single jump from repeating exactly, and using the same range across the family keeps the variation consistent without making the heavy jump sometimes sound like the quick one.
The voice pool. All five route to the same concurrency limit, typically 3–4 simultaneous jump instances. Without a shared limit, a rapid sequence of standard jumps can exhaust the pool before a wall jump fires, and the wall jump will be the one that gets cut.
The Five Variations Side by Side
The table below is the case study’s summary. It is not a parameter specification; it is the set of decisions that distinguish each variation from the base, with the axis each one changes.
| Variation | Axis changed | Key parameter change |
|---|---|---|
| Base | — | 400 → 800 Hz, 80 ms sweep, 120 ms total, 50% square |
| Heavy | Weight | 300 → 600 Hz, 90 ms sweep, 160 ms total, low-mid boost at 250–400 Hz |
| Quick | Speed | 500 → 1000 Hz, 60 ms sweep, 90 ms total, 25% duty cycle |
| Double | Contour | 600 → 900 → 1200 → 1000 Hz, 115 ms total, added shimmer at −8 dB |
| Wall | Structure | Base minus low end, plus impact (40–60 ms) and panned whoosh (100 ms) at 10–20 ms offset |
The case study shows that variation is not a matter of making five unrelated sounds. It is a matter of changing exactly one axis per variation and keeping the rest of the family constant. The five variations here share a trigger, a pitch variation range, and a voice pool; they differ in weight, speed, contour, and structure. That is enough to distinguish four events without producing four characters. The sound variation system guide covers the runtime side; the case-study point is that the derivation happens at the authoring stage, before any of the runtime variation is applied.