The question “how long should a jump sound be” usually gets answered with a single range — somewhere between 100 and 300 milliseconds — and that answer is not wrong, but it is not useful. A short hop and a full jump are not the same event with different lengths. They are different timing problems, and the duration that makes one read correctly can make the other feel wrong.
What follows is not a single recommendation. It is a set of benchmarks by jump type, each tied to the mechanism that makes that type distinct. The numbers come from documented implementations and published sound libraries, not from a preference for round figures.
The Benchmark That Applies to Everything: 50–300 ms Is the SFX Envelope, Not the Jump Envelope
Before the type-specific numbers, there is a general constraint that applies to any short game sound. A retro SFX guide states it plainly: “Music loops for minutes; SFX last 50–300 ms” . That range is not a target for every jump. It is the window within which a sound is perceived as an event rather than as a texture.
A jump that runs longer than 300 ms starts to overlap with the next event in a platformer’s input rhythm. A jump that runs shorter than 50 ms is perceived as a click, not as a launch. The type-specific benchmarks below all sit inside this window, but they occupy different parts of it, because each jump type has a different amount of time before the next event arrives.
Short Hop: 80–150 ms, Because the Next Input Is Already Coming
A short hop is a low-commitment jump. The player taps the button, the character leaves the ground briefly, and the player is likely to jump again or change direction almost immediately. The sound has to confirm the input and get out of the way before the next input arrives.
The published recipes converge on a short range. A retro SFX guide gives a jump duration of 100–150 ms, with a pitch sweep of 400 Hz → 800 Hz over 80 ms . A Japanese sound-effect reference for small hops lists approximately 0.12 seconds for a light small jump and approximately 0.20 seconds for a slightly longer small jump . The 8bit-sfx library includes jump presets as short as 0.154 seconds and 0.158 seconds, with “snappy tail” variants clustered in the 0.176–0.227 ms range .
The short hop benchmark is therefore 80–150 ms for the core event, with a hard ceiling around 200 ms. The ceiling is not about the sound’s character; it is about the input rhythm. A hop that is still decaying when the player presses jump again will fuse with the second jump and read as a single broken event. The short hop versus full jump guide covers how to weight the layers for this range; the duration benchmark is the constraint that weighting has to fit inside.
Full Jump: 200–400 ms, or 1 Second If You Separate Impulsion From Motion
A full jump is a commitment. The character will be in the air long enough that the player needs to feel the arc, not just the launch. The duration has to cover more of the jump’s travel.
There are two ways to achieve that, and they produce different benchmarks. The single-layer approach extends the sound: published guides put jumps at 100–400 ms , with the upper end used for jumps that need to feel longer. The 8bit-sfx library’s “floaty tail” jumps run from 0.314 to 0.347 seconds, which sits at the top of that range .
The two-layer approach separates the event into an impulsion and a motion. CRI Middleware’s documented implementation gives the impulsion a duration of 75–100 ms and the motion a duration of at least 1 second . That is not a contradiction of the 400 ms ceiling. It is a different structure: the impulsion is the event, and the motion is a sustained layer that follows it. The combined cue can be longer than 400 ms because the motion layer is not competing with the next jump in the same way a single extended event would.
The full jump benchmark is therefore 200–400 ms for a single-layer sound, or 75–100 ms impulsion + 1 s motion for a layered implementation. The choice depends on whether the game’s jump has a visible arc that the sound needs to accompany, or whether the jump is a discrete event that ends when the character leaves the ground.
Double Jump: 250–350 ms, Because It Fires in the Middle of an Existing Arc
A double jump does not start from rest. It fires while the character is already moving, often at the apex of the first jump or during a fall. The sound has to register as a second event without competing with the first jump’s tail.
Published double-jump assets cluster around 340 ms. A free double-jump sound is listed at 340 ms, described as “two notes rising playfully” . A Japanese reference for double jumps gives approximately 0.36 seconds . The 8bit-sfx library does not distinguish double jumps from first jumps, but its longer “floaty tail” presets at 0.324– 0.347 seconds are consistent with the double-jump range .
The double-jump benchmark is therefore 250–350 ms, slightly longer than a short hop and slightly shorter than a full jump’s upper range. The reason is structural: a double jump has to be clearly a second event, but it cannot be so long that it collides with the landing that may follow quickly. The double jump design guide covers how to make the second jump read as a different resource; the duration benchmark is the window that different-resource sound has to fit inside.
Wall Jump: 300–500 ms Total, Split Between Impact and Whoosh
A wall jump is a composite event: the push against the wall, then the launch away from it. The total duration is longer than a standard jump because it contains two distinct sub-events, but neither sub-event is individually long.
The documented wall-jump implementation in a multiplayer parkour game describes the sound as an impact (3 transients) plus a short whoosh . The impact is the push; the whoosh is the launch. A separate audio requirement describes the impact as a “distinct thud” . Neither source gives a millisecond value, but the structure implies that the impact is on the order of 50–100 ms and the whoosh is on the order of 100–200 ms, with a short gap between them.
A working total for a wall jump is 300–500 ms from the start of the impact to the end of the whoosh. The exact split depends on whether the game shows a visible contact with the wall: a longer impact suggests more contact time, and a shorter impact suggests a quick push. The wall jump design guide covers the direction and effort cues; the duration benchmark is the envelope those cues have to fit inside.
The Full Benchmark Table, With Sources
The table below consolidates the benchmarks and the implementations they come from. It is not a specification. It is a starting point for each jump type, with the caveat that the “full jump” row has two valid structures.
| Jump type | Duration benchmark | Where the number comes from |
|---|---|---|
| Short hop | 80–150 ms (ceiling ~200 ms) | Retro SFX recipe: 100–150 ms; Japanese small-jump reference: 0.12–0.20 s; 8bit-sfx snappy presets: 0.154–0.227 s |
| Full jump, single layer | 200–400 ms | Published jump range: 100–400 ms; 8bit-sfx floaty presets: 0.314–0.347 s |
| Full jump, layered | 75–100 ms impulsion + ≥1 s motion | CRI Middleware documented implementation |
| Double jump | 250–350 ms | Free double-jump asset: 340 ms; Japanese double-jump reference: ~0.36 s; 8bit-sfx long presets: 0.324–0.347 s |
| Wall jump | 300–500 ms total (impact + whoosh) | Documented wall-jump structure: impact (3 transients) + short whoosh; no single millisecond value in the source |
The wall jump row is the least precise because the source describes a structure rather than a duration. That is not a gap in the research; it is a reflection of how wall jumps are built. The impact and the whoosh are separate assets, and their individual lengths are chosen against the character’s movement, not against a universal benchmark. The jump and landing guide covers the broader movement system; the wall-jump case is the one where the total duration is a sum of two designed parts, not a single envelope.
When the Benchmark Is Wrong: Three Cases Where the Number Does Not Apply
A benchmark is a default, and defaults fail in specific situations. The three below are the ones where the published ranges are the wrong answer.
Jump sounds that carry a voice or a character grunt. If the jump includes a vocalization, the duration is set by the voice, not by the jump. A grunt that is 300 ms long will make the jump 300 ms long regardless of the platformer benchmark. The SFX duration guide covers the category ranges for voice; the jump case is that a vocalized jump is a voice event with a jump attached, not the other way around.
Jumps in a rhythm game or a music-driven platformer. If the jump fires on a beat, the duration is constrained by the beat grid. A 150 ms jump that fires on every eighth note at 120 BPM has 250 ms between onsets; a 300 ms jump would overlap the next onset. The rhythm and beat guide covers the synchronization side; the duration point is that the grid overrides the platformer benchmark.
Jumps in a game with a very slow or very fast movement speed. A character that moves slowly across the screen can afford a longer jump sound, because the next event is further away in time. A character that moves quickly needs a shorter sound, because the next event arrives sooner. The benchmark ranges are calibrated for a typical platformer pace; a game that is significantly faster or slower should shift the ranges accordingly.