⌖ Open the aim trainer
Online Aim Trainer / Guides / Research
Sound research

Your hit sound should not hurt: how we rebuilt ours from the research

Three of our hit sounds gave a tester headaches. The fourth worked. What noise research, startle physiology and a lot of measuring taught us about sounds you hear 500 times a session.

By Black Prince, founder of Online Aim Trainer · October 30, 2026 · 7 min read

An aim trainer is an unusual place for a sound. In a 60-second run you might fire 150 shots and hit 120 targets. Play a few runs and you have heard the same "you hit it" sound five hundred times. Whatever that sound is, your ears will tell you about it long before you consciously notice.

We learned this the hard way. While we rebuilt the gun and its sounds this week, the verdicts from play-testing came in order: "very weird", "ting ting, not good", "causing more headache" and, after another attempt, "still causing headache... mid-match I just want to stop training to find peace." The fourth version, the one now live, finally got a "great". This post is what we learned in between, with the measurements.

A sound you hear 500 times is a different problem

Most game sound design is judged in short bursts: does the shot feel punchy, does the kill feel rewarding. Training is different. The same short sound repeats every half-second for minutes, so small irritations add up. Three findings from hearing research explain almost everything that went wrong with our first three attempts.

1. Tones annoy more than noise

Environmental noise standards add a penalty, typically a few decibels, when a noise contains a clear tone: a hum, a whine, a beep. Listening experiments back this up. Tonal sounds are rated more annoying than broadband sound at the same level, especially when the tone is prominent and the surroundings are quiet.1,2,3 How big the penalty should be is still debated, but the direction is not.

A typical "ding" hit sound is exactly that: a clear tone, in a quiet room, repeated hundreds of times. Every one of our first three hit sounds was a note. We made it lower, softer and rounder each time. It kept failing, because it was still a note.

2. The 2–5 kHz band is the harsh one

In a 2012 brain-imaging study, people rated 74 sounds for unpleasantness, from knives on bottles to flowing water. The sounds rated worst had their energy concentrated around 2,000 to 5,000 Hz, the range where human hearing is most sensitive, and activity in the amygdala rose with how unpleasant a sound was rated.4 Psychoacoustic models of annoyance capture this as "sharpness": the more of a sound's energy sits at high frequencies, the more annoying it tends to be.5

Our first hit was a bright bell at 1,760 Hz with glassy overtones reaching well above 5 kHz. Our tracking tick, which plays about 11 times a second while you hold the crosshair on a target, was worse: 98% of its energy sat inside that band.

3. Loud and sudden means startle

The startle reflex fires when a sound is both loud and fast-rising. Studies of the reflex find that a slower rise time weakens it, and that repeatedly startling an animal quickly leads to avoidance of the place it happens.6 Most of that work is in animals, so treat the details with care, but the shape matches what our tester described: wanting to get away from the training, not just finding it unpleasant.

All three of our early hit sounds were louder than the gunshot they followed, by 2.5 to 6.7 dB, and they started instantly. Even the mellow ones meant every click delivered two sudden events, and the second was the bigger one.

What we measured

We stopped judging by ear (ours or anyone's) and measured every version on the axes the research points to: loudness, energy in the 2–5 kHz band, sharpness, and how strongly a single tone stands out ("tone prominence": how many decibels the strongest partial sits above the rest of its third of an octave).

How beep-like was each hit sound?
Tone prominence: how far the strongest tone stands above its surroundings (dB, higher = more of a beep)
Round 1: bright bell
10.3 dB
Round 2: pop + marimba note
7.0 dB
Round 3: felt mallet
5.1 dB
Now: wooden knock
0.9 dB

Each attempt was lower and softer than the last, but only removing the tone altogether fixed the headaches.

Show the numbers
GroupValueSample
Round 1: bright bell10.3 dBstrongest tone at 1,758 Hz
Round 2: pop + marimba note7.0 dBstrongest tone at 785 Hz
Round 3: felt mallet5.1 dBstrongest tone at 521 Hz
Now: wooden knock0.9 dBno clear tone
Measured by Online Aim Trainer on the in-game sounds, first 40 ms, 100 Hz to 2 kHz.
How much louder than the gunshot was the hit sound?
Hit minus shot, RMS level over the first 50 ms (dB)
Round 1: bright bell
+5.7 dB
Round 2: pop + marimba note
+6.7 dB
Round 3: felt mallet
+2.5 dB
Now: wooden knock
0.0 dB

The confirmation was the loudest thing in the room. Now it is no louder than the shot it confirms.

Show the numbers
GroupValue
Round 1: bright bell+5.7 dB
Round 2: pop + marimba note+6.7 dB
Round 3: felt mallet+2.5 dB
Now: wooden knock0.0 dB
Measured by Online Aim Trainer at in-game levels, each hit against the gunshot it played with at the time.
How much of the tracking tick sat in the harshest band?
Share of the sound's energy between 2 and 5 kHz
Before: 2.95 kHz tick
98.4%
Now: soft tap
0.5%

The tick that plays 11 times a second on target used to live almost entirely in the band people find hardest to listen to.

Show the numbers
GroupValue
Before: 2.95 kHz tick98.4%
Now: soft tap0.5%
Measured by Online Aim Trainer on the in-game sounds.

What we built instead

The new hit sound is modelled on a knock on wood, the register people chase in "thocky" mechanical keyboards and bubble wrap. A very short, soft tap of noise excites three heavily damped resonances (around 240, 560 and 1,100 Hz) that die away almost immediately, so nothing rings long enough to become a note. In detail:

The tracking tick became the same knock, smaller. And we went back to the gunshot itself, a real recording of a 9 mm round: we cut its echo from 1.25 to 0.7 seconds, turned down its harsh band and lowered it by about 4 dB. Fired two to four times a second, a long tail stacks into a constant wash of noise, and slow swells in loudness are associated with fatigue. It now sounds more like a range heard through electronic ear protection, which is also how most shooters actually hear one.

The few sounds that are still notes (the countdown, the start and end of a run, level-ups) are quiet, short and drawn from one five-note scale (C major pentatonic), which has no clashing intervals. When two of them overlap, they form a chord, never a grating beat.

What about 432 Hz or 528 Hz? You will find "healing frequencies" recommended for focus and calm. We looked. The evidence is thin: one small pilot study of 33 people reported a slightly lower heart rate with music tuned to 432 Hz instead of 440 Hz,7 and nothing supports specific numbers like 528 Hz. What does have support is broader: quiet, slow, pleasant sound helps, and natural sounds may help people recover from stress a little faster.8 So we aimed for "soft and natural", not for a magic number.

If any trainer's sounds bother you

  1. Turn the volume down before you turn it off. Loudness is the biggest single driver of annoyance. Many trainers are mixed loud to feel punchy in a short clip.
  2. Pick the dullest hit sound on offer. A "thock" or "pop" over a "ding" or "ping". If it has a clear pitch, it will wear on you faster.
  3. Watch the tracking tick. A sound that repeats many times a second, like a held-on-target tick, is the most fatiguing of all.
  4. Take the break. If you notice you want to stop, stop for a minute. In our data warm-up lasts about 30 minutes, so a short pause costs you nothing.

On Online Aim Trainer, sound effects can be switched off from your profile settings, and so can the gun itself ("Show gun") if you prefer a clean screen.

Hear it for yourself. One Flick run, with the new sounds.

Start a Flick drill →

How we measured this

What was measured. Every number in this post is a measurement of our own sounds, not player data. Each version was rendered exactly as the game plays it, at its in-game volume, and analysed with a short script (tools/sfx-measure.py in our code base): RMS level over the first 50 ms, the share of spectral energy between 2 and 5 kHz, the spectral centroid, and tone prominence (how many decibels the strongest partial between 100 Hz and 2 kHz stands above the mean of its third of an octave, over the first 40 ms). The gunshot comparison uses the version of the shot that played alongside each hit sound.

Limits. These are physical measures, not a listening study. The research they lean on was done with environmental noise, everyday sounds and, for startle, mostly animals, not with game sounds. One tester's reports drove the changes, and other people may like different sounds. The measures tell you which direction is safer, not what is best for everyone.

Sources

  1. Hongisto, V., et al. (2019). Annoyance of low-level tonal sounds: a penalty model. record
  2. Subjective evaluation on the annoyance of environmental noise containing low-frequency tonal components (2021). International Journal of Environmental Research and Public Health. ncbi.nlm.nih.gov/pmc/articles/PMC8297235
  3. Perception of tones in machinery noise and its influence on annoyance. Doctoral dissertation, Purdue University. docs.lib.purdue.edu
  4. Kumar, S., von Kriegstein, K., Friston, K., & Griffiths, T. D. (2012). Features versus feelings: dissociable representations of the acoustic features and valence of aversive sounds. Journal of Neuroscience, 32(41), 14184–14192. doi.org/10.1523/JNEUROSCI.1759-12.2012
  5. Fastl, H. Psychoacoustics, sound quality and music. Technical University of Munich. portal.fis.tum.de
  6. Götz, T., Pacini, A. F., Nachtigall, P. E., & Janik, V. M. (2020). The startle reflex in echolocating odontocetes: basic physiology and practical implications. Journal of Experimental Biology. ncbi.nlm.nih.gov/pmc/articles/PMC7075047
  7. Calamassi, D., & Pomponi, G. P. (2019). Music tuned to 440 Hz versus 432 Hz and the health effects: a double-blind cross-over pilot study. Explore, 15(4), 283–290. doi.org/10.1016/j.explore.2019.04.001
  8. Alvarsson, J. J., Wiens, S., & Nilsson, M. E. (2010). Stress recovery during exposure to nature sound and environmental noise. International Journal of Environmental Research and Public Health, 7(3), 1036–1046. pmc.ncbi.nlm.nih.gov/articles/PMC2872309