Why We Synthesized Every Sound on This Site From Scratch

Here is a fact that surprises almost everyone: most famous meme sounds are owned by somebody. The boom is from a film score. The airhorn is a producer’s sample. That iconic scream, that TV sting, that video-game jingle — each one has a rights holder, and increasingly, each one has an automated system enforcing it. When we built this site, we made a decision that shaped everything: no ripped clips, ever. All twenty sounds on the board are generated from raw mathematics by a script we wrote. This post is the why and the how.

The copyright minefield

Meme culture treats sounds as folklore — common property that everyone remixes. Copyright law does not. A two-second clip from a movie, a song, or someone’s stream is still a copyrighted recording, and platforms now scan for them at upload with audio fingerprinting. Creators know the result: videos muted, monetization redirected, strikes accumulated — often for a sound that lasted half a second.

A soundboard site built on ripped clips inherits every one of those problems and then distributes them to its users. That was the part we could not stomach: we would be handing people files that might get their videos flagged. The alternative was more work but structurally clean — build sounds that capture the archetype without copying anyone’s recording. The boom-shaped idea is not ownable; a specific recording of it is. So we make our own recordings, from scratch, out of math.

What “synthesized from scratch” actually means

There are no microphones, no samples, and no borrowed files anywhere in our pipeline. Every WAV on this site is produced by a script — a few hundred lines of code that generate audio the way 1980s hardware synthesizers did, just in software:

• Oscillators. Loops that compute pure waveforms sample by sample — sine waves (smooth and round), sawtooths (buzzy and brassy), squares (hollow and chiptune-y), triangles (soft and flute-like). At 44,100 samples per second, a “note” is just a column of numbers tracing a shape.
• Envelopes. Volume curves that give a sound its life cycle — a 4-millisecond attack reads as a strike, a slow two-second decay reads as a ring-out.
• Noise. Seeded random numbers, filtered into usefulness: high-passed for snare rattle and cymbal wash, low-passed for thumps and punches.
• Filters and sweeps. Low-pass filters darken a sound; sweeping one upward during a noise burst is the entire recipe for an EDM riser.
• Saturation. A gentle waveshaping function that adds harmonics — the trick that makes bass audible on phone speakers, which we covered in the anatomy of a meme sound.
• Delay. A feedback echo, tuned per sound — 110 ms of slap on a boom, a long 50%-feedback wash on the airhorn symphony’s concert hall.

Stack those primitives and you can build almost anything. Our airhorn is two detuned sawtooths, a square wave, and a sub-octave saw pretending to be a horn section. The sad trombone is filtered saws walking down a chromatic scale. The crickets are 22-millisecond sine pulses at 4.3 kHz arranged in the same pulse-train clusters real cricket wings produce. Every page in our sound library documents its own recipe, because we think the recipes are genuinely interesting.

The unexpected creative upside

We expected the from-scratch rule to be a constraint we tolerated. It turned out to be the best creative decision of the project. When you rip a clip, you get exactly that clip. When you synthesize, every parameter is negotiable: how deep the boom falls, how long the horn sags, how fast the wobble wobbles. Tuning a sound becomes composition.

It also means the sounds are reproducible. The script uses seeded randomness — the “random” noise in a snare hit comes from a generator that produces the same sequence every run — so the entire soundboard can be rebuilt, byte-identical, from code. The sound library is source code, which is a sentence we did not expect to write about an airhorn.

What this means for you

The practical payoff: every free sound here can be downloaded and used in your videos, streams, and edits without inheriting anyone else’s rights problems. No fingerprint database has our waveforms attached to somebody’s movie. That is the entire pitch — sounds with the cultural energy of the classics and none of the legal baggage.

Meme culture is going to keep generating legendary sounds, and rights holders are going to keep lawyering them. We think the sustainable path for anyone building in this space is the one we took: study the archetype, honor the lore — we write up every sound’s history for exactly this reason — and then build your own version, from the waveform up. The math is free. It always has been.