What Is Convolution in Audio?
The math trick behind HeSuVi, reverb plugins, and cabinet simulators — explained in plain English.
The 30-Second Version
Convolution is a way of stamping one sound’s character onto another. You take an impulse response (IR) — a recording of how a space, device, or virtual speaker colors a sharp click — and mathematically blend it with your audio. The result sounds as if it were played through that space or device. HeSuVi convolves game audio with HRIRs so headphones sound like a 7.1 speaker room.
Everyday Examples You Already Know
- Reverb plugins convolve dry recordings with IRs of real halls and cathedrals.
- Guitar amp simulators convolve with IRs of speaker cabinets and microphones.
- HeSuVi convolves each surround channel with HRIRs so two headphone drivers fake seven speakers.
- Even room correction (Sonarworks, Dirac) uses convolution to cancel your room’s coloration.
Why Convolution Sounds So Real
An impulse response captures everything linear about a system — frequency response, timing, reflections — in one file. EQ can only fix tone; convolution reproduces space and direction too. That completeness is why a good HRIR set is so convincing, and why HeSuVi’s approach beats simple ‘widening’ effects.
The Cost: CPU
Convolution is math-heavy — longer IRs cost more CPU. HeSuVi keeps it light by running inside the Windows audio engine (no separate app) and using efficient FFT-based convolution. See our CPU usage breakdown for the honest numbers story.
Convolution vs Regular EQ
Think of it this way: EQ adjusts tone (more bass, less treble). Convolution adjusts tone plus space, direction, and reflections. HeSuVi uses both — convolution for the surround illusion, EQ (AutoEq) for headphone correction — which is why the combination is so powerful.
Try it yourself: HeSuVi applies convolution in real time inside the Windows audio engine. Download HeSuVi and follow the setup guide.
Frequently Asked Questions
What is convolution reverb?
Reverb created by convolving dry audio with an impulse response recorded in a real space — as opposed to algorithmic reverb, which simulates reflections mathematically.
Does convolution add latency?
Some, inherently — but efficient FFT convolution inside the audio engine keeps it small. See our CPU/latency breakdown.
What is an impulse response file?
A WAV recording of how a system (room, speaker, HRTF position) responds to a sharp impulse — the ‘fingerprint’ convolution applies to your audio.