Convolution reverb applies an impulse response to an input signal to produce a response-based effect. A supplied IR can represent a recorded space, a physical reverberation device, a synthesized response, or a deliberately creative sound. It does not have to be a recording of a concert hall.
A fixed response represents a particular relationship under its capture or construction conditions. It is not a complete interactive model of every source and listener position in a room, and it does not automatically capture every nonlinear or time-varying behavior of a physical device.
Think of the IR as a response description
For a simple linear time-invariant model, the output is the input combined with time-shifted, scaled contributions described by the IR. A short impulse at the input produces the response itself. Real reverb responses contain many more samples and structure than a teaching example.
For an idealized 48,000 Hz response, a tap at sample 2,400 occurs after 50 milliseconds, while sample 4,800 occurs after 100 milliseconds. These timing conversions are arithmetic, not a measurement of a particular room.
A deliberately simple example
| Response sample | Time at 48 kHz | Illustrative gain |
|---|---|---|
| 0 | 0 ms | 1.0 |
| 2,400 | 50 ms | 0.5 |
| 4,800 | 100 ms | 0.25 |
Each input event would be accompanied by those shifted contributions under the stated simple model. A dense reverb IR adds a far richer sequence of arrivals and decay. The reverb-versus-delay guide explains how sparse repeats differ perceptually from a distributed tail.
What a recorded response can represent
An IR can include arrival timing, spectral shaping, and decay associated with its source-to-receiver path and processing. Moving the source, receiver, or important room elements can change that relationship. One captured path does not describe every location automatically.
The early-reflections guide explains why the initial pattern matters separately from the late tail. An IR’s channel configuration also matters; a stereo label does not by itself explain every input-to-output path.
Convolution and algorithmic approaches can overlap
In production terminology, convolution is often contrasted with a synthesized algorithmic engine. Both are computational methods. An effect can combine an IR-based stage with a separate algorithmic tail, filtering, modulation, or other processing.
The algorithmic reverb guide explains the synthesized-response side. Avoid universal claims that convolution is always more realistic or that algorithmic processing is always lighter on CPU.
Check the actual source of the response
A hall, chamber, plate, or spring response can all be used in a convolution workflow if suitable data is supplied. A creative IR can instead produce a deliberately nonphysical texture. The category describes the method, not one fixed sound.
The reverb-type guide separates physical systems, space labels, and implementation approaches. Do not infer the authenticity of a named venue or device from a preset label without supporting information.
Inspect direct-sound handling
A response can contain a direct component as well as reflected energy, while the convolution plugin may provide its own dry/wet arrangement. If a conventional parallel return includes unintended dry energy, the original source can be duplicated.
The routing guide explains the path check. Verify the IR and engine’s handling rather than assuming every file is already trimmed or prepared for the same wet-only workflow.
Editing an IR changes the modeled response
Truncation, envelopes, filtering, time scaling, and reversal can change the result where supported. A shortened response may no longer preserve the full original decay, and a sample-rate-related size change can affect both timing and spectrum depending on the method.
The reverse-reverb guide distinguishes an edited response from reversing a rendered source effect. Check the engine’s actual controls; do not claim every convolution tool supports the same transformations.
Use calculated timing without overclaiming portability
The homepage reverb calculator supplies BPM-based creative pre-delay and decay starts. It does not load or process an IR. A numeric decay start may not correspond directly to a convolution plugin’s response-length or envelope control.
At quarter note = 120, a straight sixty-fourth is 31.25 milliseconds. It can be a candidate onset gap, but the supplied IR may already contain onset timing. Check the combined behavior rather than adding delays blindly.
Choose level and spectrum in the arrangement
A convincing soloed response can still mask a source in the mix. Compare the dry signal with the wet level, onset, tail overlap, and spectral balance. Preserve the intended character while addressing the actual problem.
The reverb EQ guide explains feed and output filtering, including the ideal linear-cascade qualification. A change in route or added dynamics can make a practical comparison differ from a simple fixed convolution example.
Preserve enough output for the response
The wet signal can continue beyond the source file’s last event. Long responses or serial processing can require more render range than expected. A loop and a natural outro also need different boundary handling.
The tail export guide explains path, range, and re-import checks. The mathematical response length and the meaningful audible ending should be interpreted in the actual arrangement.
Keep capture and performance claims honest
Use suitable rights and provenance for a supplied response, and retain its source and relevant conditions. A file name alone does not prove where or how it was captured. This guide does not claim an IR recording, acoustic measurement, plugin installation, or performance benchmark was completed.
For physical room questions, the room-acoustics guide separates predictions, measurements, modes, and reflection behavior. An IR-based creative effect is not automatically an assessment of the playback room.
Frequently asked questions
What is an impulse response?
It describes a system’s response under particular conditions. In a fixed linear model, it determines how the input is convolved into the output.
Must an IR come from a real room?
No. It can represent a device, a synthesized response, or a creative sound as well as a captured space.
Does one IR model every listening position?
No. A fixed response represents particular paths and conditions. Position and channel configuration affect its scope.
Can convolution reproduce every device behavior?
A fixed linear response does not automatically capture arbitrary nonlinear or time-varying behavior. Additional processing may extend an engine’s design.
Does the site’s calculator process impulse responses?
No. It supplies numeric timing starts and separate room predictions. Use suitable audio software for IR processing.
