Build a room from its measurements and play a recording in it. The reflections are not designed, they are computed: from the size of the room, what its surfaces absorb, and where the source and the listener are standing.
Room size
Length runs from 2 m to 30 m, width from 2 m to 20 m, height from 2 m to 10 m. Size sets two things at once. A bigger room has a longer decay for the same surfaces, and its first reflections arrive later, which is the cue the ear uses to judge how large a space is. A 3 m bathroom returns its first reflection within 5 ms; a 25 m hall takes ten times as long.
Absorption
Absorption is the share of sound each surface removes per bounce.
| Surface | Absorption |
|---|---|
| Bare concrete, tile | 2% to 5% |
| Plaster, glass | 5% to 10% |
| Furnished living room, on average | 20% to 30% |
| Carpet and heavy curtains | 40% to 50% |
| Treated studio | 60% and up |
High damping makes the top end of the decay die faster than the rest, as it does in any room with soft surfaces and air in it.
The readout
Under the waveform the page shows four figures for the room as set.
- Volume, in cubic metres.
- Predicted RT60, the time for the reverberation to fall by
60 dB, from Eyring’s formula. - Source to listener distance.
- Critical distance, where direct and reverberant sound are equally loud. A listener closer than this hears mostly the source; one further away hears mostly the room.
Positions
Source and listener are each placed along the length and across the width, as a percentage of the room. The source stands at 1.5 m and the listener at 1.2 m, lower in a low room.
Moving them changes more than the balance. A listener 1 m from the source in a live room hears it nearly dry. Back the listener into a corner and the bass reflections pile up. Put source and listener at the same distance from a side wall and one strong early reflection colours everything, which is the comb filtering a real room does.
How the response is built
The first 80 ms are counted reflection by reflection with the image-source method, up to about 6,000 of them. After that the reflections are too many and too close together to hear one by one, so the tail is noise shaped to the decay time Eyring predicts, at the level the density of reflections predicts. The travel time of the direct sound is trimmed off so the room stays in step with the untreated file.
Impulse response
Download impulse response writes the room as set to a stereo 48 kHz, 24-bit WAV. It loads into the convolution reverb here or any convolution plug-in. Run it through the reverb time analyzer to check the room against its prediction: the measured figure should land close to the readout, a little shorter where High damping is raised.
Mix
At 100 the output is the room, direct sound included, which is what a microphone at the listener’s position would record. Lower values keep some of the untreated file on top, for when the room is wanted as an effect and not a place.