3D room — image-source reflections (drag to orbit, scroll to zoom)idle
Decay curve — sound level (dB) vs time
early reflections Sabine Eyring
This tool applies the image-source method to a rectangular room: every wall, the floor and the ceiling are mirrored to build virtual "image" sources whose straight-line distance to the listener equals the true length of a reflected sound path. First- and second-order images are computed exactly, giving accurate bounce points and arrival times for the room's earliest reflections, while the classic Sabine and Eyring equations predict the full RT60 decay from room volume and absorption.
Real image-source geometry driving glowing pulses that bounce off a 3D room's walls, floor and ceiling, alongside a live decay-curve graph comparing the Sabine and Eyring RT60 predictions as dimensions and materials change.
Drag the Width, Length and Height sliders to resize the room, pick a material for the side walls, floor and ceiling, then press 🔊 Trigger Impulse to watch reflections propagate and arrive on the decay graph.
A concrete-walled room with almost no soft furnishing (α ≈ 0.02) can have an RT60 of several seconds, while the same room lined with acoustic foam (α ≈ 0.8) can drop to well under half a second — that's the difference between a cathedral echo and a recording booth.
RT60 is the time it takes for sound pressure level to decay by 60 decibels after a sound source stops — the standard measure of how "live" or "dead" a room sounds, first defined by Wallace Sabine in the 1890s.
Both predict RT60 from room volume and total absorption, but Sabine (RT60 = 0.161·V/A) assumes lightly damped rooms and becomes inaccurate as average absorption ᾱ grows large; Eyring (RT60 = 0.161·V / (−S·ln(1−ᾱ))) accounts for the fact that a fully absorptive room (ᾱ→1) should have RT60→0, which Sabine's formula never reaches.
It's a geometric technique for computing specular sound reflections exactly: reflecting the source across a wall's plane produces a virtual "image source" whose straight-line distance to the listener equals the true path length of a ray that bounces once off that wall — the same idea extends recursively to second-, third- and higher-order reflections.
Sound travels at about 343 m/s, so in a typical room reflections arrive within tens of milliseconds — far too fast to see. This simulation slows the animation down by a fixed factor so you can watch individual bounce paths develop while the underlying travel times and distances stay physically accurate.
Absorption coefficient α depends on a material's porosity and thickness at the frequency of interest: rigid, smooth concrete reflects almost all incident sound energy (α ≈ 0.02), a carpet's fibres damp some (α ≈ 0.3–0.5), heavy curtains trap more via a thicker porous/air-gap structure (α ≈ 0.5–0.7), and open-cell acoustic foam's deep pore structure converts most of the rest into heat (α ≈ 0.7–0.9).
The number of image sources grows combinatorially with reflection order (6 at order 1, 30 more at order 2 for a 6-wall room), so order-1 and order-2 already capture the perceptually important early reflections; beyond that the field becomes so dense it's better modelled statistically — which is exactly what the Sabine/Eyring decay curve represents.
Each traced path's amplitude is the product of the amplitude reflection coefficient √(1−α) at every wall it bounced off, divided by the total path length to model spherical spreading loss; the result is converted to decibels and plotted as a dot on the decay graph at its true arrival time.
No — RT60 from Sabine and Eyring depends only on room volume, total surface area and absorption, not on where the source or listener sit; what does change with position is which specific early reflections arrive first and how loud they are, which is why the pulse paths update live while the decay curve's overall shape stays fixed.
Room dimensions and per-surface materials drive an exact first/second-order image-source computation; the decay graph overlays measured pulse arrivals on the theoretical Sabine and Eyring exponential decay curves.