Think about the difference between a mirror and a frosted glass window. When you hold a mirror up to a light, it creates a single bright spot — a strong, directional reflection. When you hold frosted glass up to the same light, the glow spreads out softly in all directions. Neither surface destroys the light. But the mirror concentrates it while the frosted glass scatters it.
Sound in a room behaves the same way. A flat, hard wall surface acts like an acoustic mirror — it produces a strong, directional reflection that can cause problems like flutter echo (the repeating slap you hear when you clap in a bare gymnasium) or distinct echoes in large halls. Acoustic diffusion is the frosted-glass approach: instead of absorbing the sound or reflecting it in one strong direction, a diffuser breaks it up into many weaker reflections that spread energy evenly through the room.
The result is a room that feels acoustically "full" and enveloping without the harsh flutter echoes and image-degrading strong reflections of an untreated hard room.
Why Diffusion Matters
Before diffusers became standard tools in acoustic design, engineers faced an uncomfortable trade-off. A recording studio or concert hall needed:
- A controlled RT60 (not too long, not too short)
- Freedom from flutter echo and colouration
- A sense of spaciousness and envelopment
Diffusion broke the deadlock. By scattering sound instead of absorbing it, diffusers allow a room to maintain its liveness (RT60 stays approximately the same) while eliminating the discrete strong reflections that cause flutter and colouration.
How a Flat Surface Reflects Sound
To understand diffusion, you first need to understand what a perfectly flat, hard surface does to sound.
When a sound wave strikes a flat surface at an angle, it reflects at the same angle on the other side of the surface normal — exactly like light bouncing off a mirror. This is called specular reflection. The reflected sound is essentially a copy of the incoming sound, delayed by the additional path length it has travelled to reach the reflective surface and back.
If the angle is favourable (e.g., sound bouncing between two parallel walls), specular reflections can create a flutter echo — a rapid series of repeating reflections that color the sound with a metallic "twang." In a large space, a single strong specular reflection from the rear wall can arrive at listeners in the stalls as a distinct echo separate from the direct sound, which is deeply disorienting.
Diffusion prevents both problems by replacing the single specular reflection with many weaker reflections in scattered directions.
Types of Diffusers
Surface Texture and Irregular Geometry
The simplest form of diffusion is irregular surface geometry: a bookshelf filled with books of different depths, rough-hewn stone cladding, a facade of staggered brickwork. When a sound wave encounters a surface whose features vary on a scale comparable to the wavelength of sound, it scatters diffusely. The scattering effectiveness depends on the size of the surface irregularities relative to the wavelength.
At 1000 Hz, the wavelength of sound in air is approximately 0.34 m. Irregularities of 0.1 m or more begin to cause meaningful scattering at that frequency. At 250 Hz, the wavelength is 1.4 m — irregularities need to be much larger (0.3–0.5 m or more) to be effective. This is why purely textural solutions (carpet, fabric, rough plaster) do not provide broadband diffusion: their irregularities are too small to scatter low-frequency energy.
Schroeder Diffusers (QRD and MLS)
The most scientifically rigorous diffuser designs come from the work of Manfred Schroeder at Bell Labs. In 1975, Schroeder published a method for designing wells of varying depth that would scatter sound uniformly in all directions. His insight was to use number-theoretic sequences to determine well depths, ensuring that the phase shifts introduced by each well are spread evenly across a wide range of directions and frequencies.
The most widely used type is the Quadratic Residue Diffuser (QRD). Its well depths follow the quadratic residue sequence for a prime number N:
dₙ = (n² mod N) × (λ_design / 2N)where n is the well index (0 to N-1), N is the chosen prime number, and λ_design is the design wavelength corresponding to the lowest frequency the diffuser is intended to scatter.
For example, a QRD based on prime N = 7 has wells with depths proportional to the sequence: 0, 1, 4, 2, 2, 4, 1 (these are 0², 1², 2², 3², 4², 5², 6² modulo 7). The sequence has a special mathematical property: the discrete Fourier transform of these depths has constant magnitude, which in the acoustic context means that each diffraction order carries equal energy — the sound is scattered uniformly in all directions.
A second Schroeder type is the Maximum-Length Sequence (MLS) diffuser, which uses a two-level (binary) well depth pattern derived from binary feedback shift registers. MLS diffusers are effective in a single dimension (they scatter in the horizontal plane) and are often used for side-wall treatment.
Optimised Curved Diffusers
Cylindrical and hemispherical diffusers scatter sound in an arc or hemisphere, respectively. They are less mathematically sophisticated than QRDs but are easier to manufacture and visually integrate into architectural finishes. Curved ceiling elements in modern concert halls often serve a dual purpose: they direct early reflections to seating areas while providing some diffusion of the mid-to-high frequency energy.
Scattering Coefficient: Quantifying Diffusion Performance
The performance of a diffuser is characterised by its scattering coefficient (s), defined in ISO 17497-1:2004 as the fraction of reflected sound energy that is scattered into non-specular directions:
s = 1 - (specular reflected energy / total reflected energy)Values range from 0 (perfect specular reflector, like a flat mirror) to 1 (all energy scattered non-specularly). In practice:
- Smooth painted concrete: s ≈ 0.05 at 1000 Hz
- Random bookshelf with varied depths: s ≈ 0.30–0.50 at 1000 Hz
- QRD diffuser (well-designed): s ≈ 0.80–0.95 at design frequencies
A Worked Example: Rear Wall of a 300-Seat Lecture Theatre
Problem: A 300-seat lecture theatre is 18 m long × 12 m wide × 5 m high. The rear wall is painted concrete. The acoustic consultant has measured a strong echo from the rear wall arriving at the front speaker position approximately 85 ms after the direct sound (round trip: 18 m × 2 = 36 m ÷ 343 m/s ≈ 105 ms). The echo is distinct and disturbing.
Option A — Absorb the rear wall: Install 75 mm rockwool panels on the rear wall. Average absorption coefficient α ≈ 0.90 at 1000 Hz. This eliminates the echo but removes approximately 144 m² of reflective surface from a 1,080 m³ volume, reducing RT60 from 1.2 s to approximately 0.75 s — below the 0.9–1.1 s recommended for lecture theatres by BB93 and ISO 3382-2.
Option B — Diffuse the rear wall: Install QRD panels (N = 7, design frequency = 250 Hz, well depth range 0 to 0.27 m) across 80% of the rear wall. The scattered reflections now arrive distributed over a 40–120 ms window rather than as a single strong spike at 105 ms. The echo is eliminated. RT60 is essentially unchanged because the QRD panels absorb very little energy (absorption coefficient < 0.10). The room retains its liveness.
The diffusion solution preserves the acoustic energy the room needs for music and speech projection while resolving the specific pathology (discrete rear-wall echo) that was degrading performance.
When to Use Diffusion vs Absorption
| Situation | Preferred treatment |
|---|---|
| RT60 is too long — room too reverberant | Absorption |
| RT60 is correct but flutter echo is present | Diffusion |
| Room sounds "small" and dead despite correct RT60 | Diffusion (rear/side walls) |
| Strong echo from a single surface | Diffusion (on that surface) |
| Recording studio control room rear wall | Diffusion (QRD) |
| Corridor with parallel hard walls | Diffusion on one wall, absorption on the other |
| Bass frequency build-up in corners | Absorption (bass trap — diffusers are ineffective at low frequencies below their design limit) |
How AcousPlan Helps
AcousPlan's materials library includes diffuser panels with scattering coefficients across six octave bands. When you apply a diffuser to a room surface in the simulator, the engine incorporates the scattering data into the simulation — modelling not just the change in absorption but the effect on the early-to-late energy ratio and flutter echo risk.
The room simulator allows you to compare a surface treated with absorption against the same surface treated with a diffuser of equivalent cost, letting you see the trade-off in RT60, C80, and the flatness of the energy decay curve before you commit to a specification.