How rooms behave · Fundamentals
Reflections and first reflection points
Why a bare wall smears the image, what a treated wall does instead, and how to find the spots that matter.
7 min read4 sourcesChapter 2 of 24
Schematic
Left: a bare plastered wall returns the wave almost intact, a few milliseconds behind the direct path. Right: the same wave into a porous absorber loses most of its energy as heat inside the fibres and returns as a faint remainder.
Animated illustration of the principle. Not measured data.
Sound leaving a loudspeaker spreads in every direction. Only a small part travels straight to you; the rest hits the walls, floor and ceiling and comes back. The first of those returns — the ones that bounce once — arrive so soon after the direct sound that your ear cannot separate them. It hears one sound, coloured.
This chapter is about that colouring, and about the difference between a surface that reflects and one that has been treated.
What a hard wall does
Plaster, brick, glass, gypsum board and tile all reflect mid and high frequencies almost completely. The reflection arrives a few milliseconds late and slightly quieter. Added to the direct sound, it creates a pattern of reinforcements and cancellations across the frequency range — engineers call it comb filtering because of how it looks on a graph. You hear it as a hollowness or harshness that moves as you move your head.
Reflections from the side walls also pull the stereo image wider and less precise; a reflection from the ceiling adds an unnatural sense of height to voices.
What a treated wall does
A porous absorber — mineral wool, glass wool or polyester fibre behind fabric — lets the sound wave into a maze of fibres. The air in the material is forced to move through narrow gaps, and friction turns the sound energy into a tiny amount of heat. What reflects back is a fraction of what arrived.
How much is absorbed depends on thickness and frequency. Thin panels absorb the top of the range and pass the rest; thicker panels, or panels spaced off the wall, reach lower. This is the single most common way treatment goes wrong: thin material on the first reflection points tidies the highs and leaves the low-mids largely untouched. Tested 25 mm glass-fibre board absorbs about a quarter of the sound reaching it at 250 Hz; 50 mm absorbs about 70 %.
Rule of thumb: a porous layer on a hard wall approaches its highest absorption around the frequency whose quarter-wavelength equals its thickness — about 1.7 kHz for 50 mm, 860 Hz for 100 mm — and tapers off gradually below that rather than stopping. It is a rough guide; the material and any air gap behind it also shape the curve. An air gap behind it moves the whole curve lower.
Finding the points that matter
Not every square metre of wall is equal. The reflections that reach the listening seat after one bounce come from a handful of predictable spots: the side walls between the loudspeakers and the seat, the ceiling above that path, the floor (usually the carpet’s job), and the front wall around and behind the loudspeakers.
The mirror trick finds them without any equipment. Sit where you will listen; have someone move a mirror flat along the wall; every position where you can see a loudspeaker in it is a first reflection point. For multiple seats the points spread into zones, which is why a theatre treats whole bands of wall rather than isolated squares.
Not every reflection is the enemy
In a concert hall, early lateral reflections are what make the sound enveloping — designers work to create them. In a home theatre with a decoded surround field, the loudspeakers already provide envelopment, so reflections from the room mostly get in the way. In a two-channel listening room, opinions and preferences differ, and some designers deliberately leave side-wall reflections and diffuse the rear.
The point is that reflections are managed to a purpose, not eliminated by reflex. Which is also why absorbing every surface produces a room that is unpleasant to be in.
For integrators and architectsThe numbers, the standards, and what to specify
- Geometry first
- Reflection points follow from loudspeaker and seat positions. Fix those before specifying treatment, or the treatment will be in the wrong place.
- Time window
- Reflections arriving within about 20 ms of the direct sound and within roughly 10 dB of it are the ones that colour the image. Later, quieter reflections read as room ambience.
- Thickness, not area
- Coverage of first reflection zones with 25 mm material does little below about 500 Hz (tested 25 mm board: 0.26 at 250 Hz, 0.05 at 125 Hz). Specify 50–100 mm porous absorber for reflection control, and treat thickness as the primary variable.
- Angle
- Specular reflection obeys the mirror law, so a splayed side wall or an angled panel redirects the reflection away from the seat — sometimes preferable to absorbing it.
- Ceiling and floor
- The ceiling reflection is usually the strongest untreated one in a domestic room. A cloud of absorber above the front third of the seating path, or a full absorbent ceiling in a theatre, is standard practice.
What to write into the brief
- Reflection-point treatment as 50 mm minimum porous absorber, 100 mm where space allows, with the listening-seat geometry drawn.
- Fabric finish specified as acoustically transparent — open-weave, easily breathed through — so the fabric does not undo the panel.
Sources
- 01Owens Corning, Fiberglas 700 Series Board — Product Data Sheet (ASTM C423, Type A mounting)Measured absorption coefficients for 25, 50, 75 and 100 mm board.
- 02Zhang et al., DAGA 2021 — porous absorber effectiveness and the quarter-wavelength criterionCites Allard & Atalla, Propagation of Sound in Porous Media.
- 03ISO 354:2003 — Measurement of sound absorption in a reverberation room
- 04Toole, Sound Reproduction: The Acoustics and Psychoacoustics of Loudspeakers and Rooms, 3rd ed. (Routledge, 2017)Textbook on loudspeakers, reflections and small rooms.
External links open in a new tab. Figures quoted in this chapter are taken from these sources; a Science of Sound measurement is only ever presented as one, with its file, date and instrument.
Keep reading