First reflections explained: what they are, why they matter, and how to treat them

A speaker with a textbook-flat frequency response can still sound unfocused, spatially vague, and oddly coloured in a real room. I've seen this play out on my own measurement rig more times than I can count. The culprit, almost every time, isn't the electronics. It's the first reflection.

This is the one area of domestic audio where the physics is settled, the measurement tools are cheap, and the improvement is genuinely large. Yet most people either ignore it entirely or reach straight for room-correction software without first understanding what they're asking the software to fix. I want to walk through the mechanics properly — because if you understand what a first reflection actually is, you'll make better decisions about when to absorb, when to diffuse, and when to let the EQ have a crack at it.

What a first reflection actually is

Sound leaves your speaker and travels in many directions simultaneously. The direct sound reaches your ears first. Then, a few milliseconds later, a second wavefront arrives — the one that bounced off the nearest hard surface before reaching you. That's the first reflection.

In a typical living room, six of them dominate: two side-wall reflections (one per speaker), two floor reflections, one ceiling reflection, and one from the front wall behind the speakers. The time delay between the direct sound and the first reflection is short — often 5 to 25 milliseconds depending on room geometry. Below roughly 15 ms, your auditory system partially fuses the direct and reflected sounds (this is the Haas effect, or precedence effect). The result isn't echo — it's colouration and imaging smear. The stereo image widens and blurs. Tonal accuracy suffers because the reflected wavefront adds to or subtracts from the direct sound at specific frequencies, creating comb-filter artefacts in the frequency domain.

Run a near-field impulse response measurement with REW (Room EQ Wizard — free, and the single most useful tool in this hobby) and you'll see those early reflections as discrete peaks in the time domain, clear as day, arriving within the first 30 ms window before the reverberant tail builds. The graph doesn't lie.

Why imaging is the first casualty

Your brain localises sound primarily by comparing the signal arriving at each ear: interaural time difference (ITD) and interaural level difference (ILD). A strong lateral reflection off the side wall feeds a delayed, spectrally modified copy of the signal into both ears, confusing that comparison. The image smears toward the reflecting wall. Phantom centre images lose stability. Depth collapses.

This is why I'm sceptical of the common advice to simply crank room-correction DSP and be done with it. Dirac Live, Audyssey XT32, YPAO — all of these tools operate primarily in the frequency domain. They can correct the steady-state frequency response admirably. But comb filtering from a 12 ms first reflection varies continuously with the listening position, and frequency-domain EQ cannot undo a time-domain event. The Audyssey manual itself acknowledges that correction above the Schroeder frequency (the transition from modal to reverberant behaviour) is limited to within a small measurement area. You can calibrate beautifully at the microphone position and have the image still smear 30 cm to the left of it.

The correct sequence is: treat the room first, then apply DSP to clean up what remains. Not the other way around.

Finding the reflection points: the mirror trick

The geometric method for locating side-wall first-reflection points is accurate and takes five minutes. Sit in your listening position. Have a helper slide a small mirror along the side wall at ear height. When you can see the face of a speaker in the mirror, mark the spot. That's where the first-order side-wall reflection originates. Repeat for the ceiling, moving the mirror from the speaker toward your seat along the ceiling plane.

The floor reflection point is directly between the speaker and your listening seat, offset toward the speaker by roughly a third of that distance — but in most rooms with a rug already there, the floor reflection is partially handled. If you're listening on polished concrete or hardwood with no rug, it isn't.

For the measurement-minded: fire an impulse from each speaker and use REW's waterfall or energy-time curve display to confirm those early arrivals. You can calculate the path length of each reflection from the time delay: multiply milliseconds by 0.343 metres per millisecond (the speed of sound at 20°C). A reflection arriving 11 ms after the direct sound has travelled an extra 3.77 m. Basic geometry from there.

Absorption vs diffusion: the honest tradeoff

Once you know where the reflection points are, the treatment choice comes down to absorption or diffusion — and this is where a lot of enthusiasts get it wrong.

Absorption reduces the reflected energy. At the side walls, a 100 mm panel of 48 kg/m³ rockwool (Autex Quietspace, locally made, is a common Australian product; Knauf Earthwool and Bradford SoundScreen are both available from Bunnings trade or acoustic suppliers) will give you roughly 0.7–0.9 NRC (Noise Reduction Coefficient) from about 500 Hz upward. Thicker and denser gets you lower-frequency absorption, but for first reflections in the critical midrange and treble band, 75–100 mm is enough. A 600 x 900 mm panel at each side-wall reflection point makes a measurable difference on the waterfall plot and an audible one in imaging focus.

Diffusion scatters rather than absorbs. A properly designed Schroeder diffusor (QRD or skyline type) with wells calculated for a specific frequency range redirects the reflected energy in multiple directions rather than returning it as a coherent wavefront. The result is that the room still sounds lively — you don't get the dead, claustrophobic effect that over-absorption produces — but the discrete specular reflection is broken up. Diffusion is typically more appropriate on the rear wall behind the listener than on the side walls; rear-wall diffusion preserves a sense of spaciousness without the image-smearing of a rear-wall specular reflection.

My honest opinion here: most people in smaller Australian rooms (typical suburban lounge, 4 x 5 m or similar) benefit more from absorption at the side walls and ceiling than from diffusion at those points, because the room is simply too small for effective diffusion at the relevant wavelengths. Diffusion works best in larger rooms where you have at least 1.5–2 m of distance between the surface and the listener. If you're sitting 2 m from your rear wall, a diffusor there is great. If you're 90 cm from it, you're better off with absorption.

Bass traps and the low-frequency problem

First reflections in the bass range are a separate problem. Below the Schroeder frequency — typically somewhere between 100 and 300 Hz depending on room volume — room modes dominate over reflections. The room's dimensions set up standing waves that cause large frequency-response variations across the room. Moving your head 20 cm can swing bass output by 6 dB or more. This is not a reflection problem; it's a modal problem, and it has different solutions.

Broadband absorption panels (the rockwool panels described above) start to lose effectiveness below about 200 Hz unless they're very thick — 200 mm or more, or specifically tuned as membrane or Helmholtz absorbers. Real bass trapping requires either large volumes of absorptive material (floor-to-ceiling corner traps are the most efficient geometry, because pressure maxima in room modes stack in corners) or tuned resonant absorbers. The corner trap approach is unglamorous but it works: 300 mm triangular rockwool fills in all four floor-to-ceiling corners will produce a measurable improvement in Q at modal frequencies.

This is also where DSP earns its keep. Once you have absorption reducing modal energy — particularly the decay time (RT60) at problem frequencies — room correction tools like Dirac Live can handle the residual frequency-response variance more cleanly, because they're not fighting as much stored modal energy. The best streaming amplifiers and all-in-one systems increasingly include room correction as standard, and they'll perform significantly better in a treated room than an untreated one.

What good measurements look like after treatment

If you've treated the primary first-reflection points and added corner bass trapping, here's what REW should show you compared to your before measurements:

The waterfall plot (frequency vs time, or the spectrogram view) should show significantly faster decay at mid and high frequencies — reflections that previously persisted at -20 dB for 30–40 ms should now be down below -30 dB within 15–20 ms. The impulse response should show a cleaner separation between direct sound and the reverberant tail. The frequency response, measured on-axis at the listening position, will likely be smoother in the midrange because comb-filter interference is reduced. The low end may still show modal unevenness — that's where DSP picks up the remaining work.

Subjectively: the stereo image will sharpen. Centre images stabilise. You'll hear more of the recorded reverb and less of your room's reverb — which is the whole point. The room stops competing with the recording.

Speakers like the KEF LS50 Meta (check price) or the KEF R3 Meta (check price) — both of which I've spent considerable time measuring — produce remarkably consistent off-axis behaviour that makes them relatively forgiving of room reflections compared to speakers with narrow dispersion. But even a well-controlled dispersion pattern benefits substantially from a treated first-reflection zone; the physics still applies, just with lower reflected amplitude to start with.

Practical treatment for a real Australian living room

Most of us aren't starting from a blank slate. There's furniture, a partner with opinions about the décor, and a landlord who may object to holes in the plasterboard. I've had to work within all of these constraints.

The most practical approach for a domestic space: start with two side-wall absorption panels, positioned at the reflection points, in a fabric-wrapped frame. These can be wall-hung or freestanding — freestanding panels on stands (companies like Vicoustic, GIK Acoustics, and local supplier Aurelius Acoustics all sell or can custom-build these) are renter-friendly. A thick rug between speaker and listening seat handles the floor reflection. Two corner floor-to-ceiling traps in the front corners — behind or beside the speakers — make a disproportionate contribution to bass control given their footprint.

If you want to go further: a ceiling cloud above the listening position (a large horizontal absorption panel suspended from the ceiling between speaker and listener) handles the ceiling first reflection efficiently without consuming floor space. A rear-wall diffusor, if the room depth permits, rounds out a genuinely well-treated space.

For the measurement of all this, REW is free and cross-platform. A calibrated measurement microphone — the MiniDSP UMIK-1 runs around $130 and is widely available in Australia — gets you into the game. There is genuinely no excuse not to measure. The data will tell you what your ears can't: which surfaces are the actual problem, and whether your treatment is working.

Room correction software, including the excellent implementations in home cinema processors, remains a powerful final step. But treatment and correction are complementary, not interchangeable. Fix the geometry first. Let the DSP handle the residual. The measurements, and the listening, will confirm it.

— Marcus Vale, Editor