What Sound Does When It Hits Something

A microphone is a genuinely stupid device. It has no idea what it is pointed at. All it does is measure the air pressure at one specific point in space, thousands of times a second, and report what it finds. It cannot tell the difference between a sound that came straight from the singer and the same sound arriving a fraction of a second later after bouncing off a wall. It adds them together and hands you the total.

Which means the thing you are actually recording is never just the source. It is the source plus everything the room did to it on the way. Understanding the four things a room can do is the difference between fixing a recording and guessing at it.

1. Reflection

Sound bounces off hard surfaces, and it bounces predictably: the angle it arrives at equals the angle it leaves at. That is the same rule light follows off a mirror and a ball follows off a table edge, and it means reflections are geometry, not luck. If you know where the source is and where the surfaces are, you can work out where the reflections go — and if you can work that out, you can move the microphone out of their way.

Two consequences are worth knowing because they show up constantly.

The image source

When sound from a single point hits a flat wall, the reflected wavefronts arrive at your ear exactly as though there were a second, identical source standing the same distance behind the wall. This is not a metaphor; it behaves like a mirror image in every geometric respect. It is why a reflection can genuinely sound like it is coming from behind a wall, and it is the basis of how acousticians predict where a room's early reflections will land — you draw the mirror images and follow the lines.

The pressure zone at a boundary

Right at a hard surface, something specific happens. The incoming wave and its own reflection overlap in the thin layer of air against the surface, and because they are effectively on top of each other there, they add. The pressure at the boundary is doubled — which, in decibels, is a +6 dB increase.

This is not a curiosity. It is the entire operating principle of boundary microphones — the flat plate-style mics you see taped to a stage floor or a conference table. By sitting in that pressure zone, the capsule gets a free 6 dB of sensitivity, and, more importantly, it cannot be comb-filtered by the surface it is lying on, because there is no gap between the direct sound and the reflection for a delay to happen in.

It is also why a speaker pushed into a corner sounds bass-heavy. Each nearby boundary reinforces the low end. Against one wall you get one reinforcement; in a corner, three surfaces meet and the effect stacks. Nothing is wrong with the speaker.

Curved surfaces focus, and that is a problem

A concave surface — a domed ceiling, a curved rear wall, a rounded alcove — does to sound what a satellite dish does to radio: it collects reflections and concentrates them at a point. That produces a hot spot where everything is unnaturally loud and, just as importantly, steals that energy from everywhere else in the room. In any space you have to work in, spreading reflections out is almost always better than focusing them, which is why diffusers exist and why perfectly curved architecture is an acoustician's headache.

2. Absorption

Not every reflection comes back. Some of the energy is converted to a tiny amount of heat inside the material it hits, and that portion is absorbed. How much gets absorbed depends on both the material and — critically — the frequency.

Hard, dense, sealed surfaces reflect: concrete, glass, tile, plaster, plywood. Soft, porous, fibrous ones absorb: heavy curtains, carpet, upholstered furniture, mineral wool, purpose-made acoustic panels.

The frequency dependence is the part that catches people out. Porous absorbers work by making moving air rub against fibres, so they only work where the air is actually moving fast — and high frequencies, with their short wavelengths, are far easier to catch than long low-frequency waves. The practical result is that almost every soft thing in a room absorbs treble far more effectively than bass. Hang enough duvets and foam in a bedroom and you will get a space with no top end left and exactly as much boom as before, which is the single most common outcome of a first attempt at treating a room. Fixing the low end takes thickness, depth and mass, not surface coverage — a subject with its own lesson later in the curriculum.

Even the air absorbs

Air itself absorbs high frequencies over distance, and how much depends on humidity. At normal room temperature a 10 kHz signal loses somewhere in the region of 0.15 to 0.3 dB per metre to the air alone, varying with how humid it is. Over a few metres in a studio this is negligible. Over fifty metres in an open-air venue it is the reason the back of the field sounds duller than the front, and no amount of EQ on the desk fixes it, because the loss happens after the PA. It is also a genuine distance cue your brain uses without being asked — a subject the localization lesson picks up.

3. Refraction

Refraction is what happens when a wave passes into a region where its speed is different: the wave bends.

Since the speed of sound depends on temperature, a layer of air at a different temperature will bend sound passing through it. On a normal day, air is warmer near the ground and cooler above, and sound bends upward and away from listeners at a distance. But when the layering inverts — cool air near the ground with warmer air above it, which happens over water and on still evenings — sound that would have escaped upward is bent back down toward the ground instead. That is why voices carry unreasonably far across a lake at dusk, and why an outdoor event that sounded fine at the afternoon soundcheck can behave differently after dark.

Inside a studio, refraction is effectively irrelevant — there are no meaningful temperature gradients across a control room. It is worth knowing because it explains outdoor behaviour that otherwise looks like equipment failure.

4. Diffraction

Diffraction is a wave bending around an obstacle, or spreading out after passing through a gap. Whether it happens depends entirely on the size of the obstacle compared with the wavelength — which is why the wavelength table from the previous lesson keeps earning its place.

If the obstacle is much smaller than the wavelength, the wave bends around it and carries on almost undisturbed. If the obstacle is much larger, the wave is stopped and reflected. Since low frequencies have long wavelengths and high frequencies have short ones, this means bass bends around things and treble does not.

Everything follows from that one sentence. It is why you hear the bassline from a car outside and not the vocal. It is why an acoustic screen between two musicians removes the highs from the bleed and leaves the lows, so the leakage arrives muddy rather than quiet. It is why standing behind a pillar at a gig muffles the sound rather than silencing it.

It also explains what your monitors do

A loudspeaker does not fire sound forward in a beam. Its output spreads, and how widely it spreads depends on frequency relative to the size of the driver. Low frequencies, with wavelengths far larger than the cabinet, spread almost equally in all directions — they radiate nearly omnidirectionally and take no notice of which way the box is pointing. High frequencies, with wavelengths much smaller than the driver, stay in a relatively narrow beam directly in front.

This is why the tonal balance of a monitor changes so dramatically as you move off-axis, why toe-in angle matters, and why bass reaches every corner of a room while the top end is only correct in one seat. Your listening position is not a preference. It is the only place the speaker is actually doing what its frequency-response graph claims.

What an absorption number actually means

Materials are rated by an absorption coefficient — a number from 0 to 1 describing the fraction of energy absorbed rather than reflected. A coefficient of 0 is a perfect reflector; 1.0 means everything is absorbed and nothing comes back.

The essential detail is that the coefficient is quoted per frequency band, and a single averaged figure hides everything that matters. A typical thin foam panel might absorb around 0.85 at 4 kHz and 0.10 at 125 Hz. Read as one averaged number that looks respectable. Read band by band, it says: this removes most of your treble and essentially none of your bass.

Which is exactly the failure mode of most first attempts at treating a room. Cover the walls in thin absorbers and you get a space whose treble has been stripped out while every low-frequency problem remains untouched — a room that is simultaneously dead and boomy, which is a genuinely unpleasant thing to work in and worse than the untreated version. Absorbing low frequencies requires depth and mass, because the absorber has to be thick relative to a wave that may be metres long. The acoustics specialization deals with how.

Distance: the inverse square law

Sound spreads out from a source in an expanding sphere. The same amount of energy has to cover a larger and larger surface as that sphere grows, so the energy per unit area drops off with distance — specifically, with the square of the distance.

In practical terms: every doubling of distance from a source costs you 6 dB. A guitar amp measuring 100 dB SPL at 30 cm gives you 94 dB at 60 cm, 88 dB at 1.2 m, 82 dB at 2.4 m. Halve the distance and you gain 6 dB back — the same amp reads 106 dB at 15 cm.

There is a large caveat, and it is the reason this lesson is not the last word on the subject. The inverse square law describes a free field — open space with no boundaries. A real room has reflections that do not obey it: after enough bounces the reflected energy is spread roughly evenly through the whole room, at close to the same level everywhere. So as you move a microphone away from a source, the direct sound drops by 6 dB per doubling while the room's contribution barely changes at all.

That single asymmetry is the most useful fact in microphone placement, and the recording lessons build on it directly: moving the mic is not a volume control, it is a ratio control between the source and the room.

Worked example: the desk in front of your microphone

Here is a real situation with real numbers. You are recording a voice into a microphone mounted on a desk. Your mouth is 25 cm above the desk surface and 20 cm horizontally from the mic. The mic capsule sits 10 cm above the desk.

The direct path from mouth to capsule is the straight line between them. The vertical difference is 25 − 10 = 15 cm and the horizontal is 20 cm, so:

√(20² + 15²) = √(400 + 225) = √625 = 25 cm

The reflected path bounces off the desk. Using the image-source trick from earlier: treat the reflection as coming from a mirror-image mouth 25 cm below the desk. Now the vertical difference to the capsule is 25 + 10 = 35 cm, and the horizontal is still 20 cm:

√(20² + 35²) = √(400 + 1,225) = √1,625 ≈ 40.3 cm

So the reflection travels 40.3 − 25 = 15.3 cm further than the direct sound. At 343 m/s, that extra distance takes:

0.153 ÷ 343 = 0.000446 seconds = 0.45 milliseconds

Now: 0.45 ms is far too short to hear as an echo. Your brain will not register two events. What it will register is that the microphone is summing two copies of the same voice, one slightly late — and that sum is not neutral. At some frequencies the two copies reinforce; at others they cancel. The first cancellation for a 0.45 ms delay lands at about 1.1 kHz, with further cancellations repeating every 2.2 kHz above it.

1.1 kHz is squarely in the region that carries vocal body and intelligibility. So the desk is not adding a reflection you can hear as a reflection. It is applying an aggressive, uneven filter to your voice, centred exactly where you can least afford it — and it will read on a spectrum analyser as a tonal problem with the microphone.

This is the most important idea in the lesson, and it is why the next lesson exists. A reflection that arrives too quickly to hear as a separate event does not disappear. It turns into a filter.

Reading a room in sixty seconds

You can learn a surprising amount about a space before unpacking anything. Stand in the middle and clap once, hard, then listen to what comes back.

• A short, dry thud with almost no tail — an absorbent room. Good for close work, and it will give you very little help making anything sound big.
• A ringing tail that hangs on — a reflective room. Every microphone will pick up substantially more room than you expect, and distance will become your main tonal control.
• A fluttering, buzzy zip after the clap — flutter echo, caused by sound bouncing repeatedly between two parallel hard surfaces. Very audible, very fixable: break up one of the two surfaces, or angle something across it.
• A tone in the tail — a specific note you can almost hum — a room resonance. The room is preferentially holding on to one frequency, and it will colour everything recorded in it.

Then walk the room while a speaker plays a steady bass note and listen for places where it disappears or doubles. Those positions are not faults in the speaker; they are the room reinforcing and cancelling itself at that frequency. Where you stand — and where you put the microphone — decides which version you record. There is a full lesson on why that happens and how to predict it in the acoustics specialization; for now, the useful habit is simply noticing that it does.

Putting the four together

Every recording space is doing all four of these at once, all the time:

1. Hard surfaces reflect — predictably, by geometry, and with a 6 dB pressure boost right at the boundary.
2. Soft and porous surfaces absorb — mostly the highs, unless they are thick enough to catch the lows.
3. Refraction bends sound through temperature layers — irrelevant indoors, decisive outdoors.
4. Diffraction lets bass bend around obstacles that stop treble, which is why partial barriers filter rather than block.

And underneath all of it, distance is quietly setting the balance between the sound you want and the room you did not ask for.

None of this is fixable with a plugin, because all of it happens before the microphone. It is fixable by moving things — the source, the microphone, the surfaces — which is free, and which is why experienced engineers spend so much of a session apparently just wandering around a room.

What to practice

• Clap once, hard, in the middle of your room, then again standing in a corner, then again facing a bare wall. Write down what changes. You are hearing reflection density, not volume.
• Put a hard-backed book flat on your desk under your microphone, record a few words, remove it, record again. The difference you hear is one reflection.
• Walk a speaker playing pink noise from the centre of a room toward a corner and listen to the low end build. That is boundary reinforcement, and it is why monitor placement matters more than monitor price.
• Stand a metre outside an open doorway with music playing inside, off to one side so nothing is line-of-sight. Note which parts of the music still reach you and which do not.