Direct sound, reverberant sound, and the distance between them

The one number that decides how a recording sounds

Every recording is a mix of two things: direct sound, which travelled straight from the source to the microphone, and reverberant sound, which bounced off surfaces first. The ratio between them is what people hear as “roomy” or “close”, “amateur” or “professional”.

The physics is simple and unforgiving.

Direct sound obeys the inverse square law: every doubling of distance costs 6dB. At 15cm from your mouth, then 30cm, then 60cm, then 1.2m, the direct sound falls by 6dB each step.

Reverberant sound does not behave that way. Reflections have bounced around and filled the room fairly evenly, so the reverberant level is roughly constant everywhere beyond a short distance from the source.

So as you move back, direct falls and reverberant stays. The ratio collapses. The distance at which they are equal is the critical distance, and in an untreated domestic room it is often only 0.5 to 1.5m. Beyond it you are recording mostly room.

This is why the single most effective thing you can do costs nothing: get closer. Moving from 60cm to 15cm improves the direct-to-reverberant ratio by 12dB, which is a larger change than any microphone upgrade will give you.

Absorption and isolation are different problems

These get conflated constantly and the confusion wastes money.

Absorption (acoustic treatment) reduces reflections within a room. Soft, porous materials convert sound energy to heat as air moves through them. It makes a room sound less boxy. It does essentially nothing to stop sound getting in or out.

Isolation (soundproofing) stops sound passing between spaces. It requires mass, airtightness and decoupling: a heavy wall, sealed gaps, separated structures. Foam on the wall provides none of these.

A duvet over your head will make your voice sound dramatically better and will not reduce the neighbour’s drill by a decibel. If your problem is a noise coming in, treatment is the wrong tool and there is rarely a cheap answer.

Why thin foam does not fix a boomy room

A porous absorber works by air movement, and air movement is greatest a quarter-wavelength from a hard surface. So an absorber is effective for frequencies whose quarter-wavelength is comparable to its thickness plus any air gap behind it.

Run the numbers. Sound travels about 343 m/s. At 100Hz the wavelength is 3.4m and a quarter of that is 86cm. At 1kHz the wavelength is 34cm and a quarter is 8.5cm.

So 5cm of foam absorbs well above roughly 1kHz, weakly around 250 to 500Hz, and almost nothing at 100Hz. The tiles on the wall are removing exactly the frequencies that were not the problem, while the boom (which is low-frequency room modes) remains untouched. This is why bass traps are thick, why they go in corners where pressure is highest, and why a serious low-frequency fix occupies real volume.

Thick porous absorbers with an air gap behind them work far better per pound than thin foam glued flat. A 10cm mineral-wool panel mounted 10cm off the wall behaves, roughly, like a 20cm absorber.

Flutter echo, and the cheapest test in audio

Stand in the middle of the room and clap once, sharply. Listen to what comes back.

• A short, even decay: good.
• A metallic ringing or a rapid “brrrp”: flutter echo, sound bouncing between two parallel hard surfaces. Break it by putting something soft or angled on one of the two facing walls. One wall is enough.
• A boomy tail: low-frequency modes, which need thickness.

This takes two seconds, needs no equipment, and tells you more about a room than a specification sheet.

What actually works in a spare room

In rough order of effect per effort:

1. Get closer to the microphone. Free, and the largest single improvement.
2. Do not stand in the middle of the room, and do not face a bare wall closer than about a metre. Record into the room’s long dimension.
3. Put something soft behind the microphone, in the direction it is facing past you: a duvet on a clothes rail, an open wardrobe of hanging clothes, a heavy curtain. You are absorbing the reflection that would otherwise return straight into the capsule.
4. Break the two parallel pairs: one soft or angled surface on one wall of each pair.
5. Soft floor. A rug removes the desk-floor-ceiling path.
6. Thick panels in corners, if you are building anything.

The wardrobe trick is effective and frequently mocked. Hanging clothes are a deep, porous absorber with an air gap, which is the correct construction; they work better than the foam tiles that cost more.

Measuring instead of guessing

REW (Room EQ Wizard) is free, runs on all three desktop platforms, and will measure your room’s impulse response with a sweep played through a speaker. It gives you an RT60 figure (the time for sound to decay by 60dB) per frequency band, which turns “it sounds boxy” into “there is a 0.6 second tail at 160Hz”.

For speech, an RT60 below about 0.4 seconds is comfortable and below 0.3 is good. Domestic rooms with hard floors commonly measure 0.6 to 1.0 seconds, and that is what makes a recording sound like a kitchen.

The free path

REW for measurement. Audacity or Ardour to record test clips and compare. Materials: duvets, curtains, a rug, an open wardrobe, and mineral wool or rockwool in fabric-covered frames if you build panels: the material costs a fraction of branded acoustic products and performs measurably better than thin foam.

The limitation

Treatment has a floor. You can make a small room sound neutral; you cannot make it sound like a large one, because a small room’s low-frequency modes are widely spaced and no amount of absorption removes the geometry. A 3-metre dimension supports a mode at around 57Hz and its multiples, and that is a property of the space.

And none of this addresses noise coming in. If a road, a flight path or a neighbour is the problem, acoustic treatment will not help, the honest options are expensive or involve moving, and the practical answer is usually to record at a different time of day.