How stereo creates direction: level, time and equivalence

Two loudspeakers can place a sound anywhere between them. That is not a trick of the loudspeakers but of hearing: the brain combines the two signals reaching each ear into one perceived position, the phantom source.

Two cues that carry direction

A recording can tell the listener where a sound belongs in exactly two ways:

• Level difference (ΔL): the sound is louder in one channel. Panpots and coincident microphone pairs work this way.
• Time difference (Δt): the sound arrives earlier in one channel. Spaced microphones work this way.

Both cues have natural counterparts. A sound from the right is louder and earlier at the right ear, which is what we use in everyday hearing, see How we hear direction.

How much is enough?

For two loudspeakers at ±30° in front of the listener, a few round numbers describe the behaviour:

Smaller differences give positions in between. The relationship is not a straight line: the first few decibels or tenths of a millisecond move the image quickly, and the last stretch moves it slowly.

Equivalence: combining both cues

The two cues add. A source that is 6 dB louder in the right channel and arrives 0.4 ms earlier there ends up further right than either cue alone would place it. Microphone arrangements that use both on purpose are called equivalence systems; ORTF is the best-known example, see Equivalence systems.

People often describe this as time-intensity trading, as if one cue could be swapped for the other at a fixed rate. That is a simplification. The two cues sound different (a time difference gives a more diffuse, spacious image than a level difference), and the exchange rate depends on frequency and signal. Treat any single figure as a rough guide.

Why it matters for microphone choice

• A coincident pair has only level differences, so the image is precise and mono compatible, but it can sound less spacious.
• A spaced pair has only time differences (plus whatever the patterns add), so it sounds spacious but has less precise localisation and a weaker mono sum.
• Near-coincident pairs mix the two.

Try it

The stereo recording angle tool shows both cues and how they add up for any microphone setup.

Draft: written by an AI assistant, awaiting technical review.