Phantom sources: where a sound appears between two speakers

A phantom source is the sound image you perceive between two loudspeakers although nothing is radiating from that spot. Understanding how it forms explains why stereo works, and where it breaks.

Summing localisation

When two loudspeakers play the same signal at slightly different levels or times, the ears receive both. Within a window of roughly one millisecond between the two arrivals, the brain does not hear two sources. It merges them into one image whose direction is a weighted average. This is called summing localisation.

• Equal level, equal time: the image sits in the middle.
• One channel louder: the image moves towards that loudspeaker.
• One channel earlier: the image moves towards that loudspeaker as well.

The first wavefront

If the delay between the two signals grows beyond about 1 ms, the image no longer averages. The earlier signal dominates the direction, and the later signal only changes the width and sense of space. This is the law of the first wavefront, also known as the precedence effect, see The precedence effect. At about 1.5 ms of delay, the image is already fully in the earlier loudspeaker.

What the listener must do

Phantom sources are fragile in one respect: they rely on the listener sitting in the sweet spot, on the centre line between the loudspeakers.

• Move towards one loudspeaker and its signal arrives earlier and louder. The whole image collapses towards that side.
• Sit very far off to the side and the picture is gone.
• Dull the room (more absorption) and the image gets clearer; strong early reflections blur it.

Some limits

1. Frequency dependence. Level differences work well at all frequencies. Time differences work best at low frequencies, below about 1.5 kHz, where the ear can follow the waveform.
2. Widths beyond the loudspeakers. Ordinary stereo cannot place sources outside the line between the loudspeakers. Special processing (crosstalk cancellation, wide-base techniques) can, but only for a fixed listening spot.
3. Image stability. Images that come from a level difference are generally sharper than images from a time difference.

Try it

In the stereo recording angle tool, the last section shows where a source at a given angle would appear between the loudspeakers for a chosen microphone setup.

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