Skip to content

Home theatre systems

Subwoofers and bass management

Why one subwoofer gives every seat different bass, how several subwoofers fix it, and how bass management, crossovers and the LFE channel actually work.

8 min read7 sourcesChapter 18 of 24

node · almost nothing at this noteloud herequiet herewallwallfirst axial mode — pressure swings between the walls, zero in the middle

Schematic

A standing wave between two walls. Pressure is high at the walls and zero in the middle. A seat on the null hears that note vanish; a seat at the wall hears it boom. A second subwoofer placed to drive the mode the other way can cancel it.

Animated illustration of the principle. Not measured data.

Most of the complaints about home theatre bass — boomy in one seat, thin in the next, a note that rings on — come from the room’s modes, not from the subwoofer. The cure starts with how many subwoofers there are and where they go, long before equalisation.

One subwoofer, many different seats

Below a few hundred hertz a small room behaves as a set of resonances. Toole puts the transition region in listening rooms at around 300 Hz. Below it, the bass level at each seat depends on where that seat sits in the room’s modal pattern, so the response changes significantly from one seat to the next.

That is why equalising a single subwoofer cannot fix bass for a room. As Welti put it, “equalizing the sound flat at one seat will simply make the frequency response less flat at another seat.” Equalisation is only effective once the seats already hear similar bass.

Several subwoofers

Welti and Devantier’s research at Harman showed that “through the use of multiple subwoofers the seat-to-seat variation in the frequency response can be reduced significantly, allowing subsequent equalization to be more effective”. In their study of rectangular rooms, subwoofers at the midpoint of each wall gave the lowest seat-to-seat variation, with subwoofers in all four corners almost as good and with stronger low-frequency output; four subwoofers were enough to get the best results of any configuration tried, and adding more did not keep improving things.

The principle is that subwoofers in the right places drive particular modes against each other, so the modes that cause the worst seat-to-seat differences are not excited. Real rooms are not perfect rectangles, so the positions are worked out for the room in plan and then confirmed at the seats.

Two subwoofers at the front wall, each a quarter of the room’s width in from the side walls, is a common and effective starting point when four are not possible.

Bass management and the LFE channel

Bass management sends the low frequencies from every loudspeaker channel to the subwoofers, so the main loudspeakers are relieved of deep bass and the bass is reproduced by the drivers best placed to do it. THX requires an 80 Hz crossover point in certified receivers and recommends setting the loudspeakers to “Small”; 80 Hz is the usual default because bass below it is difficult to locate.

The low-frequency effects channel — the “.1” — is a separate channel in the soundtrack, limited to below 120 Hz and replayed 10 dB louder than the other channels in its band (ATSC A/52, SMPTE RP 200). It is not “the subwoofer”: after bass management, the subwoofers play the LFE channel plus the redirected bass from all the others.

What equalisation can and cannot do

A peak — a mode that is too loud — can be cut with equalisation, and the cut helps every seat that hears that peak. A dip caused by cancellation is different: boosting it also boosts the cancelling wave, so the dip stays while the amplifier and driver work harder. Genelec’s own room-correction guidance explains that dips depend on listening position and that boosting can, at worst, leave a peak where the dip used to be.

So the order is: positions first, number of subwoofers second, level, delay and phase alignment at the crossover third, and cut-only equalisation of what the seats have in common last. Low-frequency absorption in the room — covered in the bass-trap chapter — shortens the ringing that equalisation cannot touch.

For integrators and architectsThe numbers, the standards, and what to specify
Multiple subs
Welti & Devantier (JAES 2006): multiple subwoofers reduce seat-to-seat variation and make EQ effective. Welti: four at wall midpoints lowest variation; four corners close behind with more LF output; >4 gave no consistent improvement (rectangular rooms, centred seating).
Crossover
80 Hz (THX certified AVR requirement); speakers set Small. Raise only for small satellites, and keep all channels on the same crossover where possible.
LFE
Band-limited <120 Hz, +10 dB in-band gain (ATSC A/52 §4.3.1; SMPTE RP 200 §4.7). Subwoofer level set so redirected bass and LFE both play at their intended levels.
EQ
Cut peaks shared across seats; do not boost cancellation dips (Genelec). Align sub delay/phase to the mains through the crossover before EQ.

What to write into the brief

  • Subwoofer count and positions drawn in plan, with the reasoning for the room shape, before the front wall and any riser are built.

Sources

  1. 01Welti & Devantier, “Low-Frequency Optimization Using Multiple Subwoofers”, JAES 54(5), 2006
  2. 02Welti (Harman) — Subwoofers: optimum number and locationsFree slide paper; results for rectangular rooms.
  3. 03Welti, “How Many Subwoofers are Enough?”, AES 112th Convention, 2002
  4. 04Toole, “Loudspeakers and Rooms for Sound Reproduction — A Scientific Review”, AES 2006The ~300 Hz transition region in listening rooms.
  5. 05THX — certified AV receivers (80 Hz crossover requirement)
  6. 06ATSC A/52:2018 — Digital Audio Compression (AC-3, E-AC-3) StandardLFE channel bandwidth and +10 dB level.
  7. 07Genelec — Why does GLM not fix dips in the frequency response?

External links open in a new tab. Figures quoted in this chapter are taken from these sources; a Science of Sound measurement is only ever presented as one, with its file, date and instrument.

Start a project

Want this applied to your room?

Tell us what it's for, its size and where it is. We'll say what we'd measure first and what we'd change.