“Sensitivity: 88 dB” is the figure most often used to compare loudspeakers, and also one of the most often misread. Two datasheets may both say 88 dB while one refers to 1 W / 1 m, the other to 2.83 V / 1 m, and a microspeaker datasheet may quote 0.1 m; the underlying difference can be 3 dB, 20 dB or more. This article sorts out the common sensitivity conventions and how to convert between them, derives the reference efficiency η₀ from Thiele/Small parameters, explains why a moving-coil driver converts only a few tenths of a percent of its input into sound and what it costs to raise that figure, and closes with distance and power scaling and a checklist for reading specifications.

• 112 dB1 W / 1 m at 100% efficiency
• 0.31%η₀ of the example woofer
• +3 dB4 Ω driver rated at 2.83 V
• −6 dBPer doubling of distance

1. Two Conventions: 1 W / 1 m and 2.83 V / 1 m

Sensitivity is the on-axis sound pressure level at a stated distance for a stated input, expressed in dB re 20 μPa. Two conventions dominate:

• 1 W / 1 m: the driver receives “1 W” and is measured at 1 m. The watt is not a measured real power, because a loudspeaker’s impedance varies strongly with frequency and so does the power it actually absorbs. It is a voltage derived from the rated impedance: U = √(P × R_n), which is 2.83 V for an 8 Ω driver and 2.0 V for a 4 Ω driver.
• 2.83 V / 1 m: the input is simply 2.83 V (√8 V, exactly 1 W into 8 Ω) regardless of the driver’s impedance. Since an amplifier behaves essentially as a voltage source, this convention is closer to real use.

For an 8 Ω driver the two numbers coincide. For lower impedances, 2.83 V delivers more power and the figure rises accordingly:

IEC 60268-5 (GB/T 12060.5 in China) also defines a characteristic sensitivity in a stated frequency band: the on-axis pressure is averaged over the band and referred to the reference input and a distance of 1 m. A single-frequency value moves with every ripple in the response, so a band average is better for comparison, but manufacturers choose different bands and a datasheet should state which one it uses.

2. Reference Efficiency η₀ from Thiele/Small Parameters

Above the resonance frequency f_s, and as long as the diaphragm still moves as a rigid piston (ka < 1, where a is its effective radius), the moving system is mass-controlled: the coil force is F = Bl · i and the diaphragm acceleration is a_d = F / M_ms. Mounted in an infinite baffle and radiating into half space, the far-field pressure is proportional to the volume acceleration:

Frequency does not appear: this is the flat “piston band” of the response. The ratio of radiated acoustic power to electrical power e²/R_e is the reference efficiency (Small, 1972):

Spreading the acoustic power produced by 1 W of electrical input uniformly over a hemisphere gives the link between efficiency and the 1 W / 1 m level:

Example: a 6.5-inch-class woofer

Take an illustrative set of parameters: R_e = 3.4 Ω, f_s = 55 Hz, Q_ms = 4.0, Q_es = 0.55, M_ms = 20 g, S_d = 133 cm². From Q_es = 2πf_sM_msR_e / (Bl)² we get Bl ≈ 6.54 T·m and V_as ≈ 10.5 L, and then:

Reference efficiency

η₀ ≈ 0.0031, or 0.31%. Both formulas give the same value.

1 W (into R_e) / 1 m

112.1 + 10 log₁₀0.0031 ≈ 87.1 dB.

2.83 V / 1 m

2.83 V across 3.4 Ω is about 2.36 W, giving about 90.8 dB; the pressure formula above gives the same result directly.

1 W (rated 4 Ω) / 1 m

An input of 2.0 V gives about 87.7 dB, 0.7 dB (10 log₁₀(4/3.4)) above the figure for 1 W into R_e. The “1 W” on a datasheet is normally this rated-impedance voltage.

The estimate holds only in the piston band. The example driver has an effective radius of about 6.5 cm, so ka = 1 at roughly 840 Hz. Above that, radiation concentrates on axis (directivity) while the coil inductance reduces the current, and the measured response departs from the plateau. A 1 to 2 dB gap between the T/S estimate and a datasheet’s band-averaged sensitivity is therefore unsurprising. The two serve different purposes: the estimate supports design and consistency checks, while the measured figure is the one to specify.

3. Why Efficiency Is a Fraction of a Percent: Bl, M_ms and the Bass Trade-off

Writing the first formula as η₀ ∝ (Bl · S_d / M_ms)² / R_e shows what each parameter does:

• Moving mass M_ms: efficiency falls with its square. A 20% heavier moving system loses 1.6 dB. The cone, voice coil, former, surround and the air load moving with the cone all count.
• Diaphragm area S_d: efficiency rises with the square of the area, but a larger cone is usually heavier and lowers the frequency at which ka = 1.
• Force factor Bl: efficiency rises with Bl². More Bl needs a stronger magnet system (neodymium, a narrower gap) or more wire in the gap, and the latter adds resistance and coil mass.

The price is paid at low frequencies. Since Q_es = 2πf_sM_msR_e / (Bl)², electrical damping grows quickly with Bl. Changing only Bl in the example driver gives:

A stronger motor buys midband sensitivity, but the bass starts rolling off earlier. Restoring it takes a larger enclosure or equalisation, and equalisation spends the power that was saved. The second formula states the constraint directly: η₀ is proportional to f_s³V_as, so deeper bass from the same box volume means lower efficiency. You cannot have low-frequency extension, a small enclosure and high efficiency all at once. This constraint is widely known as Hofmann’s Iron Law.

4. Scaling with Distance and Power

In a free field (no reflections) and in the far field, pressure falls inversely with distance, and level rises with the logarithm of input power. Given a 1 W / 1 m sensitivity L_s, the level at power P (W) and distance r (m) is:

Example: an 88 dB loudspeaker fed 10 W produces about 88 + 10 − 9.5 = 88.5 dB at 3 m; reaching 100 dB at 3 m would take about 143 W. A driver 3 dB more sensitive needs only half the power for the same target, and in battery-powered or automotive systems with a limited supply those few dB are often worth more than the difference in amplifier ratings.

5. Microspeakers: Reading a 0.1 m Rating

Microspeakers in phones, laptops and Bluetooth speakers are small and used close to the ear or listener, so their sensitivity is commonly stated at 0.1 m for a given power or voltage, measured on a specified test baffle or in a test box with a defined back volume. In the far field, converting from 0.1 m to 1 m subtracts 20 log₁₀(1 / 0.1) = 20 dB.

Applying the same formulas to an illustrative microspeaker (R_e = 7.2 Ω, f_s = 800 Hz, M_ms = 0.10 g, S_d = 1.2 cm², Q_es = 1.2, giving Bl ≈ 1.74 T·m):

For a microspeaker, 0.1 m is already the far field: with an effective radius of about 6.2 mm, the Rayleigh distance πa²/λ is only about 3.5 mm even at 10 kHz. What really changes the reading is the mounting. The back volume shifts the low-frequency resonance and its Q, the front cavity and sound port form a Helmholtz resonance at high frequencies, and a different test box or baffle can change the curve of the same unit noticeably. Before comparing data from different sources, make sure the test fixtures match.

6. Checklist for Datasheets and Measurements

1. Check the input reference

   Watts or volts, and if watts, which impedance was used to convert them to a voltage. A 4 Ω driver reads 3 dB higher at 2.83 V than at 1 W.

2. Check the distance and the acoustic field

   1 m or 0.1 m; half space (baffle), test box or full space; anechoic or not. Use the 20 log₁₀ distance rule only in the far field.

3. Check the frequency range

   A single-frequency value or an average over a stated band, and which band. With an uneven response the two can differ by several dB.

4. Normalise before comparing

   Convert every figure to the same voltage and distance, then account for the extra amplifier current and power a lower impedance demands.

5. Evaluate large-signal behaviour separately

   Sensitivity is a small-signal figure. Maximum SPL is also limited by excursion, distortion and voice-coil heating, so look at rated power, power compression and distortion data.