Measuring passive gain
The archive claims up to 8 dB and calls its performance lab tested. A gain figure is only meaningful alongside the comparison that produced it.
A procedure you can actually run
What you need
A measurement microphone or a phone app with a calibrated input, a tape measure, a tripod or a stack of books, and a pink-noise file. The absolute accuracy of the microphone does not matter here — both readings in the pair share the same error, and the difference is what is being measured.
The order that keeps the comparison honest
Set the source playing pink noise at a fixed volume. Place the microphone on the horn's axis at a stated distance — a metre is convenient. Record the bare device lying as it normally would. Without touching the volume, the microphone or the distance, put the device in the cradle and record again. Subtract. Then repeat the whole pair at thirty degrees off axis, which is where directivity shows itself.
Two sanity checks before trusting the result
Double the distance and confirm the reading falls by about 6 dB; if it does not, the microphone was in the near field and both numbers are inflated. Then repeat the pair with a folded towel under the device. If the gain changes substantially, part of what was measured belonged to the desk rather than to the horn.
A gain figure is a difference, never a property
Decibels are a ratio. “8 dB” on its own states nothing: 8 dB louder than what, at what distance, on what axis, over what band. The only defensible procedure is a matched pair of measurements where exactly one variable changes.
That means the device is measured bare, then measured in the cradle, with the microphone, the distance, the source material and the volume setting all untouched. The difference between the two curves is the gain. Any figure quoted without that pairing is marketing, including when it happens to be true.
The band is where honest figures separate from flattering ones
A horn does not apply a uniform lift. It raises the band it is sized for and leaves the rest roughly alone, so a single number is always a summary of something uneven. Which summary is chosen changes the answer enormously.
A peak figure takes the best third-octave band in the response. A band-averaged figure takes the mean across a stated range. The first can easily be twice the second on the same hardware. Neither is dishonest when stated; quoting a peak as though it were an average is.
Four ways a measurement overstates a horn
On-axis only. A horn narrows its output. Measuring solely at the mouth's centre line records the directivity gain at its maximum and misses that it costs output everywhere else.
Close microphone. Near the mouth the field has not yet settled into its far-field pattern, and readings run high. Measurement should be far enough out that doubling the distance costs the expected 6 dB.
Hard desk underneath. A reflective surface adds its own reinforcement. The same cradle measures differently on glass and on felt, and the difference is not the horn's doing.
Ignoring the volume control. Phone output stages compress near maximum. Measuring the bare device at full volume and the cradled one at the same setting can quietly compare a limited source with an unlimited one.
| Item | Value |
|---|---|
| Quantity | Difference in SPL, bare versus cradled |
| Must be stated | Distance, axis, band, source level |
| Peak figure | Best third-octave band |
| Averaged figure | Mean over a declared range |
| Archive claim | Up to 8 dB; band and method not published |
| Audible reference | About 10 dB reads as twice as loud |