Speaker Sensitivity


1	Average SPL	Average Difference		Breakdown
2	85.8dB	-1.5dB		Type	Ported	Sealed	PR	Dipole
3			Count	63	17	2	1
4	Number of speakers		Average SPL	85.9dB	85.7dB	86.2dB	83.0dB
5	83		Type	Standmount	Floorstander	Center	Wallmount
6				Count	66	7	5	5
7				Average SPL	85.4dB	88.5dB	88.5dB	84.8dB


1	Brand	Comment	Source

2	Ascend Acoustics	”...our sensitivity measurement is anechoic, with a microphone distance of 1 meter and an input voltage of 2.83v. We use the averaged amplitude over the range of 300Hz - 3kHz. These are the current but not-so-widely-accepted standards.”	David Fabrikant (Owner)
3	ATC	“We measure the loudspeaker on axis in essentially an anechoic environment. We apply no weighting to the response. We then take the average level through the linear region and quote this figure.”	Richard Newman (Engineering Manager)
4	DefTech	Bipole: frequency response is given as "total frequency response" with no specifics regarding the tolerance of the spec (+/- dB) nor how it is derived. Monopole: Measurements are made in our quasi anechoic chamber.
5	Dynaudio	No response
6	ELAC	No response
7	Focal	“My technical contact in France is away until the end of July, but I seem to remember it was A-weighed in anechoic...I'm pretty sure it is done at 1 kHz...I can safely say that most models are fairly efficient and easy to drive.They do not require a lot of watts, but respond better to high current, high quality electronics. For instance, you would often get better results from a high quality, hand made 75w European or American product, than you would from a 150w Japanese product.”	Michel (Support)
8	GoldenEar	“Our measurements are Anechoic.” When asked about knowing the freuqncy range: "I do not."	Christopher Volk (Eastern Region DIrector of Sales)
9	HTD	No response
10	KEF	“Generally most companies take their sensitivity rating at 1kHz – 1 meter – 2.83 volts (1 watt). Weighting doesn’t matter as it’s a constant frequency/power relationship. In room sensitivity is really not a thing as sensitivity has nothing to do with in-room acoustics but is really just how a speaker reacts to the power applied to it. Our sensitivity ratings are always somewhat conservative, meaning what we publish is typically a little lower than the actual rating – not by a lot but a dB or so maybe.”	Troy (Service)
11	Klipsch	Measurements are taken from a distance from the speaker baffle to replicate actual in-home listening and a real-world scenario...Please note that response measurements taken close to the baffle do not accurately depict a system's total sound delivery into the room, and therefore are not representative of the true-to-life experience.	Product Support
12	MartinLogan	“For floor standing ESL speakers: These are measured from approximately 3.2 meters back, with the speaker placed in a parking lot, and with the speaker toed in per the instructions. The microphone is placed at the same position as an optimum listeners ear would be placed. We then compensate the SPL measurement by the distance to calculate an equivalent 2.83V @ 1 meter reading. The Average SPL is calculated by averaging the measured SPL over the speakers rated frequency response window. For non ESL speakers: Very similar process and setup to the above, but these are measured in room, (a typical sized listening room, not an anechoic chamber) at around 2 meters back. We then perform similar compensations and calculations to get to the standardized 2.83V @ 1m specification.”	Meagan Lohmeyer (Director of Inside Operations)
13	Micca	No response
14	NHT	“We use the average anechoic SPL over a range (~500-2KHz) @ 1M, on-axis with 2.83 VRMS driving voltage.”	Gordon Chang (Engineer)
15	Pioneer	No response
16	PMC	Requested which speakers and amps were being discussed to determine whther they would be a good pairing	Brent Esry (Manager of Custom Instalation & Hi-fi)
17	Polk	No response
18	Q Acoustics	“Our measurements are carried out on axis in a anechoic chamber @ 1m, 1 kHz, 2.83Vrms”	Support
19	Revel/Harman	"Our sensitivity ratings are based on the 1 watt voltage onto the nominal impedance (so, 2.83V for 8 Ohm, etc.) curve in our chamber. The design engineer reports the sensitivity based on the curve. We calculate the nominal impedance per industry standard – 1.15 x impedance at the mid-bass minimum, rounded to the closest of 4 Ohm, 6 Ohm, or 8 Ohm. Since we have had many varying engineers over the years, I have seen various methods by the engineer in our files, including reporting peak sensitivity. Having said that though, most of the engineers call out an average sensitivity, trying to pick a value between dips and peaks over as much of the band from 100 Hz to 20 kHz as possible, depending on smoothness of the high frequency reproduction. We typically don’t look at the level below 100 Hz for this because it is so room dependent, but we do aim for close to linear in our tunings, so this band doesn’t affect the overall average anyway. So, this is more of a C-weighted visual average over the full frequency band. It is done on axis because that is typical by reviewers. Since many companies aren’t as rigorous as this and many don’t follow the proper method to calculate nominal impedance, it is better to experience the speaker than to hold this reported number as an absolute indicator of what a speaker will actually do."	Tom Pfister (Field Support Engineer)
20	Sonus Faber	“...sensitivity ratings measurements are obtained as follows: In our semi-anechoic room (coherent and certified since 100Hz), Not weighted, Average 100-20.000Hz, 1watt @ 1mt, On-axis between mid and tweeter”	Stephano (Customer Service)
21	Sony	“We would like to inform you that we do not have the information that you seek.”	Support


1	SPL reported at 2.83V @ 1 meter
2	SPL reported at 2.83V @ 1 meter
3	Measured SPL is calculated by taking the average, non-weighted SPL of the 300Hz-3000Hz region* of the on-axis	*With the exception of the Magnepan LRS, with 500Hz used per manufacturer spec; dipole speakers have substantial backwave SPL even at 3kHz, thus the anechoic SPL on-axis does not accurately represent the SPL in-room.
4
5	Data courtesy of Amir Majidimehr of AudioScienceReview
6	Data is of ~1/20 octave resolution