What Sensitivity Means
Speaker sensitivity is self-explanatory once you understand how it's measured. Start by placing a measurement microphone or SPL (sound pressure level) meter exactly one meter away from the front of the speaker. Then connect an amplifier to the speaker and play a signal; you'll want to adjust the level so the amplifier delivers only one watt of power to the speaker. Now observe the results, measured in decibels (dB), on the microphone or SPL meter. That's the sensitivity of the speaker.
The higher the sensitivity rating of a speaker, the louder it will play with a certain amount of wattage. For example, some speakers have a sensitivity of around 81 dB or so. This means with one watt of power, they'll deliver just a moderate listening level. Want 84 dB? You'll need two watts — this is due to the fact that every additional 3 dB of volume requires double the power. Want to hit some nice and loud 102 dB peaks in your home theater system? You'll need 128 watts.
Sensitivity measurements of 88 dB are about average. Anything below 84 dB is considered rather poor sensitivity. The sensitivity of 92 dB or higher is very good and should be sought after.
Are Efficiency and Sensitivity the Same?
Yes and no. You'll often see the terms sensitivity and efficiency used interchangeably in audio, which is ok. Most people should know what you mean when you say a speaker has 89 dB efficiency. Technically, efficiency and sensitivity are different, even though they describe the same concept. Sensitivity specifications can be converted to efficiency specifications and vice-versa.
Efficiency is the amount of power going into a speaker that is actually converted into sound. This value is usually less than one percent, which tells you that most of the power sent to a speaker ends up as heat and not sound.
How Sensitivity Measurements Can Vary
It's rare for a speaker manufacturer to describe in detail how they measure sensitivity. Most prefer to tell you what you already know; the measurement was done at one watt at a one-meter distance. Unfortunately, sensitivity measurements can be performed in a variety of ways.
You can measure sensitivity with pink noise. However, pink noise fluctuates in level, which means it's not very precise unless you have a meter that performs averages over several seconds. Pink noise also doesn't permit much in the way of limiting measurement to a specific band of audio. For example, a speaker that has its bass boosted by +10 dB will exhibit a higher sensitivity rating, But it's basically cheating because of all the unwanted bass. One could apply weighting curves — such as A-weighting, which focuses on sounds between about 500 Hz and 10 kHz — to an SPL meter to filter out the frequency extremes. But that's added work.
Many prefer to evaluate sensitivity by taking on-axis frequency response measurements of speakers at a set voltage. Then you would average all the response data points between 300 Hz and 3,000 Hz. This approach is very good at delivering repeatable results with accuracy down to about 0.1 dB.
But then there's the question of whether sensitivity measurements were done anechoically or in-room. An anechoic measurement considers only the sound emitted by the speaker and ignores reflections from other objects. This is a favored technique, being that it's repeatable and precise. However, in-room measurements give you a more real-world picture of the sound levels emitted by a speaker. But in-room measurements typically give you an extra 3 dB or so. Sadly, most manufacturers don't tell you if their sensitivity measurements are anechoic or in-room — the best case is when they give you both so you can see for yourself.
