February 3, 2011

How to avoid being misled by speaker sensitivity ratings

Audio enthusiasts are very often misled regarding speaker sensitivity ratings. Manufacturer's vary greatly in the degree of accuracy in the way they use this rating. They typically fall into one of three categories. Firstly, those in the minority who are brutally accurate. Secondly, those with slightly optimistic ratings and thirdly those that are downright misleading. The latter can be a problem if you need efficient speakers to work with a low powered amp. In this article you can arm yourself with sufficient knowledge to read into the specs and determine the truth. You will be able to look at the most misleading and inflated numbers and determine their true sensitivity.

Understanding voltage and watts

Speaker efficiency can be rated in two ways. Voltage sensitivity is defined as the output in decibels at one metre with 2.83V input. As I will show, this rating can be misleading because the power input varies according to impedance. 2.83V into an 8 ohm load equates to 1w of input power, but the power input will be different into any other impedance. I prefer the other method of rating sensitivity, which shows 1w input so that different speakers can be compared.

2.83V into 8 ohms = 1 watt
2.83V into 4 ohms = 2 watts (Voltage rating is 3 dB higher)
2.83V into 2 ohms = 4 watts (Voltage rating is 6 dB higher)

If you are comparing speakers, I suggest you convert the Voltage sensitivity into a 1w equivalent so a fair comparison is made. In some cases this may be difficult where a complex impedance load is presented.

How to estimate sensitivity

If you know the drivers used in a speaker and have access to the data sheet, you can estimate it's sensitivity.


In this example, the Vifa P17 is shown. It is a good example because it is a well behaved driver with it's sensitivity shown correctly. Some data sheets are also quite optimistic and you have to look at both the nominal sensitivity and the response plot and note the conditions under which it was measured. This is a text book example with a smooth response and a nominal 88 dB rating that agrees with the plot.

If this driver were used in a 2 way speaker, it could achieve no greater than 88 dB 1w1m if it were designed to be placed right up against a wall. In this scenario, the wall provides reinforcement of bass and midrange frequencies.

If it were designed as an audiophile stand mount or floorstander, then it would be placed typically into the room. This means boundary gain is given up, and a typical design would allow for this in the crossover. The result is at least 3 dB lower sensitivity. We could now expect 85 dB 1w1m. The effect referred to here is bafflestep compensation, and it is included in every well designed speaker that is not specifically intended for near wall placement.

If another identical woofer is added to cover the same range, in either a 2 way or 2.5 way design, the sensitivity can now return to 88 dB 1w1m. An equivalent voltage sensitivity would be 91 dB since the impedance would now be 4 ohms.

Estimating without data sheets

Where you only have basic specs, here are some suggested ratings.

Speakers with single vented woofers (add 3 dB for a second woofer), sensitivity 1w1m:

5" woofer ~ 65 Hz (-3 dB): 85 dB
6.5" woofer ~ 40 Hz (-3 dB): 85 dB
8" woofer ~ 40 Hz (-3 dB): 88 dB
10" woofer ~ 25 Hz (-3 dB): 85 dB
12" woofer ~ 22 Hz (-3 dB): 86 dB

For a sealed box, the extension is reduced by one octave for a given efficiency. So if the above 12" woofer were placed in a sealed box it would have the same efficiency but around 44 Hz extension.

Optimistic ratings are common

Do you find these number surprising? They are significantly lower than typical ratings published. The problem is that it's very common to show very optimistic numbers. A manufacturer faces a dilemma. If they show an accurate rating, they will be unfairly disadvantaged while others compare their speakers to inflated ratings. Suppose they were to show a rating of 85 dB 1w1m while another show a voltage sensitivity of 92 dB. Many will assume that the higher rated speaker is more sensitive, but the two speakers could in fact have matching sensitivity. I have seen examples of manufacturers who quote a sensitivity rating that is higher than the sensitivity of the bass driver used.

Hoffman's law

The relationship between speaker size, extension and efficiency is referred to in Hoffman's Iron law. You can't have small size, deep bass extension and high efficiency in one box. Where two of these three parameters are chosen, the third becomes a given. Suppose you wanted to achieve 40 Hz extension and 92 dB sensitivity from a speaker. Having chosen two of the three parameters, the size becomes a given. One would need a sensitive driver, most likely a 12" driver with a raw sensitivity of 95 dB before baffle step. A vented alignment would be required, and the box would then need to be as large as required to tune down to 40 Hz. Choosing the right driver will keep the box as small as it can be. If the box volume were restricted to one cubic foot then the choice is restricted to a 10" or 12" sealed low sensitivity driver or an 8" vented alignment with around 85 dB 1w1m.

It's also worth noting that I've discussed anechoic extension figures. Typical specs may show numbers which are based on an in-room guesstimate.

Further investigation

If you would like to investigate this topic further, you might start by downloading driver data sheets, paying attention to the nominal sensitivity, impedance, response plot and any comments on how the measurements were made. Also note that some measurements are taken at different distances. Measurement at 1 ft increases the rating by around 9 dB compared to 1m. Infinite baffle measurements exclude baffle step considerations, and some drivers are measured with no baffle at all which tends to create early roll off.

You might also like to simulate a range of drivers in WinISD, which shows the performance with 1w into half space conditions. This will give a good indication without considering baffle step.

Confused?

This topic can quickly become confusing for many who don't get involved in the technical side of audio. Understand that if you are comparing two speakers with an 8" vented woofer with about the same extension, they will have about the same sensitivity, even if their sensitivity ratings are very different.

No comments:

Post a Comment

All comments are moderated.