Almighty Subwoofers

By Stig Erik Tangen
01 Jan 1997

Acoustic measurements

Can measurements give us any clue of what a speaker sounds like? The most flawed systems can easily be selected out by measuring key parameters like frequency response and distortion. However, no speaker sound the same and no speaker measure the same. The speaker with perfect technical data does not exist, and probably never will, but a speaker that measure good usually sounds good too.

I performed measurements of frequency response and distortion on the finished subwoofers to ensure that the theory behind the system was not too far from reality. The measurements were done with a Techron TEF-20 analyzer using a Brüel & Kjær 4007 microphone, both kindly provided by an acoustic consultant, Lars Tofastrud, that happens to be a good friend of mine (VERY convenient...)

Measuring a speaker's low frequency response in a normal room is in general totally impossible! This is because the room's acoustics totally dominate the measurement. The room's resonant modes will disturb even a near-field measurement. A low-frequency measurement should be executed at a large open plane outdoors. Dragging 160 kg of speaker enclosure out in the corn field in december did not tempt me too much, so I decided to try indoors. The TEF's TDS (Time Delay Spectrometry) measurement technique allows you to isolate the speaker-response from the room's response, but on the downside this reduces the resolution at low frequencies, so a suitable compromize is necessary. About 100 ms of room decay (after the direct signal from the speaker) is included in the first three measurements shown (100 ms Time Resolution in the TEF TDS parameters). Measurements were done with the filter and equalizer included.

My room has its major resonances (both standing wave modes and structural resonances) below 100 Hz placed at 19, 23, 27, 34, 46, 65 and 80 Hz (I've measured this with the TEF). These can be spotted on the measurements below, especially the 19, 23 and 65 Hz modes stand out. No smoothing is applied on the response curves (no fakes!).

Near-field on the driver:

Near-field on the port (inside the port actually):

Near-field midways between the port and driver. The dip at 200 Hz is probably caused by a floor reflection.

The next is a measurement I did at my listening position (2.2 meters) at the same amplifier output as the above measurements. I used 500 ms time resolution on the TEF (500 ms of room decay is included in the frequency response) to capture the most significant room response. This looks extremely bad, but is in fact much better than obtainable in an average room. My room is quite heavily trimmed acoustically and its bass sounds superior to most other listening rooms. Even so, it still has a very slow decay below 50 Hz (about 1s RT60 at 23 Hz!) This measurement is indeed coloured by the slow decay.

Below is the first 600 ms of room decay that causes the rugged response shown above, measured in one rear corner of the room. Note that many strange things happens to the room modes after 200-250 ms. It's a wonder that music can sound natural in such an acoustic environment, when the room reverb changes pitch during decay. (I could write hundreds of pages on this.....)

Distortion measurements are a little easier, as the room's response will have less effect on the result. Even so, I measured in the near field (about 10 cm from the driver cone). The amp output is at a level that produces 96 dB SPL at 1 m (the amp output voltage varies a great deal with frequency because of the crossover and equalizer).

This measurement really shows what happens below the tuning frequency in a vented box! As you can see, the THD is almost 25% below the tuning frequency, but less than 1,5% at any frequency above. The distortion components are mainly of 3rd and 4th order.

About the
author

Print this
article

Email this
article