Getting The Lowdown

Whatever. To get reasonably loud, low frequency levels, you do need the speaker to have reasonable size. While it’s possible to make a small speaker have extended low-frequency response, that capability comes at the expense of ability to play very loudly.

Another engineering approach to the low-frequency/loudness/size problem is to sacrifice smoothness of low-frequency response in favor of increased mid-bass output and a sharper, somewhat higher, rolloff. This is accomplished by adding a port to the speaker enclosure. The port is tuned to provide resonant output at a frequency below the rolloff point of the woofer itself, so an additional low frequency peak is added to the response curve. This increases the speaker’s efficiency (output per watt of power), so that you get greater output from a smaller box. The downsides are that you lose more of the audio window and your accuracy of reproduction within the window may be reduced.

So, to reproduce really low frequencies (from, say, 30 to 80 Hz.) with reasonably flat response, you need substantial speaker size. You also need to consider how your speakers interact with your room. More about that later.

The Hearing of Badnesses

There is another, related, aspect to our listening needs during the music production process. We need to be able to detect and solve problems that come up during that process. We need speakers that reveal flaws in the recording.

Such speakers have pretty much the same design criteria as the “listening-back” reference speakers. They need to be accurate, which is to say they need to be transparent and neutral, to add as little as possible to the recorded signal, which in turn suggests minimal harmonic or intermodulation distortion, as well as “accuracy in time,” and “accuracy in stereo,” so that such speakers will not significantly change the time relationships of complex sounds, and that stereo pairs will match precisely. When these criteria have been met, many issues pertaining to stereo imaging, reverberant detail, microphone placement problems, etc., begin to be revealed, along with distortions of all sorts that have occurred elsewhere in the process.

Listening Ahead

For music production, we have a third, rather specialized, need. We need to be able to listen ahead in time, to predict how our music will sound in the future, over other playback systems. This is central to our production process, and it is a special production concern not shared by other listeners. Instead of concerning ourselves with how well we’ve recorded a musical event, we need to predict “how well other people in other places in the future listening over other loudspeakers will think we’ve recorded a musical event!”

The Range Rule suggests that for us to do this we should anticipate the conditions under which those people will listen, by trying to listen in the same ways that they will listen. This means, among other things, listening in crummy conditions to crummy speakers. Hence we listen to Auratones, we listen while riding around in cars, we listen in mono, we listen at low levels, etc. See my article on end-users in the May, 1994, issue of Recording.

For such an approach, we need loudspeakers that are “average” and/or “below average” and or “typical.” Such speakers are usually cheap. They are usually small. They don’t have flat frequency response and they don’t have to play loud.