EAR x 2 joins IIIIIM and I6 M2 !
My recent column pieces on IIIIIM and I6M2 (December 2013 and January 2014) seem to have struck a chord with many of my readers, judging by the stream of emails I have received. It got me thinking as to what else can be done to improve sound system performance.
One of you erstwhile readers suggested that I should have included equalization (or rather mis-equalization) on my list. I have to agree that mis-equalization of sound systems does account for many of the problems I encounter. But who is at fault? The system installer or the user/operator? Once we start on the user list of bad sound causes, then where would I stop…because this ranges from poor microphone technique to using the equipment rack as an ancillary shelf (or additional preparation area in Bars and the like) to tinkering with the signal processing?
I was assuming that the installer is competent (hmmm…maybe I won’t comment any further on that just now), which leaves the user. Just as Ford or Chevrolet vehicles have to pass a battery of safety and roadworthiness tests, their products still seem to have accidents every day! A while ago, while testing a sound system (well, OK, sorting out a real “dog’s dinner” of a mess), the installer said to me, “Well, it’s all right for you, you have all this test gear; no wonder we couldn’t solve the problem.” To which I replied, “I may have all this gear, but what told me what the problem was were my ears; the test gear is just to verify the exact cause of the problem and assist me in determining the best solution to rectify it.”
That’s really the point: So much can be learned about a system from just listening to it. Essentially, this will tell you what the problem is, though perhaps not exactly where it is. For example, if the system sounds distorted, what sort of distortion is it? Is it caused by hard clipping or is the distortion a more of a “fuzzy” sounding aberration? If it’s clipping (i.e., gross distortion), is this level dependent and does it change when different inputs are selected? Very soon, you can tell if the distortion is at the front end of the system or at the amplifier output or loudspeaker end.
A high impedance pair of headphones or a small, high impedance, (or switchable impedance) loudspeaker are all you need to test most circuits (or butt set). Frequency response problems usually are quite audible and, with a little experience, it is possible to get a handle on whereabouts in the spectrum the problem is occurring. Most people can separate sound into lows, mids and highs. Now try dividing these again, e.g., low lows or high lows, lower mid and upper mid, low highs and high highs. Half an hour’s listening to some music and speech via a good system (e.g., HiFi in the office or at home, not a sound system in a large space) and listening as you adjust a 1/3 octave equalizer or set of shelf filters can be really useful and teach you how to identify various frequency bands.
With a little training, you should at least be able to identify six segments of the audio spectrum, as noted earlier. Playing around with speech sibilance in the 4kHz to 8kHz region can also be a useful thing to learn. Another useful skill to learn is that of being able to identify half a dozen pure tones (i.e., sinewave signals). Start with the universal 1kHz signal and then learn to identify 500Hz, 250Hz and then 2kHz and 4kHz. Finish off with 8kHz. Doing this will set you up to be able to far more rapidly identify feedback and know where to start work on reducing it. You don’t even have to own a signal generator to do this; there are plenty of CD test discs out there with these frequency signals on them, or a Google search will soon turn up some online signals.
Thinking about it, I have at least three apps on my iPhone that will generate these signals (and a lot more besides). Faults, such as one of the signal leads in a balanced pair not being connected (single legged circuit), again, are simple problems to audibly identify (and easy to set up to learn what it sounds like). Short circuits on a 70V or 100V distributed loudspeaker line system are also usually easy to hear and get a potential idea of what may be happening.
Hum and noise are straightforward to identify, and normally suggest a grounding problem to be present. (It is very useful to be able to distinguish between “buzz,” “hum” and other background noise and, again, easy to set up a fault condition to experience what this sounds like).
Sound system intelligibility issues usually can be broken down into four or five main problem areas, relating to reverberation, noise, distortion, echo and level. Poor direct to reverberant ratio or excessive reverberant level is usually an easy one to spot, particularly because the reverberant tail after the speech segment finishes will be heard. Noise affecting speech, again, is obvious when you know what to listen for.
Echoes or late sound arrivals are often confused with reverberation. They are something quite different but, again, obvious when you know what to listen for (experiment with a delay line and digital reverberator or editing program to get a sense of this). Often, sound systems can exhibit poor intelligibility just because they are too quiet or too loud; both are important acoustic aspects of intelligibility, and both should be obvious to the experienced installer or audio engineer. There is, of course, EQ, which is where I started.
If you really want to get to grips with “ear training,” as the technique is sometimes called, a number of universities run courses on this (e.g., Tim Ryan’s Course at Webster University or McGill University in Montreal) or attend an AES convention and take in a tutorial from Alex Case or many of the other presenters. There are also books (courses) on critical listening, with accompanying CDs and, these days, there are also on-line materials. Where I still need an impedance meter, audio frequency voltmeter, signal generator and handheld RTA when I am fault finding a sound system, the first thing I do is to listen. Try it:
Go use your ears!