Wednesday, July 19, 2006
Distortion ain't all harmonic
In the audio world, including the subset of microphones, the term "distortion" is often talked about, written about and mentioned in the context of "something that has gone wrong". Sure, makers of amps, transducers, speakers, converters and whatever want to be able, at least some of the time, to achieve the ideal transfer function, that linear nirvana, that graphically perfect depiction of what goes on inside the "thing".
But it isn't all that simple. Distortion has many forms. Some are readily "measurable" and I put that in quotes because even these can be prone to errors of quantification and interpretation. Other forms of distortion are very subtle, too subtle for the best instruments to identifiably quantify. But the human ear, if it is trained or connected to the highly adaptive brain, hears and knows about many forms of distortion that aren't on the specification sheets of any products or test instruments.
The particular type of distortion that most often gets reported is Harmonic Distortion. According to Fourier's Theory of Trigonometric Series, which among other things states that any arbitrary waveform is comprised of one or more sine waves that are mathematically related, we are able to measure the artifacts that occur when a transducer, for example, is excited by what we presume to be a pure sine wave. It can be a valid way to measure "fidelity". But, there are a lot of things that can get in the way of this measurement and that can even mask the results, like averaging, which can "restore" a sharp square wave, for instance, round it off, and make it look more like a sine wave.
Very precise, calibrated equipment is needed to make good distortion measurements, in the less-than 0.1% range, and all parts of the circuit must be fully characterized and in most cases compensated for their own contribution to distortion.
The typical setup is to use a calibrated distortion analyzer and a signal generator and calibrated loudpeaker. A sine wave is generated and pointed at the device under test, in our case, a microphone. The assumption that the sine wave shape will be perfectly preserved is then put to the test by some comparison method. In the case of the H-P gadget in the picture, the original sine wave is notched out, and the remainder is measured, and called harmonic distortion.
You might be one step ahead here and wonder if that loudspeaker might also contribute to the distortion measured. Yes it can. And also the coupling, the placement, the room, loudness, reflections and standing waves, the sensitivity of the pickup device (a sensitive microphone is more prone to detect, correctly, the ambient distortion than a less sensitive microphone) the stand, the mount, the just about anything that can be moved or that changes the way the mic picks up the sound. All of these can and do contribute to the overall distortion.
But you know that. Mics have to be placed, sometimes quite precisely, for just the right sound. And loudpeakers are very frequently driven to "saturation" (distortion) in the production of music, especially for electric guitar.
Audio distortion and tone may be perceived with instruments, or with your ear-brain setup. If you need numbers, a graph might do. If you need a sound, the ears usually win.
Posted by Bob Crowley at 8:32 AM