Clipping: What is it and what's the big deal?
When a signal has waveform peaks that exceed the electronics' power-supply rails or signal swing, that waveform will get chopped off, or "clipped." This condition can occur for a number of reasons, but the most common is applying too much gain to a signal. Clipping is undesirable, as it is an obvious alteration, or distortion, of the waveform.
The audibility of clipping depends upon the degree. We're all familiar with the sound of severe clipping as experienced with a boom box on weak batteries or the typical car radio turned up too high. The common mechanism there is an amplifier limited to modest battery voltage. Light clipping, on the other hand, is much more difficult to characterize. The audibility of light clipping depends greatly on the signal being clipped. A sine wave, or pure tone, as long as it is below a few kilohertz, will generate audible harmonics, or overtones, discernible as additional frequencies. If a 10 kHz tone is clipped, the first harmonic is 20 kHz and not very audible.
This should not be interpreted to mean that clipping is not important at high frequencies. Real-world music signals typically contain high- and low-frequency signals simultaneously. Often, the high-frequency signal that clips is riding on top of a low frequency. This modulated clipping creates low-frequency intermodulation products that are very audible. The more common example of this is nasty FM radio stations that overcompress. Compression raises the low-level sibilants (s and f sounds) until they clip at later stages. This is aggravated by widespread use of pre/de-emphasis, where high frequencies are boosted and then later cut to reduce noise. While this relies on the fact that high frequencies do not naturally occur at high levels, it's not nice to fool mother nature.
In general, the more complex the sound, the less audible the clipping. Light clipping of transients will in fact increase the apparent high-frequency content and give the sound a little more edge. One obvious exception is the human voice. While the voice is a very complex signal, our hearing is highly attuned to voice frequencies, and we have daily reminders of what undistorted speech sounds like.
OK, what does this mean to me? It depends upon your application. If you're making an archival recording of a flautist (flutes like to put out clean sine waves), you need to record as cleanly as possible. If you're mixing Axel Motif's YMCA rap gig, you might be improving the sound to let it clip a little. As with everything else in life, there are tradeoffs. If you run your signal too far below clipping, you will not get a good signal-to-noise ratio. Most Peavey products are calibrated to fire the clip LEDs a few dB before clipping, so occasional firing of a clip LED means you are hitting the channel hard enough to get a good S/N and not distort the sound. If you're mixing on that other guy's mixer, by the time you get clip LEDs, your signal is toast. Always use your ears to back up your eyes.
To help you in your task of accurately reproducing sound without clipping-induced distortions, we incorporate DDT into our power amplifiers. DDT senses clipping (or current limiting) and reduces the gain to prevent audible distortion. In mixers we provide metering and clip/OL indicators. Recently, competitors' advertising has made a big deal about "headroom." One would be led to believe from the frequency of such statements that they somehow have more headroom than others—not so. Most contemporary mixer designs share similar limitations, namely, the power-supply voltages. The voltages are further defined by the integrated-circuit technology. Most general-purpose ICs are manufactured with a 36V process; this limits the power supply to +/-18V max. The more conservative approach, +/-15V rails, costs a mere 1.5 dB of headroom for increased reliability. There will be small to nonexistent headroom differences between competing mixers. Why the perceived difference?
One dirty trick competitors use is cheating the clip indicator to understate how much a signal is clipping. As light clipping is not very audible, and the ad said they have more headroom, they get away with it. One prominent manufacturer (who shall remain nameless) goes as far as not sampling after the fader gain stage in his input channel. This means you could be clipping as much as 10 dB before the LED lights up. This same manufacturer also doesn't sample at multiple points (we sample at preamp, EQ, and post fader). Any cut in the EQ will allow the preamp to clip before showing up on the clip indicator. (Note: when using PFL to gauge headroom, keep in mind that it won't show the 10 dB of gain in the fader stage.)
Just for fun, if you know someone who foolishly bought a non-Peavey mixer, get them to try this test. While talking into a microphone plugged into an input with the fader up at maximum, slowly increase the mic preamp gain while listening to the output. If your friend is unlucky enough to have the wrong mixer, you will hear severe clipping distortion long before the clip light indicates. This falls under the category of fooling all of the people some of the time, and some of the people all of the time.
It would be easier (and cheaper) for us to remove the multiple-point-sampling clip circuitry, but it isn't the right thing to do. We have cultivated a long-term relationship with our customer base by giving them honest, accurate data and will continue to do so. We believe they can't fool all of the people, all of the time. The marketplace will figure out who's telling the truth.