There is considerable confusion among microbroadcasters with regard to the an accurate standard for setting modulation deviation maximums on their stations. The purpose of this paper is to describe a method which is highly accurate and will keep you within legal levels.
One of the major problems most microbroadcasters experience with setting peak modulation levels stems from the methods often employed. Let's look at these methods and see where the deficiencies lie:
Adjusting by Ear
The most primative approach used by those first starting out and lacking even simple measuring equipment. A major drawback here is no reference for peak levels. Depending on the type of limiter employed (if even such a station uses one), the results can be wildly skewed depending on the peak to average ratio of the program content. Without a limiter, levels sound very low, so the tendency is to set them higher. So goes it with poorly-designed limiters as well.
Adjusting by VU Meters
This is slightly better, but still prone to errors, as most VU meters will not read true peaks. Most are average responding meters. Again, depending on the peak to average ratio of the program material, the operator could tend toward setting levels too high, simply to satisfy what he/she deems the desired range of meter reading.
A Better Way
When we understand the nature of audio waveforms, we begin to form an interesting picture of how the human ear perceives sound, particularly the intensity of volume. The ear-brain system itself, is a relatively slow mechanism, which relies more on the long-term components of an acoustic wave to interpret loudness, rather than the extremely short spikes or transients that exist in percussive sounds for instance.
Modern broadcast equipment, in order to increase the perceived loudness without actually increasing the peak modulation levels, must actually modify the waveform to some extent. The most popular method is to employ peak clipping. The clipping of short duration peaks results in a net gain of several decibels in loudness, do to the newfound headroom and resulting ability to increase average levels to fill that space. Well-designed broadcast limiters yield a louder program level without overmodulating the channel. When the ameteur microbroadcaster attempts to make his/her station sound as loud as a commercial station, without a peak-modifying limiter, he/she is venturing into dangerous territory, because the peak levels haven't been removed down to the average region of the waveform. The result is frequent, brief overmodulation, which affects adjacent channels in the form of spike noise. This is why it is so necessary to set peak modulation levels by a means that displays the true peak levels.
Using the Oscilloscope
Here's the only way, short of buying a commercial modulation monitor, that provides a reliable measure of peak modulation levels. Using a good-quality oscilloscope with calibrated graticules on the screen and a high-quality FM tuner, you can establish a "brick wall" boundary for your station's modulation.
The typical FM tuner puts out about 2-2.5 volts peak-to-peak when receiving the average FM broadcast. The .5V/div scale on your 'scope's vertical deflection amplifier is probably the one most likely to provide the best view of the signal levels in this case.
First, you need to establish a standard. This will take some time, possibly hours of monitoring different commercial FM stations' audio. What you're after is a maximum peak level to which you'll draw an imaginary line as your boundary. We can safely assume that commercial broadcast stations are going to keep their peaks pretty close to 100% modulation (more like 95-98%) and that these peak deviation values will provide you some references that you should never exceed with your own audio. Do not use a college or non-commercial station as a reference. Such stations often overmodulate and get away with it for months until a surprise FCC inspection results in a NAL. The big 50,0000-watt stations have a lot at stake and tend to play it safe with levels. You will probably notice this after staring at modulation waveforms for various stations over many hours.
You'll also notice short-term averages, and occasional brief deeper spikes in some program material. The points which the short-term spikes reach are your "100%" peaks. When you set your own limiter output levels, monitor your station's audio in this fashion and adjust it downwards until you get a 100% peak every few minutes or so during the loudest program peaks you have available. It's a good idea to max out the studio mixing board levels to way beyond what you'd normally use, just to make sure you're seeing the absolute maximum levels you'll ever put out.
Watch the levels from your station. Tune around and watch the other stations' levels frequently. Become accustomed to the patterns and how they look with speech and different types of music. Adjust your levels so that they never exceed those you observe on the commercial station.
I strongly recommend studying and building the Modulation Monitor circuit in my schematics section. This method is absolutely accurate, when properly calibrated, and will garantee awareness of whether modulation levels are correctly set.
Now It's Not Loud Enough!
Okay, now you've discovered one of the differences between ameteur and commercial broadcast limiters. This is where the fun begins. Your levels are legal, but now you need to get clever to gain the volume. This is where you need to study methods of waveform modification that allow you to increase your perceived loudness. Take a look at the limiter schematics in other sections of this web site. They will give you some ideas and a blueprint from which to begin. At least for now, you will have an understanding of this very important issue, which will help you to prevent interference to adjacent channels and reduce the possibility of a complaint.
Please make every effort to be a responsible broadcaster. The more knowledge you have, the better your capabilities to reach this standard and very important goal.
Authored by your friendly "Peg-legged" Bovine One