Right now, over-the-air broadcast radio is fighting for every single listener. Beyond the competing stations in the market, there are other choices looking to cut in on radio’s territory, including streaming to personal devices and connected dashboards in cars.  Holding the attention of the audience requires more than just exciting formatics – you’ve also got to address the subtleties of what may cause a listener to remain locked in, or to tune away.

Aside from the station format and how it’s executed, the other key element is how the station sounds technically. Here’s where audio processing plays a huge role. The old “crank it up and rip the knob off” positioner might be a classic slogan, but creating a great sounding station is more than just revving up the processing with the goal of being the loudest big boy on the dial. Yes, it is possible to sound competitively loud AND possess good audio quality with modern processors, but there’s even more to it. There is an annoyance factor that can occur with processing, and this essay will detail what that is, and provide a new solution to counter it.

Annoyance Factors

In the technical domain of signal processing, there are a number of factors that govern how the resulting audio sounds. The two key items are rooted in forms of distortion.

Harmonic distortion (THD) is typically the sound one hears where the audio appears “broken”. In an audio processor, THD is affected by the level of clipping which is employed for competitive loudness. Too much clipping may help burn a hole in the dial, but the quality will usually suffer. (Yes, there are distortion controlling mechanisms in many processors, but as with anything, there IS a limit.)

Intermodulation distortion (IMD) is created when a dominant portion of the audio spectrum affects another portion of the spectrum through modulation. A simple example of this is the basic audio compressor. When set to very fast attack and release rates, the compressor will modulate the resulting audio to the point where it sounds busy, or as some may say, “bizzy.” This occurs because the control signal of the compressor is operating at or near the frequency of the audio spectrum, and it is adjusting the audio so fast that the listener’s ear perceives it as an annoyance.

"Constipated audio"

This annoyance can occur if the time constants in compressors or limiters are set incorrectly,  and the result can give the impression of the audio “breathing” due to heavy depth of compression coupled with very fast time constants. I know there are some folks in our industry who actually like or prefer this signature, but research has shown that this affected sound turns off most listeners, especially women! Yes, processing set in this manner will generate extremely dense audio. Yet, when the brain perceives a dense signal, it reacts negatively and subtly begins to tune out. In my own listening, I refer to this as “constipated audio.”

Clipper Systems

While this discussion thus far has focused on the dynamic aspects of processing, the critical mass occurs in the final section, which is usually referred to as the final limiter, or clipper. Digital signal processing techniques have improved the final limiting function manyfold over their analog counterparts, but there have still been limitations to what these systems are capable of…until now!

As mentioned previously, we have mechanisms that can monitor, control, suppress, and filter THD. These functions help enable competitive loudness and minimize the distortion side effects. While very much a benefit, they do nothing to reduce any annoyance due to clipper-induced IMD.

The challenge presented here occurs when clipper action in various parts of the spectrum pushes the rest of the audio in and out of the clipper system, whether it needs to be clipped (truncated) or not. Since all of this activity is happening at the rate of the audio signal, meaning there is an instant attack and an instant release function, the generated IMD can build up quite significantly. In our broadcast systems, there is usually a pre-emphasis network employed, which boosts the high-frequency range. This further pushes the presence and treble frequencies into heavy limiting at almost all times. The resulting audio can be very dense and annoying to the ear.

Enter the FFT

Over the years, processing algorithms have evolved using the Fast Fourier Transform (FFT for short). We refer to this as managing a signal in the frequency domain, as compared to the time domain. Time domain signals are continuous waves, as would be observed on an oscilloscope, where frequency domain signals are independent of time, and managed within frequency bins. Through creative design, it has been possible to further reduce THD in the broadcast clipper system. Yet, the problem of IMD annoyance remains, and with the FFT, it can even get worse! How so? Read on.

One benefit the FFT offers is the ability to manage the frequency spectrum in a very finite manner. As mentioned, the FFT is a frequency domain function that utilizes many “bins” to store and manage audio information. These bins are set in the FFT function as a power of two. So, for example, we can create an FFT function that has 1024 bins, which is 2 to the power of 10. Given the audio spectrum of 15kHz for FM stereo, and with a sample rate of 192 kHz, we will divide the sample rate by 1024. In this example, 192 kHz divided by 1024 results in each bin being 187.5 Hz wide. As you can see, this is very precise and offers a lot of flexibility in what can be managed and adjusted over the range of the entire audio spectrum. It also provides exceptional filtering capabilities. So far, so good…but not so fast!

As with most aspects of life, there is no free lunch. While the FFT provides great filtering and THD management, it suffers with regard to how transparent audio transients are handled. For example, let’s take a well-recorded audio track from Steely Dan, the song ”Babylon Sisters.” It opens with four very distinct tom-tom drum hits, which contain great transients. When passed through the FFT clipper system, the drum transients get lost and/or muffled instead of remaining transparent. This occurs due to the nature of the FFT transform, and (without going into extremely deep technical jargon to explain why) it is not possible to preserve the sharp transient response. This is known as “transient smearing,” and it affects IMD perception as well. While the FFT provides very low THD, the level of IMD is still in the subtly annoying range.

A Fresh Approach

Although extensive research has been conducted on FFT methods, it is not possible to eliminate the loss of transients (transient distortion, for short), or the aforementioned IMD, from the FFT-based clipper system. So after considering the main benefit of the FFT, which is the precise filtering gained by using the “bins,” we’ve created a new approach devised in the time domain. This replicates the benefits of FFT filtering and employs new algorithmic elements to suppress clipper-induced IMD. An entirely new system evolved which provides excellent THD management, IMD suppression, AND the preservation of transient transparency!

The result is an open, clean, and competitively loud signature without transient distortion. In simple terms, the audio is easier to listen to and very natural-sounding. This invites longer listening, as the subtle audio annoyances that prompt listeners to tune away have been removed.

The recently-introduced Omnia.XII FM/HD audio processor employs this new clipping system, along with brand-new algorithms in the dynamics (compressor and limiter) sections, which monitor and reduce IMD as well. The end result is a fresh approach to audio processing that delivers high quality, competitive loudness, without the sonic annoyances that other systems create, and significantly raises the bar for improving listener retention and increasing TSL (time spent listening). You’ll definitely hear the difference, and your listeners will too.

Frank Foti
June, 2026

Ready to hear Omnia XII for yourself? Just contact your Omnia dealer.