Once a sound has been transduced (transformed into electrical energy for the purpose of recording or transmission), the characteristics of that sound can be electronically manipulated. These characteristics include pitch, loudness, duration, and timbre. Thus, the term audio processing refers to the art and science of making changes to an audio signal to improve or enhance the original sound or to create an entirely new sound based on the original. There are many technical and creative reasons for audio processing in radio.

     In the early days of radio, if the audio going into the transmitter was too loud, the transmitter could be damaged. Even today, because of the potential for interference to other stations caused by overmodulation, the Federal Communications Commission (FCC) has strict rules about modulation limits. Audio processors continuously maintain a station's compliance with these rules.

     There are also many creative reasons to process audio. Consider the following examples: a commercial producer needs to transform the talent's voice into that of a space alien. In another commercial, the voices sound a little muffled; rerecording the spot through an equalizer to increase the midrange can make the voices sound louder. A third spot as recorded runs too long; it can be shortened by redubbing it through a digital signal processor using the time compression function. These are examples of problems that can easily be solved with the right audio processing in a production studio. Radio stations want their sound to be clean and crisp, bright, and distinctive. Rock-formatted stations targeting teens and young adults usually want to sound loud, regardless of the particular song being played. These are examples of the types of needs addressed by the processing equipment in the audio chain before the signal goes to the transmitter. A description of the basic characteristics of sound identifies the component parts that are manipulated during audio processing.

Characteristics of Sound

     Sound is created when an object vibrates, setting into motion nearby air molecules. This motion continues as nearby air molecules are set into motion and the sound travels. This vibration can be measured and diagrammed to show the sound's waveform. The characteristics of a sound include its pitch (frequency), loudness (amplitude), tonal qualities (timbre), duration (sound envelope), and phase. A sound is described as high or low in pitch; its frequency is measured in cycles per second or hertz. Humans can hear frequencies between 20 and 20,000 hertz but usually lose the ability to hear higher frequencies as they age.

     The subjective measurement of a sound's loudness is measured in decibels (dB), a relative impression. The softest sound possible to hear is measured at 0 dB; 120 dB is at the human threshold of pain. The range of difference between the softest and loudest sounds made by an object is called its dynamic range and is also measured in decibels. A live orchestra playing Tchaikovsky's 1812 Overture complete with cannon fire will create a dynamic range well over 100 dB. The amplitude, or height, of a sound's waveform provides an electrical measure (and visual representation) of a sound's loudness. Timbre is the tonal quality of sound; each sound is made up of fundamental and harmonic tones producing complex waveforms when measured. A clarinet and flute sound different playing the same note because the timbre of the sound produced by each instrument is different. Timbre is the reason two voices in the same frequency range sound different. The sound envelope refers to the characteristics of the sound relating to its duration. The component parts of the sound envelope are the attack, decay, sustain, and release. Acoustical phase refers to the time relationship between two sounds. To say that two sounds are in phase means that the intervals of their waveforms coincide. These waves reinforce each other, and the amplitude increases. When sound waves are out of phase, the waves cancel each other out, resulting in decreased overall amplitude.

     Individually or in combination, the frequency, amplitude, timbre,·sound envelope, and phase of the audio used in radio can be manipulated for technical and creative reasons. The characteristics of the audio created for radio typically need adjustment and enhancement for creative reasons or to prepare the audio for more efficient transmission.

Processors Manipulate Audio Characteristics

     Equipment used to process audio can generally be classified using the characteristics of sound described above. There are four general categories of audio processing: frequency, amplitude, time, and noise. Some processors work on just one of these characteristics; others combine multiple functions with a combination of factory preset and user-adjustable parameters. Some processors are circuits included in other electronic equipment, such as audio consoles, recorders, or microphones. Processing can also be included in the software written for a computer-based device such as a digital audio workstation.

     An equalizer is a frequency processor; the level of specific frequencies can be increased or decreased. A filter is a specific type of equalizer and can be· used to eliminate or pass through specific narrow ranges of frequencies. Low-pass, band-pass, and notch filters serve specific needs. Studio microphones often contain a processing circuit in the form of a roll-off filter. When engaged, it eliminates, or "rolls off," the bass frequencies picked up by the microphone.

     Amplitude processors manipulate the dynamic range of the input audio. Three examples of amplitude processors are compressors, limiters, and expanders. A compressor evens out extreme variations in audio levels, making the quiet sections louder and the loud sections softer. A limiter is often used in conjunction with a compressor, prohibiting the loudness of an input signal from going over a predetermined level. An expander performs the opposite function of a compressor and is often used to reduce ambient noise from open microphones. Most on-air audio processing uses these types of processors to refine the audio being sent to the transmitter. Recorders often have limiter or automatic gain control circuits installed to process the input audio as it is being recorded.

     A time processor manipulates the time relationships of audio signals, manipulating the time interval between a sound and its repetition. Reverberation, delay, and time compression units are examples of processors that manipulate time. Telephone talk shows depend on delay units to create a time delay to keep offensive material off the air. Commercial producers use time compression and expansion processing to meet exacting timing requirements.

     Dolby and dbx noise reduction processing are methods of reducing tape noise present on analog recordings. The Dolby and dbx systems are examples of double-ended systems: a tape encoded with noise reduction must be decoded during playback. These types of processing become less important with the shift to digital audio.  

     Until the 1990s most processing was done using analog audio. Individual analog processors, each handling one aspect of the overall processing needs, filled the equipment racks in production and transmitter rooms. Equalizers, reverb units, compressors, limiters, and expanders all had their role. Digital processors were introduced during the 1990s. These processors converted analog audio to a digital format, processed it, and then converted the audio back to the analog form. Most processing today has moved to the digital domain. These digital signal processors allow for manipulation of multiple parameters and almost limitless fine adjustments to achieve the perfect effect. Modern on-air processors combine several different processing functions into one unit.

Audio Processing in the Audio Chain

     Virtually every radio station on the air today uses some type of processing in the audio chain as the program output is sent to the transmitter. The technical reasons for processing the program audio feed date to the earliest days of radio. Engineers needed a way to keep extremely loud sounds from damaging the transmitter. The first audio processing in radio was simple dynamic range control done manually by an engineer "riding gain." The operator adjusted the level of the microphones, raising the gain for the softest sounds and lowering it during the loudest parts. During live broadcasts of classical music, the engineer was able to anticipate needed adjustments by following along on the musical score. Soon, basic electronic processors replaced manual gain riding.

     Early processing in the audio chain consisted of tube automatic gain control amplifiers and peak limiters. The primary purpose of these processors was to prevent overmodulation, a critical technical issue with an amplitude-modulated signal. Operators still needed to skillfully ride gain on the program audio, because uneven audio fed to these early processors would cause artifacts, such as pumping, noise buildup, thumping, and distortion of the sound. Early processor names included the General Electric Unilevel series, the Gates Sta­ Level and Level Devil, and Langevin ProGar.

     Broadcast engineers generally consider the introduction of the Audimax by Columbia Broadcasting System (CBS) Laboratories to be the birth of modern radio audio processing. The Audimax, introduced by CBS in the late 1950s, was a gated wide-band compressor that successfully eliminated the noise problems of earlier compressors. The Audimax was used in tandem with the CBS Volumax, a clipper preceded by a limiter with a moderate attack time. In 1967 CBS introduced a solid­ state Audimax and the FM version of the Volumax, which included a wide-band limiter and a high-frequency filter to control overload due to FM's pre emphasis curve.

     The reign of the Audimax was challenged in the early 1970s with the introduction of the Discriminate Audio Processor by Dorrough Electronics. This broadcast compressor/limiter divided the audio spectrum into three bands with gentle crossover slopes, compressing each band separately. Broadcast engineers began to make their own modifications to some of the internal adjustments, adjusting for specific program content and personal preference.

     In 1975 Orban Associates introduced the Optimod-FM 8000, which combined compressor, limiter, high-frequency limiter, clipper, 15-kilohertz low-pass filters, and stereo multiplex encoder into one processor. This unit allowed for higher average modulation without interference to the 19-kilohertz stereo pilot signal. The Optimod-FM 8000 was replaced by what soon became the industry standard, the Optimod-FM 8100. A digital version, the Optimod-FM 8200, was introduced in 1992. The Optimod-AM was introduced in 1977.

     The development of these processors was driven by the need for a reliable method of maintaining compliance with the FCC transmission and interference rules while allowing for creative use and adjustment of processing for competitive advantage. Along with maintaining compliance with regulatory constraints on modulation, interference, and frequency response, engineers and programmers are always looking for ways to make their stations sound better than and different from the others. Some stations have taken creative processing to extremes. During the 1960s WABC in New York was well known for the reverb used on disc jockey voices during music programs.

     A station programming classical music has processing needs different from those of an urban format station. Preserving the dynamic range of an orchestral work is critical, whereas maximizing the bass frequency and loudness enhances the music aired on the urban station. Today's processors allow for this kind of flexibility in adjustment based on format and on specific goals for the sound of the station. Audio processing plays an important role in radio stations' competition for listeners. Stations targeted toward teens and young adults want to sound louder, brighter, and more noticeable than their competitors. This is where audio processing becomes something of an art. Programmers and engineers cooperate to adjust processing to attract and maintain listeners. This is a subjective process that involves trial-and-error adjustments and critical listening by station management. There is a fine line between compressing audio to boost overall loudness and creating listener fatigue. Low time-spent-listening numbers in the ratings may not be the fault of poor programming as much as of overprocessed audio.

Audio Processing in the Studios

     Much of the audio sent to the on-air processor has already been processed, perhaps as it was originally recorded, dubbed in production, or mixed with other sources in the air studio to create the program output.

     One of the most common forms of audio processing in the studio is equalization (EQ), which is the increase or decrease of the level of specific frequencies within the frequency spectrum of the audio being created. Many audio consoles, especially those used in the production studio, have equalization controls on each channel to allow for adjustment of the EQ of each individual audio source. At a minimum, there are controls for low-, medium-, and high-frequency ranges, but many consoles divide the frequency spectrum into more parts. The EQ controls can be used for various creative and technical purposes. Examples include matching the frequency response of different microphones so they sound the same, creating a telephone effect by decreasing the low and high frequencies of the audio from a studio microphone, adding presence to the voices in a commercial by boosting the midrange, or eliminating hum on a remote line by decreasing the low end. Equalization can also be done through an outboard equalizer; the source or console output can be routed to the equalizer for processing. These units usually divide the frequency spectrum into intervals of one-third or one-half of an octave. Each band has a slider to increase or decrease the amount of EQ on that band. Filters, a specific type of equalizer, can be used to eliminate specific narrow ranges of frequencies. Low-pass, band-pass, and notch filters are usually used to eliminate technical problems with the audio or to keep unwanted audio frequencies from getting to the transmitter.

     A well-equipped production studio has a number of processing options available to producers. Until the development of digital signal processors, every effect came from a separate unit. Although many of these single-function processors are still in use and are still manufactured, digital multiple-func­ tion processors are the norm today. These are generally less expensive than the on-air multifunction processors, and a number of manufacturers provide many different models and options in their studio processor lines. Most units offer a number of factory preset effects with user-adjustable parame­ ters. These units also allow users to create and store their own combinations of effects. The Eventide Ultra-Harmonizer, for example, provides _pitch changing, time compression and expansion, delay, reverb, flanging, and sound effects as part of its inventory. The major advantage of these multifunction units is their ability to combine effects. For example, pitch change can be combined with chorusing and reverb. Flanging can be combined with stereo panning. Given the opportunity for user-created presets and parameter adjustments, the possi­ bilities are almost limitless.

     These same types of digital effects are also integrated in the software of digital audio workstations and editors. Audio pro­ cessing can be added after a recording is made on a multitrack editor. The complex waveform of each track can be processed using the same type of multiple-effects options described above. An announcer can be made to sound like a group of elves through the addition of chorusing, pitch change, and reverb; each track can be processed independently. Because the changes are not made to the original sound files, any of the modifications can be easily undone and the original audio remodified.

     Microphones in the production and air studios often receive special, full-time processing. An analog or digital microphone processor typically provides compression, limiting, de-essing, equalization, noise reduction, and processing functions designed specifically to enhance vocal characteristics.