Signal Processing

Powerful computers, DAW software and plug-ins have irrevocably altered the face of the audio landscape. Addicted to the convenience of automation, unlimited takes, performance restoration and infinite manipulation, few continue to splice tape. Yet curiously, in the wake of this revolution, analog processing enjoys a greater popularity than ever before. Analog EQs and compressors offer a multitude of functions, personality and flavors that plug-ins cannot deliver. Modern units are often built from the ‘bones’ of legacy products that were first designed in order solve specific problems, but go much farther by leveraging contemporary components and latter-day research.

EQUALIZATION

Equalization was developed as a solution for the telephone industry, who’s long carrier lines were unable to maintain high quality sound. In 1915, George Ashley Campbell working at AT&T patented the first Electric-Wave filter. By combining high-pass and low-pass filters he created a band-pass filter that could manipulate voices sent over telephone lines. These early equalizers were built directly into the circuits of the telephone transmission and receiving equipment. At the time, there was no forethought of ever building stand-alone equalizers.

In 1925 AT&T founded their research and development department, Bell Telephone Laboratories. That same year, Western Electric conducted the first electrical recording session using microphones to convert acoustic energy into voltage and feeding that voltage to the cutting lathe. Adding a circuit path on the way to the needle opened up the possibility of processing the audio signal. Soon, Bell Labs engineers Joseph P. Maxwell and Henry C. Harrison found a use for equalization in phonograph recordings. By 1926 they had already recognized that equalization would be a great way to counteract low-frequency over-modulation, which lead to overlapping grooves being cut into Western Electric discs. They devised a system where low frequencies below 250 Hz would be attenuated on their way to the recording head. A complimentary low frequency boost would be added to the circuits of a consumer’s player which would restore the low frequencies to their natural level during playback.

This concept is known as pre-emphasis and de-emphasis, and would eventually develop, seeing high frequencies also boosted during recording, and attenuated during playback. This became common in the record industry and before long, engineers were trying the same thing in sound movie theaters. In the 1930’s, John Volkmann, working for RCA, had created an independent equalization device which is recognized as the first operator-variable equalizer. His tool, featuring a set of selectable frequencies with boosts and cuts was designed to improve the sound of playback equipment in various movie theaters.

RCA, Bell Labs, Langevin, and Cinema Engineering would continue to develop standalone equalizers which would be used not only for reproduction in theaters, but throughout production and post production of films. These tools would become commonplace in music recording studios and radio broadcasting facilities as well, with notable contributions like the Pultec EQP-1 passive equalizer coming into the fold as early as 1951. The flavor of its active counterpart the EQP-1A is still one of the most popular program EQ’s today. In 1952 P.J. Baxandall published his paper “Negative-Feedback Tone Control” in which he introduced his circuit providing bass and treble increases and decreases using only potentiometers, and not switches, to shift from boosts to cuts. The smooth musical equalization curves Baxandall created would later be captured by the Dangerous Music BAX EQ.

This idea of different equalizers having their own unique flavors would define research and development in the late sixties and early seventies. In this era, recording consoles were custom-built for studios, each one catering to the desires of house engineers. Because of this, different studios each had their own unique sound. By 1967, Saul Walker introduced the API 550A equalizer, whose bandwidth is inherently altered relative to the amount of signal boosted. This EQ, like others of its time featured a fixed selection of frequencies, and variable boost or cut controls at those frequencies.

At the same time Bob Meushaw was developing an equalizer that was not limited to frequency presets, but instead offered infinitely sweepable frequencies within a certain range. His friend George Massenburg would further this idea, incorporating a fully “parametric” equalizer into a console designed for ITI Studios in Maryland. This design, though somewhat noisy, was the first to feature fully-sweepable frequency, gain, and dedicated bandwidth controls for equalization bands. Massenburg would continue to improve upon this design with his GML equalizers. Now, countless modular equalizers and console EQ’s feature different flavors and option based on these classic designs.

COMPRESSION

Limiting and compression would see a similar journey from humble problem solvers to signature sounds. The very first peak limiters were built in-house by radio broadcasting engineers. Early on, over-modulating of AM signals was prevented simply by manual level-riding. Consequently, overall broadcast volumes had to be extremely low in order to ensure that unpredictable spikes would not exceed the legal broadcast limit. Eventually, the desire to present louder broadcasts inspired the development of safety limiters that could reduce the hottest peaks in a signal.

Engineers at different stations hot-rodded amplifiers to achieve this result, making custom gear. By late 1937, due the popularity of these types of devices, manufacturers began to sell these devices commercially, with Western Electric issuing the 110a limiting amplifier, and RCA following suit with the 96-a. While crude, simple, devices like these were trusted merely as last resorts after traditional level-riding, they helped establish the signal processing industry.

Truly reliable peak limiting did not come until 1947, when the GE BA-5 was introduced. Dubbed a “feed-forward” limiter, the BA-5 design takes a split from the input signal and feeds it into the detector circuit of a variable-gain amplifier (VGA) through which the original signal is passing. The VGA responds to increases in incoming voltage at the detector by attenuating the output of the amplifier. This circuit, common in many modern compressors, allowed radio stations to broadcast at higher levels with the confidence that they would not over-modulate. It became so popular that RCA built off of the design creating a very similar BA-5. One unfortunate side-effect of using these early feed-forward limiters was the fact that they were prone to audio artifacts. The truth is, they would prevent fines but at the cost of audible distortion and thumping effects. RCA’s next model, the BA-6A would be among a new generation of compressors aimed to solve this problem by using updated tube designs.

By the 1950’s the GE-6383 variable-mu tube was introduced and would change the face of audio forever. This one-of-a-kind circuit would become famous for its ability to alter its gain characteristics in response to changes in incoming signal, allowing a remarkably smooth performance relative to other compressors of the time. This tube became the basis of a whole new school of compressors that would become standard in radio broadcasting and record cutting for decades.

In the 1960’s engineers sought out new, innovative solutions to detect and control gain in audio signals. In 1965, Teletronix introduced the revolutionary LA-2A optical compressor. Using electricity from the input signal, an electroluminescent device was illuminated and this light operated a photo-electric resistor. This photo-resistor method continues to be regarded as an incredibly natural and fluid-sounding compression scheme. Shortly after Bill Putnam acquired Teletronix, adding their inventory to his UREI brand, he introduced the 1176 FET limiter, which used transistors instead of tubes or optical circuits for gain control. No stranger to innovation, having designed a wealth of gear for his Chicago-based Universal Recording studio empire, Bill Putnam would oversee the development of many circuits which would reinvent modern compression. From the solid state LA-4, to the low-noise 1176LN, to his original 610 tube recording console, many of his designs are still alive and well in modern studios.

TIME-BASED PROCESSORS

By the late 50’s, and early 60’s, processors creating artificial delays, reverberation and other time-altering effects were making their way into recording studios. Many people are familiar with tape delays, in which sound would be recorded to a tape loop with one head, and played back with another shortly afterwards. Even before that, the first delay processors were being designed by Bell Labs in an attempt to recreate the echo incurred over long-distance telephone lines. Their defunct design, which saw audio vibrating through springs and picked up at the other end, caught the interest of tone-wheel organ manufacturer Laurens Hammond. In the late 1930‘s, he was trying to bring the sound of a cathedral pipe organ to the living rooms where his tone-wheel organs usually resided. Building off of the Bell Labs invention, he introduced the first spring reverb, a giant device, housed in the organ’s tone cabinet.

While spring reverbs became popular in organs and guitar amplifiers their sound failed to make an overwhelming impression on recording engineers. The more common practice in studios was to use echo chambers to add reverberation to a mix. Bill Putnam is known to be the proprietor of this technique, actually taking splits from his console to feed signal to speakers in echoey rooms, which would be mic’ed and returned to his console. Though this concept gained popularity in a good number of studios, but there were plenty that didn’t have the real estate to dedicate to those types of spaces.

A solution came in 1957, when the Germany’s EMT introduced an idea which played on the classic spring reverbs but achieved a much greater frequency response and superior overall sound. Instead of recording sound vibrations through springs, their EMT 140 reverb featured a thin sheet of metal suspended by springs inside of a wooden enclosure. Sound from a speaker would vibrate the metal plate, a microphone would record the vibrations, and a damping pad would allow control of the overall decay time. Plate reverbs were big, heavy, and required a quiet space in which to reside, but were still more practical than echo chambers in the eyes of many.

Like plate reverbs and tape delays, other effects have histories that predate familiar circuit-based versions, extending back to a time when they were purely mechanical processes. Flanging, for example is known to have been introduced by Phil Spector. Attempting to thicken a vocal by playing it from two tape recorders, he pressed his hand against the tape flange on the second machine to slightly offset the positions of the two decks. Due to the irregular pressure applied, the speed of the second machine wavered relative to the first creating a modulated phase shift. Studio tricks like these were common in the experimental recording era of the sixties, but eventually, cool, weird sounds became a regular expectation in mixes. The need for consistent access to these types of effects created a demand for electronic signal processors.

In 1969 Phillips engineers Sangster and Teer presented a solution to electronically delay audio signals by storing and passing the voltage through a series of capacitors. The timing of the voltage passage was electrically clocked at a speed which was operator-variable. This passage of voltage down the line conceptually mirrors a fire-fighting “bucket brigade” passing buckets of water from hand to hand, thus this was the name given to this type of circuit. The sound of the Bucket Brigade Device (BBD) was made famous by devices like the Electro-Harmonix Memory Man delay pedal, the Roland Dimension D stereo widener, and flangers like the MXR Flanger/Doubler and Mu-Tron Flanger. This first wave of purely circuit-based TBP’s were known for their organic character, which resulted from natural signal degradation throughout the BBD chain. In this same era, a new type of delay circuit that would be less colorful, and more accurate was already in the works.

When Dr. Francis Lee at MIT devised the first digital delay circuit in 1969, his plan was to use it for use in heart monitors and possibly even speech-learning tools. He formed Lexicon with engineer Chuck Bagnashi for that reason, but his teaching assistant, Barry Blesser, had the idea of running audio through the circuit. The successful experiment caught the attention of Gotham Audio who saw digital delay as a potential pre-delay circuit for the EMT plate reverbs that they were distributing in America. Gotham licensed the design, and in 1971, introduced the pro audio industry to the Delta T-101, the very first digital signal processor.

Lexicon would stay true to their original course offering a language-learning aide in the Varispeech Model 27Y, designed to correct pitch after a tape was slowed down, in order to study the words spoken. Used by itself, the device became popular in the music world providing digital pitch-shifting, a few years before the famous Eventide H910 harmonizer. Meanwhile, Barry Blesser left his MIT associates to work with EMT and created the sci-fi looking EMT250, the first production-model digital reverb, introduced in 1976. Lexicon would finally define their legacy as makers of some of the world’s finest digital reverbs with their 224 Reverb in 1978.

As more complicated processing algorithms were developed, and digital recording became available, the studio-in-a-computer days dawned in the early nineties. A program called Deck from OSC offered the first multi-track recording software, which was soon licensed to Digidesign, giving birth to Pro Tools. From there, the idea of “plugins” adding functionality to an existing program had already been happening in other places, like photo editing software. That said, when Waves introduced the Q10, the first audio DSP plugin for Pro Tools, the industry was forever changed. Eventually, any popular signal processing hardware device was rivaled by a software version. The whole thing spawned the endless debate: Do we still need hardware? Or can we mix entirely “in the box?”

MID/SIDE PROCESSING

Click here for a full article on Mid/Side Processing

THEN AND NOW

That conversation is still fresh today, and many would agree that whatever sounds the best is the right solution. That said, to one degree or another, most engineers are adopting a hybrid workflow, fusing the benefits of a DAW with the sound of their favorite hardware. Whether flangers, plate reverbs, limiters, analog EQ’s or full-on mixing consoles, the common belief is that passing voltage through electronics creates an organic sound that can never truly be expressed by a predictable mathematical formula. Dangerous Music is fully vested in creating solutions for this modern, hybrid studio.