The Allison Research Kepex 500 was designed as a companion to the Allison Research Gain-Brain compressor/limiter. Both used the same compact modular format, with the Kepex 500 measuring 178mm high, 38mm wide and 183mm deep including its connector lugs. Rather than being a conventional self-contained rack unit, the module slid into a dedicated metal case using a printed circuit edge connector.
Construction was substantial. The electronics were neatly arranged on a high-quality printed circuit board, with components clearly identified and four internal preset controls labelled according to function. Around 47 transistors were used, many of them associated with the metering circuitry. The layout was considered sufficiently straightforward to make servicing practical when the correct technical documentation was available.
The front panel was simple, although not particularly precise. Three rotary controls handled Range, Decay and Threshold, accompanied by Key and In/Out pushbuttons. Above them was an eight-segment illuminated display showing the amount of gain reduction.
The biggest ergonomic weakness was the calibration. The small knobs provided poor position indication and only a handful of marked settings. Returning the unit to a previously established threshold or range setting with any accuracy was therefore difficult. For a processor that often required considerable experimentation to find the correct operating point, this was a significant inconvenience.
How the Kepex worked
The Kepex 500 operated by reducing the level of material falling below a selected threshold. Rather than simply switching a channel completely on or off, its downward expansion characteristic could progressively reduce lower-level information.
The Range control determined how much expansion was permitted, potentially reducing unwanted material to inaudibility. Threshold established the level below which expansion began, while Decay controlled how quickly attenuation occurred after the wanted signal fell below that threshold.
The specified active expansion range exceeded 60dB. Expansion began at approximately 2:1 and progressively increased towards 4:1 at maximum expansion. Threshold could be adjusted from -35dBm to +20dBm.
Attack was extremely fast, specified at less than 20µs, with no front-panel adjustment. Measurements confirmed that the Kepex responded exceptionally quickly. This was important when processing drums and other percussive instruments because a slow gate could remove or soften the leading transient that gave an instrument its definition.
Release was continuously adjustable, with marked settings of 0.05, 0.2, 1 and 5 seconds. Measurements produced actual times of approximately 0.06, 0.3, 1 and 6.5 seconds respectively. The differences were considered of little practical significance in normal studio operation.
Cleaning up multitrack recordings
Where the Kepex 500 really demonstrated its usefulness was on multitrack material containing microphone spill.
When used during reductions of eight-track pop recordings, it proved capable of substantially reducing low-frequency breakthrough between channels and percussion leaking into microphones positioned around quieter instruments. By setting the Kepex so that predominantly the wanted peaks passed through, an engineer could prevent much of the unwanted background sound from contributing to the final mix.
This could make a noticeably tighter balance possible. It was also useful for preventing the colouration and reverberation associated with unwanted spill from appearing on a processed channel.
The technique was not automatic, however. Finding the point where unwanted material disappeared without the processing itself becoming obvious required care. Too much expansion could produce abrupt changes in level, while too little defeated the purpose of using the device.
Changing performances created another problem. If an instrument altered significantly in level or playing style during a song, the Kepex settings sometimes had to be adjusted during the mix. This was very much a hands-on processor rather than something an engineer could necessarily set at the beginning of a session and forget.
When correctly adjusted, though, the Kepex could rescue material suffering from levels of spill that might otherwise seriously compromise a mix. Its action was considered considerably less appropriate to classical recording, where obvious changes in ambience or natural decay would be much more intrusive.
External keying and creative effects
The Key facility was one of the Kepex 500’s most forward-looking features. Pressing the Key switch disconnected the normal internal control signal and allowed another signal to determine the expansion envelope.
This opened possibilities extending far beyond noise reduction.
An oscillator connected to the key input could impose rhythmic pulsations onto another sound. Very low-frequency waveforms could create pronounced level patterns, while careful adjustment of the oscillator frequency could produce rhythmic effects synchronised with the music.
Speech could also be used as the key signal. The speech itself did not appear at the Kepex output, but its amplitude envelope could control another audio signal. Reversing the arrangement and feeding speech or impulse sounds into the key input while passing white or pink noise through the audio path generated synthetic percussion-like sounds.
Gunfire-style effects were also created in this way, particularly when the processed result was combined with reverberation. Adding tape delay extended the possibilities still further. Even in the early 1970s, the Kepex was therefore being used not simply as a corrective processor but as a sound-design tool.
Metering
The eight-stage display provided effectively instantaneous indication of attenuation, with markings from 0 to 30dB. The display incorporated enough persistence for short-term changes to remain readable and was easily visible even under fairly bright conditions.
Accuracy was slightly less impressive than the claimed ±1.5dB, with measurements suggesting that ±2dB would have been a more realistic figure. Nevertheless, the display provided a useful visual indication of exactly how hard the Kepex was working, particularly when several units were being adjusted across a multitrack mix.
Technical shortcomings
The Kepex 500 was not technically flawless, and contemporary measurements revealed several significant weaknesses.
Its published frequency response was ±1dB from 20Hz to 40kHz. Into a high-impedance load, performance was respectable, remaining within approximately ±0.1dB between 1kHz and 20kHz, although it was already 1.4dB down at 20Hz.
The situation changed dramatically when the unit was required to drive a 600Ω load. A small output coupling capacitor produced substantial low-frequency loss, including approximately 6dB attenuation at 50Hz. A second unit displayed exactly the same behaviour, confirming that this was a design characteristic rather than a fault in the original review sample.
Input impedance presented another problem. Measurements between approximately 2.6kΩ and 2.9kΩ were considered too low for straightforward bridging of conventional 600Ω studio lines.
High-level distortion could also become severe when driving 600Ω. One unit produced around 6 per cent total harmonic distortion at +18dBm, while a second sample measured 16.3 per cent at the same output and began clipping at approximately +16dBm. At more moderate levels performance improved dramatically, with the second sample measuring only 0.25 per cent at +15dBm and 0.11 per cent at 0dBm.
Importantly, these limitations were much less troublesome when the Kepex was connected to a higher impedance input. During practical mix sessions using a 10kΩ load, the severe bass loss was not audible and distortion was considerably less problematic.
The unit’s power requirements were another inconvenience. In addition to a 24V DC supply for the electronics, it required approximately 100V DC solely to operate the neon indicators. Later dynamics processors would make such arrangements look unnecessarily cumbersome.
A distinctive studio tool
The limitations of the Kepex did not prevent it becoming remarkably widespread. By the beginning of the 1980s, its reputation was such that the Kepex name had become almost interchangeable with the term noise gate in professional recording circles.
There is also documented evidence of the Kepex being used to tackle exactly the sort of problem for which it had been designed. During work associated with Van Morrison’s Tupelo Honey, Kepex processing was used in an attempt to reduce leakage present on the recorded tracks. The sensitivity could be adjusted so that unwanted ambience disappeared between wanted material, illustrating how the processor was being applied to major recording projects during the period.
Its character remained appreciated long after more sophisticated gates became available. Roy Thomas Baker later continued to use older Kepex units alongside newer Drawmer gates, specifically commenting on the distinctive behaviour produced by their distortion. That observation is particularly interesting given the measurable distortion of the original hardware. What appeared on the test bench as a technical weakness could, in the right production context, become part of the reason an engineer chose the device.