elysia alpha compressor 2-Channel Discrete Class-A Rackmount Analog Compressor

Manufacturer's Description

The alpha compressor is a modern classic and becomes, since the release in 2008 a mighty tool for professional dynamics processing. Featuring innovative functions and providing absolutely uncompromising sound quality, it sets new standards for enhancing any kind of audio material. It offers M/S processing, parallel compression, limiters, as well as sidechain and audio filters, just to name a few.

Every single of its many aspects – sound, features, circuitry, components, design, materials, manufacturing – has been thought over and over until there was just nothing left to improve. Basically, every single part of this fine compressor is custom, and most of these are made according to our own designs and specs.

Only the best components the market has to offer are good enough for the alpha, and it shows. Each alpha compressor is manually built to order and tested by ourselves, here at our headquarters in Germany.

The alpha’s benefits sum up in an audio quality that is beyond all doubt. Even when extreme settings are used, the sound always stays clean and powerful. The alpha achieves perfect processing results for all different styles of music and assures the ‘this is it’ feeling of a finished song. Its enhanced functions exceed the potential of normal compressors by far and give you unimagined options for mastering and mixing.

The circuitry is a completely new development which provides an amount of openness and clarity which can hardly be found a second time. Every part of the alpha’s signal processing is based on single transistors running in constant class-A mode, and even the sidechain and power supply use discrete circuits. All components are of the highest grade available and exclusively chosen according to their sonic excellence.

M/S technology is commonly known as a variant of stereo microphoning. This technique uses a microphone with cardioid pattern for the middle signal (M) and another one with bi-directional pattern with an offset of 90° for the side signal (S). The main advantage of this technology is its mono compatibility. FM radio stations use M/S technology for transmitting stereo signals exactly for this reason.

To create M/S signals, the left and right channel of the stereo sum are added to generate the mid (M), whereas the side (S) is created by subtracting the right from the left channel:

M = L+R
S = L-R

To decode an M/S signal back into stereo again, M is added to S for the left channel and S is subtracted from M for the right channel:

L = M+S
R = M-S

The integration of an M/S encoder and decoder into a compressor generates new potentials that classic linked stereo compressors can hardly offer. One of the main advantages is the possibility to process the middle and side signals separately. This way you can make the center more compact without corrupting the original stereo spectrum, for example.

Of course it is also possible to enhance the presence of the side signals in an already finished mix. The stereo width can be influenced fast and effectively, too, and it is also possible to compress specific parts of a mix that could not be selected in a stereo mix as precisely as it is possible in M/S mode.

This function makes it possible to switch the feed of the sidechain alternatively behind (feedback) or in front (feed forward) of the actual compressor section.

It has an enormous effect on the character of the compressor: While processing in feedback mode is smooth and even, switching into feed forward mode will result in a clearly stronger and harder kind of compression.

From the technical point of view this function mainly influences the characteristic curve of the ratio value.

In feedback mode it goes up to a moderate ratio of 1:2.5. In contrast, the feed forward mode provides much higher values which also allow limiter settings and even negative ratios (i.e. loud signals will be reduced even more).

Small changes in dynamics can generate a high amount of gain reduction.
The values of the attack controller also change noticeably in feed forward mode, as they are almost twice as high as the values shown on the scale.

Attack and release are very crucial factors for the operations of a compressor. Choosing the right time settings is very important, but depending on the dynamic progress of the source material this can be a difficult task – no matter if single tracks or complete mixes are processed.

If a very short attack time is chosen, the compressor is able to catch the short peaks, but on the other hand the sustaining signal will also be processed, which might result in audible distortion. Longer settings reduce distortion significantly, but then the compressor is too slow for catching fast impulses.

This is where the Auto Fast function comes into play. For example, if you set the attack to 80 ms and then engage the Auto Fast mode, the attack time will be shortened automatically on fast and loud signal impulses. The compressor reduces the signal quickly and prevents it from slipping through.

Then the attack time directly and automatically returns to its original setting. In Auto Fast mode the compressor can be very fast, but only when it is really needed. This function influences the attack parameter on short and loud impulses only; in all other cases the original setting of the controller has priority.

The separate Auto Fast for the release controller behaves in a similar way. The release also often forces the user to accept compromises when searching for the right setting. If it is set it too fast, distortion will occur, if it is too slow, drive and loudness are lost. In Auto Fast mode the compressor adapts to the currently right setting automatically.

This filter specialized in changing the overall sonic character of a track in fine nuances. It features two controllers per channel and is capable of flexibly producing convincing results in no time at all. Whenever a classic shelving filter would be too limited and a fully parametric filter would be too much, the Niveau Filter is the efficient and elegant solution.

Its main task is changing the proportions between high and low frequencies. It works like a pair of scales: Dependent on the gain setting around a selectable center frequency, the high frequencies are boosted up to +3 dB while the low frequencies are simultaneously attenuated by -5 dB maximum. Turning the gain controller into the other direction will cut the treble and boost the bass instead.

The filter type used for this application is an all pass variant with a flat frequency response that changes its phase according to the setting of the frequency controller. If the signal that went through the all pass is mixed to the original signal, everything in phase will be boosted while out-of-phase signals will be reduced.

Because the filtered signal is mixed to the original, the genuine structure and impulse-response remains almost completely intact. None of the amplifiers is shortened in its frequency response, resulting in an open and dynamic sound. Boosting and cutting the selected frequency-areas at the same time makes it much easier to influence the character of a track (‘bright’ vs. ‘dark’) than with standard equalizers.

The sidechain filter allows frequency-dependent shaping of the compression process by giving specific frequency areas a stronger or weaker influence on the detection circuit.

If the SC gain controller is set to HP (High Pass), the filter will act like a 6 dB high pass and the reaction of the compressor on bass frequencies decreases. The setting LP (Low Pass) turns the filter into a 6 dB low pass and the compressor reacts primarily on low frequencies.

The combination of the sidechain filters, M/S matrix and different attack settings enables you to make very selective changes and – depending on the source material – even allows to process single instruments or voices in finished mixes.
Parallel Compression

Parallel compression, also known as ‘New York‘ compression, is a technique based on mixing a dry signal with a heavily compressed identical signal.

It is thought to maintain the subtleties of a performance while stabilizing the dynamics.

The mix controller of the alpha compressor makes it possible to cross-fade between the unprocessed and the compressed and filtered signals. This allows parallel compression right in the box and supersedes additional routings in favor of a better signal quality.

Now you can use even extreme compression settings without killing a track by winning the loudness war. By mixing just a part of the compressed signal to the original, the major portion of the initial dynamic structure remains intact.

Let’s have a look at the control logic of this interesting feature: If only the compressed button is pushed, you will hear the compressed signal only. Otherwise, if only the direct button is pushed, only the unprocessed signal routed from before the compressor section will be heard. If both buttons are activated, the mix controller will become active, and if none of the buttons is pushed, the channel will be muted.

In practice this lets you switch between unprocessed, compressed or mixed signals very fast without having to change the position of the mix controller.

The left and right channels (stereo mode) or the middle and side channels respectively (M/S mode) can be listened to separately – again you have the choice between the original, the compressed or the mixed signal. To mute the other channel, just deactivate its associated direct and compressed buttons.

One of the biggest problems of digital audio technology is setting the right recording level on AD converters. The main challenge is that there is no level reserve beyond the maximum of 0 dBFS which could catch short peaks. In the moment a signal is digitally overdriven it is damaged irrevocably, because the original structure can hardly be reconstructed later.

Depending on the source signal, this type of peaks will produce audible distortion, as the signal is cut off and does not correspond to the original source anymore. In addition, this kind of clipping produces lots of new harmonics that do not always fit into the desired musical context.

The Soft Clip limiter has been developed to solve these specific problems. It is specialized in catching short and transient-like signals reliably. The technical principle is different from a classic ‘brickwall’ design which completely forbids further level increases beyond a certain threshold.

Instead, it is working similar to an analog tape machine driving loud impulses into saturation, acting like a ‘natural’ limiter. The Soft Clip limiter is based on discrete transistors, and just as with tape, its characteristic saturation curve results in rounding peaks instead of cutting them off. Especially when the source material contains variable peak values, the Soft Clip limiter comes in very handy.

Each of the two channels features an additional transformer that can be switched into the signal chain after the mix stage. In principle, these are classic output transformers,
but here they are not used for balancing and galvanic isolation, but as an additional means of sound shaping.

If you like to add that certain amount of ‚iron‘ to your sound, just push the transformer button and there you are. Because of the mastering approach of the alpha compressor, this feature is much more of a subtle audio shaping feature than a glaring sound effect.

In stereo mode, the buttons for both channels should be activated or deactivated at the same time. Depending on the source material and personal taste, you can, however, add the transformer sound to single channels when working in M/S mode.

The gain reduction meter is a very important visual tool for evaluating the operation of the compressor in addition to what your ears tell you. A lot of devices make use of sometimes more, sometimes less precise VU meters. Because of the inertia of the needle these meters are only useful with moderate time parameters.

Another popular form of meter is the LED chain. Unfortunately it has a disadvantage, too: When the standard driver units are used, the change between two values happens abruptly. A single LED in the chain can therefore only show an imprecise value in a defined interval. Hectic flicker indicates that the actual value must be somewhere in between.

The alpha compressor solves these problems by using an analog dynamic variant that combines the benefits of both VU meters and LED chains. This meter is based on LEDs, too, but a special circuit design makes it possible to show intermediate values by modulating the brightness of the LEDs.

This means a true analog way of showing the operation of the compressor: very fast, but with smooth transitions. The user gets an important tool for precise gain reduction monitoring – finally the relationship between acoustic and visual perception feels just right.

A crucial part in the development process of a compressor is to design its control element which reduces the audio signal controlled by voltage. This is also where the main technological concepts show the most obvious differences.

The alpha compressor uses elysia’s fully discrete and temperature compensated Passive Current Attenuator (PCA). This circuit transforms the incoming signal into a current which is then reduced controlled by voltage. The triggering can be compared to that of a VCA with a predictable characteristic curve. The core consists of sixteen discrete transistors which are kept at a defined temperature by an exclusive heating system, avoiding unwanted fluctuations.

Unlike industrial ICs, the PCA works purely passive – it has no internal amplifiers and only reduces incoming signals. Not until the signal leaves the PCA, a downstream amplifier stage converts the current into voltage again, and the make up gain is added by a separate stage. The characteristic curve of the control voltage has been optimized for mastering applications, resulting in precise but gentle gain reduction performance. If stronger and more audible results are on the agenda, the switchable feed forward mode is a perfect alternative.

Further advantages of the PCA are its enormous bandwidth and its extremely fast control speed. Even huge level changes are done in just a few microseconds – an important premise for a fast and flexible compressor. The PCA plays an important role in the overall sonic character of the alpha compressor. It enriches the music with natural overtones by producing a healthy amount of K2 and K3. The sound becomes richer in character and inspires especially digitally mixed productions with that certain something.

Some delicate circuit components can be influenced by the surrounding temperature easily. The main reason for this circumstance is the discrete transistors that can react very sensitively to variations in temperature (that can – depending on the place of installation and operating time – happen by all means).

With the T16 Heater elysia presents a system that keeps constant conditions and reduces thermal fluctuation to a minimum. This system was inspired by high-precision measuring instruments. It features up to 16 discrete transistors in a massive copper ring which is warmed up to a definite temperature.

A surrounding ceramic cap the isolates the copper ring and keeps it from cooling down quickly and prevents heat emission into the housing at the same time. Once the system has reached its working temperature, it only needs little current to keep its status at the same level.

An electronic control circuit is responsible for only a small variance of just a few degrees. The procedure is known from high end tube gear: the alpha compressor should be granted an adequate warm-up-time in order to experience it in its best form.

The alpha compressor features high-quality conductive plastic potentiometers for all its parameters throughout. The 21 steps make a precise recall very easy, which is very convenient not only for mastering purposes.

This potentiometers have a very long lifetime. Since the release of the alpha compressor in 2008, we have not had to replace any single potentiometer.

Moreover, these potentiometers are also considered to be the best for audio in terms of sound quality. In order to keep the component tolerance as low as possible, we measure all potentiometers and match them as stereo pairs with a software routine we have written exactly for this purpose. The obvious benefit of this effort is that the left and the right channel behave exactly the same.

Through this optimization step, it may happen that some potentiometers feel slightly different, since different values are used.

To achieve maximum sound quality with a high end product like the alpha compressor, it is very important to use a perfect internal audio wiring, too. If the wrong material is used, audible losses will be the definitive consequence. After several experiments we tested the VOVOX sound conductors which we liked immediately. The difference is obvious and a noticeable improvement; therefore all alpha compressors are equipped with this cable in series since May 2007.

VOVOX sound conductors transmit the whole frequency range with minimal losses. The sound becomes brilliant and tangible. At the same time, the bass range becomes more powerful and precise. Due to their special design, these cables are able to transmit hard peaks very directly. Music becomes more brisk, dynamic and powerful.

During recording and replay, VOVOX sound conductors enhance the spatial reproduction of music. Even in very complex situations, the sound remains transparent with clear contours. Often a significant enlargement of the stereo base is observed. As the sum of many different effects, there are also noticeable improvements to the ‘charisma’ of music: The clarity and immediacy of the sound signals offer an aura of calmness, solidity and sovereignty.

To explain the advantages of discrete circuitry we’d like to present a comparison that might seem a little odd at first: Audio technology vs. cooking! If somebody uses instant meals exclusively, he will have to accept whatever comes out of the box. A creative cook, however, focuses on his own special recipes and ingredients.

In this respect, integrated circuits (ICs) are pretty similar to packet soups: they are cheap, mainstream and they simply do not match haute cuisine. So if you want to design an analog audio device 100% according to your own demands and ideas without any compromise, there will be no way around a discrete design.

The alpha compressor follows this philosophy consequently. Not only all audio paths, but also the power supply, sidechain and even the driver for the LED meter have specially been designed for this unit, based of the exclusive use of discrete components. A truly unique recipe!

In addition, our mastering compressor operates in permanent class-A mode. This means that the transistors are always conductive, resulting in the absence of crossover distortion and providing a pristine sonic base: The general sound character is always wide, open and transparent.

Specifications

Frequency response

• <10 Hz - 200 kHz (-0.5 dB)
  

THD+N @ +15 dBu, 20 Hz – 22 kHz

• Stereo mode (Direct): 0.0039 %
• Stereo mode (Compressed): 0.009 %
• M/S mode (Direct): 0.014 %
• M/S mode (Compressed): 0.034 %

Noise floor, 20 Hz – 20 kHz (A-weighted)

• Stereo mode (Direct): -95.8 dBu
• Stereo mode (Compressed): -89.3 dBu
• M/S mode (Direct): -95.6 dBu
• M/S mode (Compressed): -92.3 dBu

Dynamic range, 20 Hz – 22 kHz

• Stereo mode: 122 dB
• M/S mode: 118 dB

Maximum input level

• Stereo mode: +28 dBu
• M/S mode: +23 dBu

Maximum output level

• Stereo mode: +27 dBu
• M/S mode: +28 dBu

Impedance

• Input: 10 kOhm
• Output: 68 Ohm

Power Consumption

• 100 W max

Dimensions (W x H x D)

• 483 mm x 133 mm x 405 mm
• 19″ x 5.3″ x 16″

Weight

• 16 kg / 35 lb