LA-2A Style Compression: How Tube Optical Compressors Actually Work
There is a reason optical compressors from the 1960s are still being used, copied, modified and reinterpreted today. It is not because they are particularly accurate dynamics processors. In fact, compared with a modern VCA compressor, an optical design can look rather primitive: its timing is imprecise, its response changes with the signal, and there may be very little for the engineer to adjust.
That is also exactly why it works.
The classic LA-2A approach combines a tube amplifier with an optical gain-reduction element. Instead of forcing every transient through a fixed attack and release envelope, the compressor reacts differently depending on how loud the signal is, how long it remains loud and what happened immediately before it.
This behaviour is usually described as program-dependent compression. The term is often repeated in discussions of optical compressors, but it is worth looking at what it actually means electrically and why it matters when working with real audio.
What makes an optical compressor different?
All compressors need some element whose gain can be controlled. The difference between compressor families is largely a question of how that job is done.
A VCA compressor uses a voltage-controlled amplifier. A FET design uses a field-effect transistor as a variable element. In a vari-mu compressor, the operating point of a tube itself is changed to produce gain reduction.
An optical compressor takes another route.
A portion of the incoming signal is sent to a detector. The detector drives a light source, and that light falls on a photosensitive element. As illumination changes, the electrical resistance of the photosensitive element changes as well. This resistance change is then used to control the level of the audio signal.
In simplified form:
audio signal → detector → light source → photosensitive element → gain reduction
The interesting part is that the optical element does not behave like an ideal electronic control device.
It takes time to react. More importantly, it does not necessarily return to its original state at the same rate at which it entered gain reduction. Its recovery can also depend on how strongly and how long it was illuminated.
Once that happens, attack and release stop being just two fixed numbers.
What “program-dependent release” really means
This is probably the most important part of an optical compressor's behaviour.
Imagine a short transient that produces 5 dB of gain reduction for only a moment. The optical element is illuminated briefly and then begins recovering.
Now take a sustained vocal phrase or a long bass note that also reaches 5 dB of gain reduction, but holds the compressor there for several seconds.
The meter may show the same maximum reduction in both cases, but the recovery afterwards can be quite different.
After the short transient, the compressor may release relatively quickly. After sustained compression, part of the recovery can take considerably longer.
This is why describing an optical compressor as having, for example, a single “one second release time” does not tell the whole story. In many opto cells there is an initial recovery followed by a much slower tail, and the proportion between the two depends on the preceding signal.
From a circuit designer's point of view this behaviour is inconvenient. From a musical point of view it can be extremely useful.
The compressor is effectively carrying a small amount of memory from one moment of the programme into the next.
Attack is only part of the story
Optical compressors are often described simply as “slow compressors”. That is an oversimplification.
The optical element does require some time to respond, so the very beginning of a transient can pass before maximum gain reduction develops. But what the listener perceives is determined by much more than the nominal attack time.
The shape of the gain-reduction curve matters. The behaviour of the detector matters. The amount of compression matters. The previous few seconds of signal matter as well.
This is one reason an opto compressor can reduce the level of a vocal quite substantially without making consonants and articulation disappear. It does not necessarily clamp down on every short peak in the way a very fast FET or VCA design can.
That does not make it better. It makes it suitable for a different job.
What the tubes are doing
The presence of tubes tends to dominate discussions of vintage compressors, sometimes more than it should.
It is tempting to explain the sound of a tube opto compressor by saying that the tubes make it “warm”. Electrically, that is not a very useful description.
In an LA-2A style architecture, the tubes provide amplification around the optical gain-reduction system. Depending on circuit topology, operating level and loading, the tube stages can introduce gradual nonlinearities and harmonic content as signal level increases.
Transformers, where they are used, contribute their own behaviour too.
But the compression envelope itself is strongly influenced by the detector and the optical element. Replacing an opto cell with another device having different response characteristics can change the way the compressor feels even if the surrounding tube circuit remains identical.
So the characteristic behaviour is better understood as the result of the complete system:
detector + optical element + tube gain stages + operating levels + circuit topology
Reducing all of that to “tube warmth” misses most of what is happening.
Why optical compression often sounds unobtrusive
The word “smooth” is used so often in audio marketing that it has almost lost its meaning. In the case of an optical compressor, however, there are concrete reasons why gain reduction can be less obvious to the ear.
The compressor does not instantly jump between two gain states. Reduction develops progressively. Recovery is often non-linear and can happen in several stages. After sustained compression, the slower part of the release can keep the level from rushing back up between phrases.
As a result, the listener may hear a vocal or bass part becoming more stable without clearly hearing the compressor moving.
This is an important distinction.
A compressor designed to control drum transients may deliberately make its action part of the sound. An optical levelling amplifier is often doing the opposite: changing the envelope while drawing as little attention as possible to the fact that it is changing it.
Optical, FET, VCA and vari-mu: different tools for different problems
Compressor types are sometimes discussed as though one topology has to win. In practice, the more useful question is what kind of level change needs to be controlled.
FET compressors can react extremely quickly and can be excellent when transient control or aggressive compression is part of the desired sound.
VCA compressors offer predictable timing and precise control. That makes them useful on buses, drums and applications where repeatability matters.
Vari-mu compressors tend to increase their effective compression ratio progressively as the signal is driven harder. Like optical designs, their behaviour can be strongly program dependent.
Optical compressors are particularly good at levelling signals whose amplitude changes over a longer musical timescale: a vocal phrase, bass notes, clean guitar or sustained keyboard parts.
There is considerable overlap, of course. A good engineer can use any of these devices outside its stereotypical application.
The topology does not dictate the source. It mainly tells you something about how the compressor is likely to react when the source changes.
One weakness of the traditional opto approach: low frequencies
A compressor detector does not hear music the way we do. It responds to electrical level.
This can become a problem with low-frequency material.
A kick drum, low synth note or strong bass fundamental can contain enough energy to drive the detector hard even when that part of the signal does not sound excessively loud.
The compressor then reduces the level of the entire signal.
On a bass guitar this may be exactly what is wanted. On a full mix or a complex keyboard signal, it may not be.
A large kick transient can pull down vocals, guitars and everything else at the same time. Once the kick disappears, the gain comes back up. That movement is what is usually described as pumping.
This is not specifically an optical-compressor problem, but the relatively long and program-dependent release of an opto design can make the effect particularly noticeable.
One practical solution is to filter the signal feeding the detector while leaving the actual audio path untouched.
A modern interpretation: the RSE Audio RA-2A
The RSE Audio RA-2A (RA2A) was designed around this general tube-and-opto philosophy, but it was never intended to be a component-for-component copy of a vintage LA-2A.
The aim was to retain the behaviour that makes a tube optical compressor useful while giving the circuit a little more flexibility in a modern studio.
The RA-2A uses a fully tube audio path together with a tube-based sidechain and detector. Gain reduction is performed by a custom optical element whose recovery was selected to give a relatively quick initial release followed by a longer tail.
That second part is important. If the compressor is only touched briefly by a peak, it should not remain buried in gain reduction for an unnecessarily long time. But after a sustained loud passage, a slower recovery can prevent the level from jumping back up abruptly.
This is one of those areas where looking only at a specification such as “release time” tells very little about how the compressor will actually behave on music.
Compressor and Limit modes
The RA-2A retains a deliberately simple control concept.
In Compressor mode, gain reduction is progressive and intended mainly for levelling. In Limit mode the effective compression ratio is considerably higher, allowing firmer control of peaks.
Limit mode should not be confused with a modern digital brick-wall limiter. The optical element is still present, so the timing remains dependent on the programme rather than becoming an instantaneous peak ceiling.
Why the Comp. HPF is in the sidechain
The Comp. HPF was added specifically to deal with the low-frequency detector problem described above.
It is important to understand what this switch does not do.
It does not remove bass from the audio output.
The filter is located in the detector path. When it is engaged, the lowest frequencies contribute less to the control signal that determines gain reduction.
Without sidechain filtering:
large low-frequency signal → strong detector response → more gain reduction across the whole signal
With the Comp. HPF engaged:
large low-frequency signal → reduced detector response → less unnecessary gain reduction
The original bass remains present in the audio path.
On a vocal recording this feature may barely matter. On a stereo keyboard, bass-heavy programme or mix bus it can make the compressor behave quite differently.
The intention is not to make the detector ignore the low end completely. It is simply to stop the lowest octave from being allowed to make every compression decision.
Where optical compression tends to work well
Vocals
Vocals are probably the most obvious application because the problem is usually not isolated peaks. It is the changing level of complete words and phrases.
A singer can move several decibels between two consecutive phrases without anything being wrong with the performance. A levelling compressor can reduce those differences before they become a fader-riding problem later in the mix.
With moderate gain reduction, an opto compressor can do this while leaving much of the initial articulation intact.
This is also why looking only at the gain-reduction meter can be misleading. Five or six decibels on an optical compressor may sound far less dramatic than the same number produced by another topology.
Bass guitar
Bass is perhaps an even better demonstration of program-dependent behaviour.
A bass line contains notes with different fundamentals, different sustain and often quite different electrical levels. The objective is usually to keep those notes sitting in roughly the same part of the mix without flattening the beginning of every note.
Optical compression is naturally suited to that job.
The slower part of the recovery can also help prevent the level from rising too quickly between sustained notes.
There is no rule saying that bass must be compressed with an opto circuit, but there is a good technical reason why this combination became common.
Electric guitar
A heavily driven guitar amplifier is already a compressor to some extent, so additional compression may not always be necessary.
Clean guitars are different. Arpeggios, picked chords and sustained melodic parts can benefit from a few decibels of levelling, particularly when the aim is to keep them present without making every pick transient louder.
Keyboards and synthesizers
Electric pianos, analogue synths and other sustained sources can work very well with optical compression, but they also demonstrate why detector filtering can be useful.
A synth patch may have a strong low fundamental together with relatively quiet midrange information. A conventional detector can respond mostly to that fundamental. Reducing its influence in the sidechain allows the compressor to follow the complete sound more naturally.
Mix bus
An opto compressor on the mix bus is not automatically a good idea.
That is worth saying because optical compression is sometimes treated as though it improves anything simply by being inserted.
On the right mix, one or two decibels of slow, program-dependent gain reduction can create useful movement and gently control level. On another mix, the same circuit may simply react to the kick drum and make the whole arrangement breathe in an unwanted way.
This is where a sidechain high-pass filter becomes particularly practical.
The correct setting is the one that improves the mix when properly level matched, not the one that causes the gain-reduction meter to move the most.
How much gain reduction?
There is no particularly meaningful universal number, but some starting points are useful.
For gentle levelling, 1–3 dB of gain reduction is often enough to change the stability of a source without making the processing obvious.
At around 3–5 dB, compression becomes more substantial but can still sound quite natural on vocals and bass.
Optical compressors can sometimes tolerate 5–10 dB or more surprisingly well, especially on individual sources. Whether that is useful is another question.
Rather than chasing a number, it is usually better to increase compression until the source begins to sit where it should, compensate the output level, and compare with bypass.
Level matching matters. A signal that is 1 dB louder will often be perceived as better even when nothing else improved.
Is every LA-2A style compressor an LA-2A clone?
No.
“LA-2A style” has gradually become shorthand for a broader design concept: tube amplification, optical gain reduction, relatively simple controls and program-dependent timing.
Two compressors following that principle can still be electrically very different.
They may use different tubes, transformers, optical elements, detector circuits, operating voltages and output stages. Even apparently minor changes to the opto element can produce a noticeable difference in attack and recovery.
For that reason, the RSE Audio RA-2A Tube Opto Compressor is better described as RSE Audio's interpretation of this design philosophy rather than a reproduction of one particular vintage unit.
The basic idea is familiar. The implementation is our own.
Why this old idea still works
The original tube optical levelling amplifier comes from an era when designers did not have digitally controlled VCAs, look-ahead processing or a screen showing a gain-reduction curve.
Yet the underlying problem has not changed very much.
A singer still moves closer to the microphone. One bass note still lasts longer than another. A keyboard patch still has too much energy in its lowest octave. A mix can still need two decibels of control without sounding obviously compressed.
An optical compressor handles these situations in a slightly unusual way. It does not react only to the signal that exists at this exact instant. The behaviour of the optical element means that the previous part of the programme influences what happens next.
From a measurement perspective, that makes it less predictable.
From a musical perspective, predictability is not always the most important property.
That is why the basic concept has survived for more than half a century, and why there is still room to develop new circuits around it rather than simply reproducing the old ones.
Learn more about the RSE Audio RA-2A Tube Opto Compressor →