Before we start, it is of utmost important to know what we are talking about. Both plug-ins claim a certain relationship with a much celebrated unit in the world of professional audio, the Teletronix LA-2A.
Conceived by Jim Lawrence in the 60's, this opto-electrical tube compressor was the first in its genre and can still be found in many studios. The secret of its longevity lies in its smooth sound, largely due to unusual variations in time constants. In fact, the average attack time is 10 milliseconds, but it varies according to the frequency of the signal. And that's where the magic happens: the release time is not constant! It takes between 40 to 80 milliseconds to release half the signal, plus between 1 and 15 seconds to release the remainder.
This is achieved thanks to an at-the-time unheard-of mechanism, which essentially boils down to the T4 module that combines an electroluminescent panel and a photosensitive cell. The T4 not only determines the gain reduction, but also the attack and the release. To be more specific, the brightness of the panel that lights up the photocell increases when the input signal increases, and the photocell controls the gain reduction. So, if the input signal is high, the light is brighter and consequently the compressor... compresses more. But the coolest part is that the photocell has a "memory." This means that as long as the signal remains above the threshold, or if the compression is heavy, the cell "remembers" and returns to its idle state more slowly.
Despite the complexity that it conceals in its innards, the use of the LA-2A couldn't be simpler. It has a Gain knob to adjust the output level (after the compression) and a Peak Reduction knob to adjust the processing. The latter controls at the same time the compression ratio and the threshold level above which signals are processed. There's a switch to choose between Compress (roughly a 4:1 ratio) and Limit (infinity ratio) mode. And finally, it also has a three-way switch to select whether the VU-meter displays the gain reduction or the output level (+4 or +10). However, there's a subtle detail on the back panel: an R37 pre-emphasis screw... Originally Teletronix designed this little gem for radio broadcast. But the FM band generates a 17dB hump at 15 kHz. To avoid overdriving the transmitters, the R37 pre-emphasis setting applies a filter that reduces the low frequencies in the sidechain circuit of the input signal so the compressor is more sensitive to high frequencies.
So, now that we know the original a little better, let's take on its virtual renditions.
Let's start by taking a look at what Cakewalk has to offer. Originally conceived for the ProChannel module in Sonar, then as a Rack Extension for Reason up to 2012 (RE-2A), the CA-2A T-Type Leveling Amplifier is now available for all other DAWs in VST and AU format. On PC you need at least Windows 7 SP1 (32-bit or 64-bit) and an Intel Core 2 Duo processor. Mac users need at least OSX 10.7.5 and a 64-bit Intel Core 2 Duo processor. On both platforms it requires 1GB of RAM, 100MB of space on your hard disk and a 1280x800 display.
Regarding the plug-in itself, Cakewalk states that they carefully modeled the tubes as well as the now famous T4 module, but without specifying the exact period of the LA-2A reproduced, which is very important, as we will see later. On the interface side we find, from left to right, the Limit/Compress switch, the R37 pre-emphasis control, the Peak Reduction and Gain knobs, the display selector for the VU-meter (which offers the possibility to monitor the output volume on a dBFS scale), and finally the power switch. Do note that the R37 isn't in the "Flat" position by default. It is halfway, so to speak, which means part of the lower spectrum will be ignored automatically by the detection circuit... a rather odd choice given that the factory setting on the hardware unit is "Flat." Among the hidden features that are not available on the original, the CA-2A provides an external sidechain input to trigger the processor with another track (to achieve effects like ducking or pumping), and a "Fast Reset" mode that allows a faster attack time to avoid "pops" after a long silence. By default, the plug-in is in Classic mode, emulating the response of the original, "wake-up pops" included.
Now let's see what the competition has to offer.
This is not Universal Audio's first attempt at virtualizing vintage units. Actually, the UAD family of DSP cards owns its big success, at least partially, to one of the first plug-ins to recreate the LA-2A in the digital world.
Still available today, the Legacy LA-2A displayed the typical behavior of the compressor, but without the R37 setting or the modeling of the transformer nor the input/output distortion, all for the sake of sparing DSP resources. But with the expansion of the product range with the UAD-2 cards, DSP resources aren't much of a problem anymore! That's why the brand now offers the LA-2A Classic Leveler Plug-in Collection bundle with the promise of a more authentic sound. Available in VST, AU and RTAS formats since the launch of UAD Software version 6.5, the bundle requires at least Windows 7 64-bit or Mac OS X 10.6.8 Snow Leopard, 2GB of space on your hard disk, a 1280x800 display, and a UAD-2 card or an Apollo interface. Do note that v6.5 does not support UAD-1 cards. Therefore, if you still have one of these cards around (like I do), you might as well kiss it goodbye...
The bundle includes three different modellings of the LA-2A:
• The Teletronix LA-2, based on one of the earliest models
• The Teletronix LA-2A Gray, that recreates a unit from the mid 60's (the Jim Lawrence era)
• The Teletronix LA-2A Silver, which takes after a unit from the late 60's (the Bill Putnam era).
All three offer the R37 pre-emphasis control, the Gain and Peak Reduction knobs, the VU-meter display selector, and the Power switch. Only the Gray and the Silver feature the Limit/Compress switch. The LA-2 doesn't have a limiter mode.
Now that the players have been clearly defined, it's time to get serious!
The test was performed on a PC with an Intel Core 2 Quad Q6600 2.4 GHz running Windows 7 64-bit and a UAD-2 Quad. We used the Reaper v4.32 (64-bit) as our main sequencer together with a TC Electronic Konnekt Live audio interface.
The plug-ins we review here are based on different versions of the LA-2A, so it makes no sense to simply copy the settings from one to another in order to compare them. That's why we proceeded in the following way: we applied the same gain reduction to all of them (measured with the VU-meter) and then we compensated with the output gain to have the same level (measured with the DAWs virtual mixer). All examples follow the same order: "dry" signal (unprocessed), CA-2A, LA-2A Legacy, LA-2, Gray, and Silver. Please note that when using the R37, the Legacy version only appears as a reference since it does not feature this particular control.
We'll begin with the LA-2A's favorite domain: vocals. Let's apply a gain reduction of -5 dB.
- 01 voix dry 00:44
- 02 voix ca2a 00:44
- 03 voix legacy 00:44
- 04 voix la2 00:44
- 05 voix gray 00:44
- 06 voix silver 00:44
A first observation: the controls of the different versions are way too disparate to obtain a comparable result. All display the typical behavior of Teletronix's gem and provide a nice foundation to the vocals without making the compression seem too obvious. However, the UAD bundle seems a bit more "colored" in the mid-lows and the Silver model seems to give a slightly higher loudness.
Now let's try with the same compression level but with the R37 pre-emphasis control fully on.
- 01 voix dry 00:44
- 07 voix r37 ca2a 00:44
- 03 voix legacy 00:44
- 08 voix r37 la2 00:44
- 09 voix r37 gray 00:44
- 10 voix r37 silver 00:44
As expected, the compressors are less responsive to the energy of low frequencies, so the gain reduction is applied to the high-mids and the highs. This results in a slightly less aggressive sound in all cases, while the Universal Audio seem a bit rounder. Note that the effect of the R37 feels more evident on sibilant sounds, mitigating them, which can be useful if you do not have a proper De-Esser.
Let's take a more violent approach now with a gain reduction of -10dB and pushing the output gain a bit.
- 01 voix dry 00:44
- 11 voix 10 ca2a 00:44
- 12 voix 10 legacy 00:44
- 13 voix 10 la2 00:44
- 14 voix 10 gray 00:44
- 15 voix 10 silver 00:44
The compression is much more evident, but the results of the UAD Legacy and the CA-2A are quite usable, even if the latter lacks a bit of volume. On the other hand, the UAD bundle shows another face with a very large output level on average and some distortion! That's the famous modelling of the transformers and the input/output distortion that Universal Audio boasts about, and it really rocks! Sure, with such extreme settings on vocals it is hardly useful (although...), but it suggests some nice options. It is also worth noting that the volume of the LA-2A Silver is clearly higher than its siblings.
Enough vocals, let's move on to a Wurlitzer piano and reduce -10dB.
The rather slow attack time makes the first chord clack, which is especially evident on the CA-2A. It's the classic "pop" signature of the beast after a long silence. Fortunately, the CA-2A has the "Fast Reset" mode to deal with this problem. This effect is less present on the Silver, which seems to have a faster attack time. The compression itself is much less evident with the new UAD bundle than with the two others. In any case, all five plug-ins provide a nice sustain to the piano and that's due to the famous and typical release of the LA-2A. Once again, Universal Audio's newcomers display a coloration in the low-mids and the Silver is louder than its siblings.
The results are similar with the R37 in action. The highs are more compressed, which emphasizes the mids.
It is now the turn of a bass line to experience a -10dB gain reduction.
This bass line is relatively antsy and would probably need a compressor with a fast attack time. The Cakewalk version does quite well and that's due to a significantly faster release time than the others. This can also be seen directly on the VU-meter with a frantically dancing needle, which makes reading difficult. The LA-2A Legacy has a much harder time and the twang of the bass becomes a little soft. Moreover, we must also confess that we had to crank the gain to the max in order for it to have the same level as the others. The LA-2 provides a very nice color, but its time constants are definitely too slow for this kind of phrasing. The VU-meter confirms this with a darting needle. The speed and coloration of the last two seem more adequate here, especially the Gray.
Now it's the turn of guitars. The first sample is a rhythmic part to which we applied -5dB of compression with the R37 filter on to tame a bit the shrilling highs. The second one is a sort of riff to which we applied the same compression level but without any filtering.
- 19 rythm 01:09
- 20 riff 01:08
The transparency of this type of compressor is impeccable. We are certainly not winning a lot of level because the peaks pass through, especially with the CA-2A and LA-2, but we gain a certain consistency that makes placement in the mix easier. For this sort of very light processing, all five plug-ins are very similar in this regard.
Although at first glance these toys don't seem very apt to record drums, we still gave it a try. First the bass drum at -7dB and no R37. We then activated the filter.
- 21 kick 01:15
- 22 kick r37 01:15
The results are surprising. The slow attack time allows the attacks to go through and we are still able to gain some power, which is very nice. The filter makes it possible to get a less muffled sound. We most also mention that the Legacy struggled to reach -7dB. The Gray and Silver models perform surprisingly well with a beautiful medium volume, probably due to a faster attack.
The results are confirmed with the snare at -10 dB and the R37 on. The attack and resonance improve, except with the LA-2 whose slowness is not really suited for this type of sound.
Finally we applied -7dB of gain reduction on the drums bus, first without the filter and then with the filter.
- 24 drum 01:08
- 25 drum r37 01:08
The CA-2A and the Legacy aren't very comfortable with heavy processing that becomes too easily audible. That being said, the use of parallel compression could be interesting. The new UAD bundle comes out really well, especially with the use of the filter which prevents the ride from standing out too much. The three results are widely usable with a very fine grain and a significant loudness gain. The choice of which one of the three to use will be a matter of taste.
By the end of the test we were tempted to use the virtual LA-2A on the mix bus. We first did an “enhanced rough mix” with compression, EQ and a few effects, but no automation. The result was then passed through the five plug-ins with a very soft gain reduction (-3dB).
- 26 mix dry 01:28
- 27 mix ca2a 01:28
- 28 mix legacy 01:28
- 29 mix la2 01:28
- 30 mix gray 01:28
- 31 mix silver 01:28
Nothing extraordinary here. We noticed a certain volume gain but it seems that it entails a slight loss of stereo width. Do note however that the color added by the LA-2, Gray and Silver is very pleasant and they all provide a higher volume than the others.
Before concluding, some words on the more practical side. Cakewalk's CA-2A is very efficient in terms of CPU resources. During our test, no instance of the plug-in ever used more than 2% or 2.5% of our Intel Core 2 Quad Q6600 2.4 GHz processor. Suffice to say that with a newer machine you will have enough power to deploy many more instances than you will ever need. As for the Universal Audio plug-ins, you can stack up to 96 stereo instances of the Legacy with one UAD-2 Quad, which ought to be more than enough. However, the new models are much more resources hungry... With the same card you will just be able to use up to 16 stereo instances (or 24 mono), and without using any other plug-ins on the UAD-2. So, keep in mind the capacity of your system if you do not want to be caught short.
Who won? Only you can tell. And we invite you to not only listen to the sound samples on a good audio system, but also to try the demos with your own sounds. Personally, we believe that the LA-2A Legacy has aged a bit when compared to the rest. While it certainly respects the temporal characteristics of the celebrated original, it offers less functions and lacks some character with respect to the others. Nevertheless, it will prove very useful whenever you are short on DSP resources. For its part, the CA-2A (€99) seems very comfortable in almost any situation. Moreover, its R37, Fast Reset and sidechain options are very welcome. As for the new UAD bundle ($149, introduction price; afterwards $299), the sound character it provides is undoubtedly an asset and the diversity of each plug-in is very useful to deal with all sorts of situations.
Download the audio files (in FLAC format)