EMT 240 VS PLUGIN (English)

SUBJECT OF TEST: EMT 240 vs  plate reverb PLUGINS

There is a question that we often ask ourselves: how close do my reverb plugins sound to the true hardware units and particularly plate reverbs? A lot of AB tests were made on compressors but we found nearly nothing regarding reverbs online.

This is a comparison between our EMT 240 gold foil and 9 plugins that emulate plate reverbs. For our test we have chosen three convolution reverbs and six algorithmic reverbs.

We are a studio based in Paris since 2006 that uses a lot of plugins  not only because they save us a lot of time but also because we like the way they sound.

You’ll find the complete analysis and methodology on our web site melodiumstudio.com

EMT 240 GOLD FOIL:

The EMT 240 uses a metal plate to create reverb. It’s funny to imagine that the vibrations of the metal plate will give us a similar result as wall reflections in a room. We might think that using a very thin metal plate (18 microns which is 5 times thinner than a sheet of paper!) would give a brighter sound with more treble but in fact it has quite a dark sound. The high frequencies decline quicker on the metal plate than the low frequencies so it seems to be a wise choice to use a metal plate to mimic the natural behavior of sound absorption in the air. 

When listening to the samples, the EMT 240 sounds a bit darker compared to all of the other plugins (except for the Arturia 140 and Altiverb 140). The reverberated sound frequency response of the EMT is approximately linear between 115Hz and 8kHz. From 2KHz-8KHz we can see that it remains linear but at a lower level. We noted an important lack of high frequencies in the reverberated sound starting at 8KHz. Is this lack of definition the haunting thing that we like? This is much more complicated and you will find some answers to this in the conclusion.

Surprisingly, the EMT 240 is the only reverb that has high frequencies in the attack (0-35ms) than these digital emulations. A possible explanation could be that it takes a certain amount of time for the metal plate to diffuse the lower frequencies while on the other hand a very short time to diffuse the high frequencies. It acts almost like a deesser but offset in time which can be quite practical on vocals! However, regarding the body and the tail of the reverberated sample, the EMT 240 remains the darkest.

REVPLATE 140 (ARTURIA):

It’s an emulation of the EMT 140 (a reverb machine with less treble than the 240). This plugin is surprisingly unbalanced with a lot of mediums (between 400Hz and 2KHz), a small deficit in the bottom and characterized by a large lack of high frequencies. It’s the only plugin that has less treble than the hardware EMT 240. The rest of the plugins we tested were the opposite (containing a larger amount of high frequencies). Compared to the EMT 240, the sound was much more muffled and boxy. The energy of the reverb is concentrated in the medium frequencies. For example, on the voice sample the energy of the reverb starts to descend at 4KHz in comparison to the EMT 240 which starts to descend at 8KHz. It’s probably closer to the EMT 140 than the 240. Nevertheless, it’s certainly a useful tool to fit vocals in the mix because it only gives the reverberated frequencies that we need.

CONVOLOGY XT (IMPULSE RECORD):

This plugin is sold with a preset that is an emulation of the EMT 240 with impulse responses however, it sounds very different from the hardware unit. The sound has a very metallic quality. There is a lot of energy in the high and medium high frequency range (between 1KHz-8KHz~). In contrast to the EMT 240 which begins to descend around 2KHz, this plugin stays relatively linear throughout the high medium frequency range. Surprisingly, this plugin doesn’t sound close to our EMT 240 despite having an impulse response preset based on a real hardware unit.

VINTAGE VERB (VALLHALLA):

We tested the mode plate 1970s which is supposed to be the closest preset to the EMT 240. We observed this plugin had a high cut at 8KHz but yet it contained more energy in the high medium range (1KHz-8KH~) than the rest of all of the plugins (except the Convology XT). This gives it a much shinier and brighter sound than the hardware EMT 240. When we tested it on the voice sample, it added much more sibilance (around 5dB more at 8KHz) than the hardware EMT. We wanted to put a deesser right after the Valhalla! However, it also contained a greater amount of sub bass frequencies (below 100Hz).

We decided to use the Vintage Verb from Vallhalla because it is a well known plugin. However, Vallhalla also offers another plugin specifically for plate reverb called the “Vallhalla Plate”. We tested this plugin on “steel” plate mode and the sound was closer to our EMT 240 than the Vintage Verb. It has a darker sound with less sibilance.

LITTLE PLATE (SOUNDTOYS) :

This was one of three plugins that sounded closest to the EMT 240 (or least furthest). It was modeled after the EMT 140 but its sonic character is much closer to our hardware EMT 240 than the Arturia 140. When tested with a white noise sample, it was approximately linear from 60Hz-4KHz with a slight drop after 4KHz while our EMT was linear from 115Hz-8KHz. Despite this difference, this plugin is among one of the closest to the EMT 240. When tested with a drum sample, it greatly accentuated the sub frequencies around 60Hz while the hardware unit accentuated primarily the low medium around 200Hz. Hopefully there is a cut off on the little plate!

LUSTROUS PLATES (SLATE DIGITAL / LIQUIDSONICS) :

We tested this plugin on the “steel” preset and it sounded as close to the hardware EMT 240 as the Little Plate. It also has more low bass frequencies than the hardware unit and when tested with a white noise sample, it was practically linear from 20Hz-8KHz (except for a slight bump around 1KHz). In comparison to the little plate which only has three parameters, this plugin is much more complete and diverse with parameters for adjusting modulation, EQ, damper, pre-delay, and stereo width. Thats enough to get close to the frequency response and stereo field of the EMT 240.

IR1 (WAVES)

This is a convolution based reverb plugin that’s already about ten years old. For the test we used the “plate” impulse response setting and adjusted the decay time to match the EMT 240. We discovered this plugin to be pretty close in sound to the EMT like the two plugins described above. The frequency response was similar to the EMT 240 from 150Hz-1KHz but we noticed a small bump around 1KHz and another 4KHz which contributed to a brighter sound. Like the Vallhalla there is a high cut around 10KHz, but it’s still brighter than the EMT 240. Beyond that, we had a feeling that we could hear a vibrating plate.

R-VERB (WAVES)

Another plugin we tested was the R-Verb from Waves. Although it was not as close as the others we discussed, we achieved a close sound with the “large dark plate” preset. This plugin has an EQ integrated which contained a cut around 2KHz to darken the sound but also a large bump at 600Hz which totally unbalanced the plugin with an over pronounced subs. It’s a very good cheap Swiss knife that has been proven reliable for a long time. For the record, the H-verb from Waves was discarded from the test because the plate presets were generally very far from reality!

ABBEY ROADS PLATE (WAVES)

We tested the Abbey Roads plate reverb from Waves and it came quite close to the sound of the EMT 240 a bit like the Little Plate. More specifically, it’s an emulation of the four 140 models that we can find in the famous Abbey Road studios.

We tested the four options (A, B, C, D) and the plate setting A was the one that sounded closest to our EMT 240. Although it contained more energy in the high medium frequencies (more “s” sounds in the voice sample) it still resembled the dark signature sound of the EMT.

ALTIVERB 7 (AUDIO EASE)

We also decided to test the Altiverb plugin, which is the industry standard plugin for convolution reverb. However, the Altiverb doesn’t offer an EMT 240 preset so we tested the EMT 140 preset instead. As expected, it sounded like a 140: much darker than our 240. Arturia 140 and the Altiverb are the only plugins that took such a risk. The difference between left and right side within the Altiverb was more pronounced than the other plugins which gives us greater depth.

SOUND SAMPLES :

CONCLUSION:

We noticed that none of the plugins tested sound exactly alike to the hardware EMT 240, and some of them sounded drastically different. The ones that came closest to the overall sound of the EMT 240 were the Littleplate, Lustrous plate, IR1, and Abbey Roads Plate. The IR1 and R-Verb from Waves also sound quite close but not as close as the three plugins we just mentioned. 

We discovered that reverb is not only a question of EQs and the distribution of energy over time (attack, body, and tail). For example one of the closest plugins, the Little Plate has nearly the same frequency response and dynamic distribution (especially during the tail) than the hardware EMT, but despite those very close figures, we had the feeling that we could hear two different space qualities. So we started to dig for more. Here is what we found. But before we dive in the differences, let’s have a look of how a reverb functions.

HOW DOES A REVERB FUNCTION?

When we hear reverberation (natural or artificial) we have the impression that it’s an echo that goes very fast, so fast that we can’t hear the silence between these echoes. To recreate these fast echoes artificially, there are two main techniques. There are algorithmic reverbs which are based on mathematic calculations while convolution reverbs are based on recorded samples (impulse responses) mainly to replicate famous rooms (opera houses, theaters, concert halls, etc.). However, these techniques can be applied to replicate any situation (spring, plate, rooms).

CONCLUSIONS OF THE COMPARATIVE TEST

LEFT AND RIGHT SIDES

When we observed the waveforms of each reverb, we noticed obvious differences between left and right sides which clearly appeared not to be correlated to phase. We noticed also that each reverb had its own way of distributing the energy through time. There are certain plugins that added a boost in energy during a fraction of a second.

RANDOMNESS

One characteristic that is shared by all of the plugins is that their reverberated sound is predictable. All of the plugin constructors use their proper algorithm but we observed a variety of behavior with the EMT which brought us to the conclusion that it probably has a random behavior due to the qualities of its metal plate.

When sending one input source into plugins (no matter what plugins), you  always obtain the same reverbed sound. However, when we fed the EMT the same sample five times, we obtained five different reverbed sounds, a bit like when you hit a key on the piano multiple times.

Some of the plugins had parameters to replicate this quality of randomness. For example, on the Little Plate, you can add an element of randomness with a “modulation” switch but nevertheless it stays consistent and predictable.

HARMONICS

  The randomness of the EMT goes further than left and right differences. We noticed differences in the amount of energy within the fundamental and the harmonics between left and right sides (when tested with a sin wave). Just like the differences in left and right, there is a random behavior with the harmonics created. When we sent the same input source through the EMT 240 multiple times, it gave us a different result in terms of harmonic distribution every time. Unlike the plugins which did not add any harmonics or too little to mention, the EMT had anywhere from 3-4 harmonics when fed a pure sin wave.

ANOTHER UNIQUE EMT240 CHARACTERISTIC

Another characteristic we observed specifically on the EMT 240 during the drum sample test was the time it took for the reverberated sound to build. During the first three seconds the reverberated energy continued to grow until it stabilized at a constant level. This phenomenon could be due to the nature of the physical plate or due to the re-injection algorithm. It takes time for the plate to fully resonate the sound. This could be seen clearly when looking at the waveform as well. We only noticed this phenomenon to a lesser extent in the Convology XT reverb while the rest of the plugins maintained a constant level throughout.

FINAL THOUGHTS

The differences between the EMT 240 and the plugins are very noticeable when considering the following parameters:

• differences in frequency response
• differences in energy distribution over time (attack, body, and tail)
• differences in left and right sides
• differences in behavior: predictable behavior for all plugins and random behavior in the EMT240 (generally and harmonically).
• differences in build-up time (3~ seconds for the EMT)

It is possibly the combination of all these factors that give us a different sensation when comparing the EMT to plugins. Aside from the Convology XT plugin which was presented as an emulation of the EMT 240, the other plugin developers just propose “plate reverb” or EMT140 emulation preset. The ability to precisely emulate the EMT 240 will therefore not be judged.

In conclusion, the best way to obtain this EMT 240 “signature sound” is to use the hardware unit! All of these reverb plugins are very useful and practical however if you are searching for the closest sound to the EMT240 we would recommend the Little Plate (very popular algorithmic reverb),the Lustrous Plates (not as well known with plenty of parameters), the Abbey Roads Plate, or the IR1 from Waves. But remember they act like a robot, unlike the EMT 240 that act more like an animal…