Elemental Audio Systems plugins

Available from Elemental Audio Systems’ online shop only in RTAS, VST and Audio Unit format for mac OS9, mac OSX and Windows 98, ME, 2000 and XP. Demos are available from the website. Tested primarily in Logic Platinum 5.5.1 running under Windows XP, and more briefly in Adobe Audition 1.5, Live 3 for Windows XP, Logic Pro 6.4.3 and TC Spark XL 2.8, both running under mac OSX 10.2.8.
Reviewed by David Mondrup

Firium? Eqium?? Neodynium??? No, this is not label printing on packages brought home from the pharmacy; we're talking about effect plugins here, all coming from Elemental Audio Systems (from here: EAS), a company that specializes in delivering plugins that don't take tradition at face value. The processors delivered are two EQ's and a compressor, in other words of the bread'n'butter variety, yet they don't necessarily give us user interfaces like we're used to see them in these types of effects. Since software opens up a whole new host of possibilities when it comes to GUIs when compared to hardware units, it seems the guys and gals at EAS has found it fitting to grab the opportunity to re-think how to work these beasts. In this review, we'll take a look at this new approach, as well as the quality of the processed output from these effects.

Common concepts

The plugins share the graphical approach to signal processing. The signal processing is represented graphically in an easy-to-understand manner, even if the representation presents us with some new concepts in some cases. These graphic representations are what takes up the most prominent space in the plugin interface. All the plugins are editable directly in the graphics simply by clicking and dragging the graphs. For preciser editing, and as a comfort for those more used to other ways of controlling these processes, they are also controllable by sliders with more traditional labels on them.
Like so many other serious effects manufacturers out there, EAS supplies us with some cool features that has become more or less standard with commercially available effect plugins. In this department we find the ability to save presets to files that will load in the same effect in any other host, no matter what format the plugin is currently used in, and under what platform. Another feature is swappable A/B workspaces for quickly comparing tweaks made to the effect.
So much for what these effects have in common; here's some more on how they differ:

Firium

Firium can be very briefly described as a graphic stereo EQ featuring 2 x 50 bands. It was built for mastering, but can obviously be used for other tasks too. A recent update saw a mono (1 x 50 bands) version included, to be used as a channel insert effect on mono channels.
The name "Firium" comes from the use of Finite Impulse Response (FIR) filters, as opposed to Infinite Impulse Response (IIR) filters. Long lectures can be held on the difference, and fortunately for this reviewer, a lot of these lectures are already freely available on the www, like at this site: http://www.dspguru.com/info/faqs/fir/basics.htm. Let it suffice here to say that the upside to FIRs are the elimination of phase distortion inbetween the EQs bands, as one would typically experience with IIRs. The downside is that FIRs are more CPU intensive than IIRs, though I wouldn't exactly call Firium a CPU hog, and the process introduces plugin latency, which your host may or may not be able to handle properly.

The interface that greets you when launching Firium is mostly taken up by the filter graph, which can be configured to include filter handles (sliders, if you wish), and spectrum analyzers on the audio coming into Firium, as well as the audio coming out, providing an easy way to see the difference your processing makes to the source audio. All readouts in the EQ display will also show the difference between left and right, if there is any. Surrounding the EQ display, you'll find lots of controls, though only one of them, the gain control, has any influence on the sound not reflected in the display. The rest of the controls are there to help you shape the EQ curve. There are controls to help determine how neighbouring control points should react when dragging one point up or down on the EQ display, making the EQ behave somewhat like a parametric one - or not, depending on how you set it. There are different ways to setup how one channel should react to changes made to the other channel, sliders that will do different actions to the overall shape of the curve, and spring back to centre position afterwards. There are buttons to invert and reset the curve. There are tools to scale the EQ curve horisontally as well as vertically. And, to top it off, Firium even includes it's own multistep undo history. All these tools make it very easy to do what can otherwise be a daunting task; drawing well working EQ curves on a graphic EQ using a mouse. After all, even if most graphic EQs give you a way to draw the curve rather than set one slider at a time, how many can draw a soft curve using the mouse (except for pro graphic designers). Having the curve the way you want it is pretty easy once you get accustomed to using the tools provided here.
Now, let's take a look at some of the other features in Firium.

Incorporating the ability to automate a graphic EQ must present the developer with some headaches; of course one could always provide the user with the ability to automate the individual sliders, but in a unit with 100 sliders like this one, making actual use of the automation would be so timeconsuming that it would be practically unusable. EASs answer to this is called states. Clicking an arrow brings you to an overview of 50 thumbnails showing different EQ curves, each called a state. In practice, this allows you to have up to 50 different EQ curves stored in one instance of the plugin, and you can then automate your way from one state to another. To add to the functionality of the states display, individual states are easily moved or copied between the places, and to quickly make states that helps you go gradually from one state to another, the fill feature is invaluable.

Another feature that marks Firium for mastering use is the EQ matching ability. Firium can learn and store up to 4 different EQ curves from audio sources, and use these to give the destination audio a frequency response as close to the source audio as possible. This is great for giving you a starting point for mastering your track to sound like that million dollar commercial track that you achieve to get your results close to, but can also be used for more experimental stuff, like applying the EQ curve from one instrument to another. This can help you achieve starting points for your EQ curves, or you can just get some outright weird results.

In the first example, I've taken the EQ curve of a dark beatbox pattern, recorded live in my studio, and applied it to a sampled doublebass track, in order to thin out the rather fully mastered bass sound. The beatbox comes first, then the unprocessed bass, and finally the result.

In the second example, you'll hear first a pretty bright piano pattern, then a high pitched synth patch, and finally the EQ curve of the piano applied to the synth to give it some of the piano's brightness.

In the third example, I've applied the EQ curve of the piano to the beatbox track, just for the hell of it. The result is, well, unusual.

Finally, I tried applying EQ curves of professionally mastered commercial tracks to one of my own raw mixes (an example track I made for my students some months ago). You'll hear the unmastered track first, then with the EQ curves of 3 different housetracks* applied, one at a time. Trying out different EQ curves from different sources is easy, as one instance of Firium will store four different source EQ curves, and switching between them is a simple matter of one mouseclick.

I have to admit that I used to have a hard time seeing the relevance of the graphic EQ in a DAW environment as anything but the plugin manufacturers insistence on bringing all traditional hardware audio processor types to the DAW scene too. However, EAS has with Firium come a far way towards bringing the graphic EQ to make sense, even in a DAW context. The ease of use is second to none, and the feeling that some of us know from hardware graphic EQs, that you can more or less shape the sound by grabbing a handful of sliders and manipulate them intuitively, is not only reproduced, but has even been improved with the many editing features.

Nothing is ever absolutely perfect, and I'll round off the description of each plugin by a short wishlist of additions that I'd like to see included in upcoming versions:
- as it is, the only parameter available for automation is "states". This makes sense after having used Firium for a while, and with a little practice it can take you anywhere you want to go, but I would like to see automation access to the gain parameter as well.
- to make even better use of the "states" concept, it would be cool to be able to set a transition time when switching between the states. This transition time should ideally be automatable too. This would allow for completely smooth transitions between different EQ curves, and give you even more benefit of the 50 states available.

Eqium

Equim is an EQ too, but this time a multiband EQ. Multiband EQs are found everywhere in modern DAWs and are probably among the effect types most common out there. They work by having a number of different filters that can be configured individually and will then be added on top of one another to produce an EQ curve to shape the audio going through it. Typically you'll get a low and a high shelf filter to deal with the bottom and top of the audio, and 2-4 parametric EQs to make curves in the mid ranges. In hardware units, the number of filters have been limited by the number of components going into the box, and the number of knobs the designer wanted to put on the front of the machine. Most multiband EQs have adopted the fixed number of filters approach from their hardware counterparts, but Equim differs; the manual claims that the EQ supports unlimited numbers of filters, as long as you still have CPU cycles left. Wanting to try out this claim, I reached 200 filters before deciding not to go on. I figure this should satisfy most users. Eqium features 11 different filter types, and it's primary strength is the ability to combine any number of any of these filters as you wish. After having played around with them for a little while you will find that this opens for you to shape your EQ curve in all sorts of ways, and you can actually gain a lot of inspiration simply from playing around. A filter type that I haven't seen very often before is the harmonic filter, available in 2 different strengths. They affect the harmonics, either even, odd or both, of the fundamental frequency. Ideal for dealing with ground hum and other such electric or digital artefacts, they can also be used very efficiently for elimination of standing waves, or they can be brought to use as creative tools in their own right, especially with gain values above 0.

As in Firium, the graphic representation of the EQ curve takes up the predominant space of the GUI, and can be customised in quite a few ways. The workflow is pretty simple; start out by selecting a filter type from the list, then configure it, adding new filters as needed. The created filters are represented 3 different places in the GUI, and editable in all 3 places, though not all features are editable in any of the places, so you'll need to move around the GUI a bit to get around all features for any given filter. One place in which to edit the filters is (of course) in the graph itself, where the filter's shape can be dragged using crosshairs at the hz/gain setting for the appropriate filter. Another place is the list of all created filters, where stereo settings can be made, and filters can be deleted and bypassed if needed. The 3rd place to edit the filters is the filter parameters area, where the relevant parameters for the currently selected filter can be edited using sliders, or, since version 2.1, numerical input. As said, if you mouse around all 3 places, all settings (except one - see below) for a created filter can be edited, but no single area of the GUI gives you the possibility of editing all features of the selected filter.
As mentioned, there is one uneditable parameter; you will look in vain for any way to change the type of any filter once it's created. This is of course easy to work around by simple deleting the filter and make a new one of the preferred type, but it makes a bit cumbersome to eg. try out the 2 different shelf filters to see which works better with the same settings.

With the high degree of flexibility involved in the concept of Eqium, EAS has once again had to be pretty creative to make the plugin automatable in a practical way. The answer to the task of making any number of filters automatable is called handlers. Each filter can be associated to one handler, which is used for automation access only. Automate the handlers parameters, and you will see the associated filter respond to the automation. A good, flexible and wellworking solution.

It has to be said that while the level of documentation for EASs plugins is generally high, with lots of clear instructions on how to work the plugins as well as the background for that particular processor type, the Eqium manual is lacking severely in the background department. Apart from a description of the unique Harmonic Parametric filters, there is no explanation of the different filter types and how they work, which leaves the beginner out in the cold. One could always argue that a well versed DAW user should know a parametric filter from a low pass filter, but there is in my opinion no excuse for including 2 different shelf filters without at least explaining the difference between them.

My wishlist for future versions of Eqium would include:
- A GUI update to make all areas include full editing access to all parameters, including the filter type, of any selected filter. If this means reducing the number of areas to the graph and an enhanced filters list, then so be it. It would still make the GUI more consistent in my opinion.
- An update of the manual to include explanations of the different filter types.
Now, these would seem to be pretty serious flaws, and they are, but after having played around with the effect for a while, it still seems to be highly intuitive and easy to work with. The flexibility is very inspirational, and I could actually see Eqium become my preferred track equalizer, given a little time. The flexibility alone is well worth the asking price, and so, even if the GUI and the manual lacks in some areas, Eqium is a unique piece of software.

Neodynium

This is EASs newest offering, and is quite a new grip on dynamic processing, both when looking at the GUI and the concept. Dynamics is one of those areas in signal processing that needs rethinking the most, and EAS has given us food for thoughts with Neodynium. Once I learned how to interpret the GUI, I found it to be the most intuitive dynamic processor I've come across so far, and I predict we will see a lot of products copying this approach very shortly.

I have so far seen two other ways to work dynamic processors; the most widespread is the one where you set a treshold and a ratio, defining a range of audio that should be preserved, and what to do with the audio falling outside the range. This is such a widespread concept that I wont bother you with examples of other plugins; anyone who has ever worked with the compressor will have come across this in one form or another. The other way is the input/output graph, where you in a graph define that any audio that comes in at eg. -10db should come out at -3db, and so on for all other levels, as in the graph below here (further commented in a little while). This gives complete flexibility to treat dynamics as we wish, but is a bit hard to get your head around, mostly because you'll have to get used to envisioning level in 2 dimensions (horisontally and vertically) at the same time, something we don't do anywhere else.

Neodynium offers a whole new way of seeing this, maintaining the graphic approach of the I/O chart, yet keeping levels the way we're used to see them. When watching levels, we're used to see things vertically. Faders work vertically; led level meters displays levels vertically; and the same thing happens in Neodynium, in what EAS appropriately calls the I/O map. In the left side of the map, input levels are shown vertically, and the curved lines in the middle illustrate how levels are altered before they reach the output level, displayed vertically. Very simple, once you get your head around it. The screenshots featured here show compression with threshold at -20db and ratio set to 4:1 in Adobe Audition's I/O graph and Neodynium respectively.

Apart from this new way of dealing with the GUI, Neodynium also offers ways of processing dynamics that are not normally seen in run-of-the-mill compressor plugins. In fact, even if EAS keeps referring to Neodynium as a compressor, I will keep calling it a dynamic processor, since it can be used to compress, limit, gate, expand and de-ess as well as quite a few things that you can't do with any of these types of processors. The invention being used here is what EAS calls zones (4 of them), which can be referred to as a particular input level range, which is then mapped to a particular output level range. Speaking in traditional terms, this could be seen as 4 different treshold levels, each with it's own ratio of expansion/compression (ranking from 0.5:1 - 40:1). But as before, once you understand the graphic representation, this is much easier to see than to explain - look for yourself in the first Neodynium screenshot presented here.
The controls offered in Neodynium allows you to get very surgical with your audio. The level processing is programmed in the I/O map, using the handles on the sides or in the middle of the lines dividing the zones. If you're more comfortable with the traditional way of working dynamic processors, you can also keep to setting treshold values on the input side of the I/O map and adjust ratio using the sliders underneath. Attack and release can be set for the whole process, or for the zones individually, allowing you to get very precise in avoiding pumping or breathing effects that can otherwise cause so many problems.
There's a whole section dedicated to the key audio, that is, the audio that triggers the process. This can be set to be the source audio itself, either channel of the source audio, or some other track via a sidechain function (which is unfortunately not available in all plugin formats, due to platform limitations). The key audio can be analysed in advance (depending on plugin delay compensation capabilities of the host), and it can be filtered through a freely configurable 3 band EQ, with functionality reminiscent of the Eqium plugin - set the key to source audio and set it to filter out anything but 6-8 khz (sibilants for most vocals) and you've got yourself a deesser. Or maybe you just want it to react to those too loud kickdrum beats? Use the key filter to isolate the kick, and your wish will be granted. And you can of course audition the key, both pre and post filter, and even post filter run through the I/O map if you wish.

The 4 zones concept really allows you to do precise dynamical changes to your audio that would otherwise be hard to achieve; here are a few examples;

In the first example, I've treated an accordion recording that contains long held notes as well as faster playing. The object is to expand the dynamic range of the held notes, to make them leave more space to the rest of the arrangement. The red zone preserves the top 10 db, so that the louder parts will pass through unaffected. The orange zone expands whatever comes in the -10 - -17 db range, in other words the more silent held notes. The yellow zone acts as a ceiling towards the green zone, which gates away background noise. The audio file contains first the unprocessed, then the processed audio, and the I/O map used can be viewed to the right.

The second example is taken from the same track. Wanting to create a dragging rhytmic sound, I played the bottom of a metal bucket with a dishwasher brush. This is fine, except the dynamic range of the original recording mostly goes from very quiet to extremely quiet. In Neodynium, the input is first attenuated by 7 db to bring it within a workable range. Then, the orange zone compress the loudest parts, the actual hits, a bit, while the yellow zone enhances the dragging sounds almost up to the level of the hits. The green zone distances the background sounds, while not doing away with them completely, and the red zone deals with the rather loud thump when the brush hits the bucket first. Look at the screenshot to the left to see for yourself.

As metering is of the utmost importance when doing dynamic processing, Neodynium includes 3 different brands of meters; one is a set of standard led meters for input and output, including an attenuation meter. These give a good overall view of what is happening, and this is also the place to look out for overs. They can also be used to reference what you see in the other meters, which are smarter but require some getting used to. The cloud meters right next to the traditional peak meters essentially show the same, but do so giving you a clear impression of how the energy is distributed across the audio's levels, much like an RMS meter. To top it off, you've got 3 different meters to show the key's level; 1 for the key input, and 2 for the key output, pre and post attack/release settings respectively - this makes it very easy to get a visual grip on how your attack and release settings affect the processing, and is a great help in setting these.
To round off the features found in Neodynium, you'll find input and ouput trims, and right before the final output, a brickwall limiter. Sporting 4 settings in total, with one of them being "off", this seems like the only part of Neodynium that could benefit from being made more configurable. That said, it is clearly intended for no other purpose than catching accidental overs, which it does fine, and since the rest of the plugin offers so many ways to deal with levels, you hardly miss more configuration options in the limiter department.

As stated earlier, the level of the documentation is generally high, and exceptionally so in the case of Neodynium. Every parameter is very clearly and precisely explained, as is the case with all the manuals, but the thing that makes this manual stand out in particular is the long introduction, which gives a detailed overview of compression in general, and the special Neodynium concept in particular, over no less than 18 pages. The manual even offers some ways to overcome the limitations of some plugin formats that rules out sidechain as an option for the users of these formats.

So, is there anything to wish for in future versions of Neodynium? Yes, I can think of a few things, but these are more in the "ideas for new developments" department than in the "you oughta fix this" department. Neodynium is already a very fine plugin as is.
My wishlist includes;
- more settings in the limiter area (after all). Possibly the limiter would be more flexible if the "fast", "medium" and "slow" buttons were replaced with a slider to set the release time?
- the option to mix or even reverse levels somehow - even if that would mean yet another GUI rework. You can do a lot of very detailed work with the I/O map as it is, but the placement of levels between input and output stays the same through the process. If one snare hit is lower than another in the source audio, it still won't be higher in the processed signal. Somehow incorporating full freedom to level processing, allowing any input level to be routed freely to any output level, regardless of how other levels are routed, would in my opinion be a very interesting option for the more experimental sound designers. That said, I feel pretty sure that this has been considered already, and that the current range of Neodyniums functionality is based on a choice on behalf of the developers, a choice that many mixing engineers would probably agree on.

Audio quality

As with the GUI, the output of the effects do not appear to try to emulate well known tried and tested hardware signal processors. You will not find any options to tweak, or even enable or disable, tube warmth or tape compression emulation algorithms, cause they're not there. This is simply not that kind of product. Instead, you hear exactly what you ask for from the processors, delivered as transparently as possible. If you ask Eqium for a 5 db attenuation of the audio around 1,5 khz with a bandwidth of 500 hz, you'll hear frequencies in the 1 khz - 2 khz area raised, with 1.5 khz exactly 5 db up, and no other change to the source audio. Since phase distortion between bands is a problem with many graphic EQs, one could believe that a 2x50 band EQ like Firium would be nothing short of a sonical nightmare, but fortunately the FIR technology works as intended, and you only hear exactly what you ask for - no audible artefacts to be detected. The same goes for Neodynium; if you ask for signals at -10 db to be brought up to -5 db, you'll get just that. Arguably (and ironically) exactly what the makers of many hardware devices over the years have been striving for, while making those devices that have now become the targets of many other software developers, including the sound that we have come to recognise as the way a recorded piece of music should sound, shaped by the flaws introduced where the hardware manufacturers failed their mission.
This obviously leaves the responsibility for avoiding distortion, too harsh filters and other such engineering "errors" at the user, as the effects don't assume what you want, but deliver what you ask for. This is in my opinion a very welcome responsibility to get. As a firm believer that one should ask from their devices what they do best, and that tubes belong in devices driven by them and not in a software algorithm that wont ever reproduce it fully satisfyingly anyway, EAS's goal and achievements is very much in line with my philosophy, and these plugins are a great help in getting what I want from my computer.

Conclusion

Elemental Audio Systems is clearly on a mission. A mission to bring signal processing finally and firmly into the computer realm, where effects are made not to be an on-screen replica of something already made and well known in hardware, but an intuitive programming option functioning on the current platform's (ie, the computer's) premises. As such, one would have to declare mission completed. Even if the GUIs may take some getting used to, this is only to be expected when someone comes along and declares their intention to do away with traditions. However, the occasion is just right, or, since computers in music production has been around for quite some time now, long overdue. After having played around with these effects, I discovered that it took surprisingly little time to use these as intuitively as many other apps that were originally made for computers with nothing else in mind. It is hard not to wonder why it's taken so long for someone to come along with effects like these.
If you're looking for tools that you can count on to deliver exactly what you ask from them in an intuitive way, these are very much worth checking out. Add to this that the plugins come amazingly cheap for effects at this quality level, and Firium, Eqium and Neodynium is a highly recommendable investment, well worth the asking price.

* the tracks used for demonstrating EQ matching in Firium as sources are:
- DJ Spiller: Batucada
- Can 7: Cruisin' (both taken from the compilation album "Destination Rio" by Can7)
- Black & Brown: Cool Affair, remixed by Eric Kupper (taken from the compilation album "Groove Sanctuary" by Raw Deal)