Chromaphone – How the Magic Happens
Chromaphone is AAS’s most recent masterpiece that produces a vast variety of sound. We examine how this is accomplished in this close-up look.
by David Baer, Nov. 2014
In this article, we’re going to take a close-up look at Applied Acoustic System’s Chromaphone synth. I’ve wanted to do a piece on this marvelous instrument for a while now – I’m a big fan of AAS and how they employ physical modelling mathematical processing to produce all manner of fascinating sounds. But the world really doesn’t need yet another review lauding the praises of Chromaphone. There are already plenty out there that you can readily find with your favorite search engine.
Instead, we’re going to examine at the sound engine in Chromaphone and study how it can be programmed. First we’ll undertake a general review of the major components of the sound engine (that is, we are not going to spend any time on effects, preset browser, etc.). After that, we’ll examine three presets that, between them, nicely illustrate much of Chromaphone’s capabilities.
Sound in Chromaphone is produced by virtual resonators which are set in motion (i.e., resonating/vibrating) by virtual exciters. There are two exciters which can be used individually or in parallel. These can stimulate one or both of two resonators. Many effective presets are included that use just a single exciter stimulating a single resonator – simple enough. The plot thickens when we consider that one of the resonators can become an exciter for the second resonator. This is called “coupled mode”. In it, one or both of the exciters stimulate the first resonator, which in turn stimulates the second. We have the following illustrations from the excellent Chromaphone user manual:
A Little Excitement, Please
The two exciters are Mallet and Noise. There’s plenty to say about both of them.
We’ll start with the Mallet. The concept is straightforward: a mallet is like a drum stick with which an object is struck or a human hand tapping on a surface. But it can also be like a drummer’s brush stick. The primary control for the mallet is Stiffness, which takes values from 2 Hz to 5 KHz. How frequency relates to stiffness is far from obvious, but it’s not really important for the sound designer. If you’re simulating a drum stick, set the stiffness high. If it’s the palm of a human hand, then that obviously merits a setting of less stiff than a drum stick.
Next to the stiffness control is Noise, allowing values from 0% to 100%. This is not to be confused with the other exciter, also called Noise – we’ll get to that in a moment. The easiest way to envision the effect of mallet noise is to consider a drummer’s brush stick. It doesn’t land on the target surface with a single thump but with a lot of micro-thumps within a very brief period of time. The following image shows a recorded result of Noise at 0% 50% and 100%. This picture tells the whole story.
Next to Noise is Color, which is enabled only if Noise is other than 0%. This is set from 0% to 100%, but it actually controls the frequency of a high-pass filter on the noise source.
The overall intensity of the mallet is set with Volume in units of dB. Volume, Stiffness and Noise can be individually modulated by key position and velocity.
Finally, we have the Direct control (which applies to the Noise exciter as well). This ranges from minus infinity to plus 30 dB. With it, we can specify how much exciter noise is to be included in the generated sound. For many percussive type sounds (tuned or otherwise), some amount of direct mallet contribution is a requirement. But for other sounds, especially other-worldly, ethereal sounds that Chromaphone can excel at, keeping the direct component out of the final sound is part of the solution.
Now, let’s look at the lower exciter, Noise. Unlike mallet, which strikes immediately at MIDI note-on and contributes nothing thereafter, Noise is a continuous source of excitement as long as the MIDI key is depressed, subject to an ADSR envelope.
First, we have a choice of four filter configurations: none, low-pass, high-pass and high-pass plus low-pass. We have a Volume control, but we do not have a volume envelope to use. The ADSR envelope can, however, modulate the frequency of the filter which can accomplish much the same thing.
Probably the most common setup in presets calls for the HP plus LP filter combination. With this set, the control to the right of Frequency is Width (specified as 0 to 10 octaves). If just high-pass or low-pass is set, the Width control turns into a Q control (specified as 0% to 100%).
Finally, there’s density, with values ranging between 1 and 50K. This controls the frequency with which noise impulses occur. At low settings, think wind chimes. At high settings, think a drummer’s brush stick sliding over the surface of a snare drum surface. Somewhere in the middle, think Geiger counter.
Modulation options abound. Key position, velocity, an LFO and the ADSR envelope can modulate any of the Volume, Frequency and Density parameters.
The other main part of the picture is the resonators, which have the primary job of producing the sound (although remember that the exciter contribution and also be mixed in via the Direct control).
We have the following selection of resonator types (this directly from the manual):
- String: a perfectly elastic string
- Beam: a rectangular beam with constant cross-section
- Marimba: a beam with variable section
- Plate: a rectangular plate
- Membrane: acoustic membrane as in a drumhead
- Open Tube: a cylindrical tube with both ends open allowing one to obtain the complete harmonic series (even and odd harmonics)
- Closed Tube: a cylindrical tube with one end closed allowing one to obtain only odd harmonics
- Manual: In this mode, one can adjust the frequency of one to four different partials
Any are available for use in the top and the bottom resonator and they mostly share an identical set of controls. In coupled mode, the top resonator is the exciter of the bottom one.
First, we have Pitch, which can be set to between minus 48 semitones (four octaves) and plus 48 semitones. The upper resonator also has a pitch envelope which is not present in the lower resonator.
Under the icon of the selected resonator type we find four dots (not applicable to tube resonators). These are a quality parameter. One dot specifies the simplest modelling (and the least CPU). With each additional dot engaged, the complexity increases.
Decay and Release do what you would expect. Decay can be set from 5 ms to 100 seconds. Release is set to a percentage of the Decay value.
Then we have the mysterious Material parameter. What it actually controls is the rate at which partials decay with respect to the fundamental frequency. Values range from -3.0 to +1.0 with a default setting of -1.0. At -1.0, the first partial (one octave above the fundamental) will decay twice as fast as the fundamental, the next partial will decay three times as fast, and so on. At zero, all partials decay at the same rate. At positive settings, high partials decay more slowly than lower ones.
The Tone parameter also dictates behavior of partials, but in this case it’s the amplitude of partials relative to the fundamental. The values range from -12 dB to +12 dB per octave. Natural instruments will usually have a slightly negative setting. But there’s one more wrinkle …
Hit Position also has an effect on the behavior of partials. This parameter ranges from 2% to 50%. For a string resonator, for example, it controls where the string is set in motion. 2% is very near the end of the string; 50% is at the exact center. Significantly, we can modulate this parameter with a random source to add a “humanized” quality to the performance.
The two Tube resonators have somewhat different controls available. Since these are not used in the presets we’ll dissect, I’ll leave understanding them as an exercise to the reader. The documentation fully explains this special case.
Finally, we have the Balance control. If coupling is disabled (and both resonators enabled) it simply controls the mix level of the two resonators. With coupling enabled, things are not so easily explained. The Balance control, in this case, dictates the ease with which the top resonator sets the bottom in motion. When coupled, the resonators exchange energy. Actually what I wrote earlier about the top resonator being the exciter of the bottom in coupled mode is not the whole story. The resonators actually excite each other; the degree to which this happens is dictated by the Balance control. Confused? OK, let’s just call this a “set by ear” parameter and leave it at that. Even if you fully understood what was going on in the simulation computations, you’d probably just end up setting by ear in any case.
Case Study One – Glass Glock
The audio clip demonstrating this was made using a fragment of a MIDI file I found on-line – I don’t have a performer’s name to credit. It (and the remaining audio demo clips to follow) are recorded with zero additional processing such as EQ, compression, etc. You are hearing unadulterated Chromaphone output.
This sound is exactly the sort of thing you’d expect from an instrument named Chromaphone, and the preset is programmed in a very straightforward fashion. We have a mallet exciter stimulating a single resonator, a Marimba bar. The Noise exciter and the lower resonator are disabled.
The mallet is set to quite stiff and includes a fair amount of mallet noise that is LP-filtered fairly strongly. The expected modulation from velocity is present on volume, but velocity also moderately influences stiffness, where playing a key harder gets you a stiffer simulated mallet strike.
The Direct control, at -49 dB, is set to include a small amount of mallet striking noise in the sound, but probably less than you might expect is needed.
The resonator type is Marimba and it is set to maximum complexity. Given the high-pitched quality of the sound, you might expect that the pitch is set up by 12 or even 24 semitones, but it’s left right at zero. The high-pitched quality of the timbre comes courtesy of the Tone control, which calls for a 3 dB per octave increase in partial amplitude. So, the third partial (at two octaves about the fundamental pitch) will be about twice as loud as the fundamental.
Material is set to -0.34, dictating that higher partials decay more slowly than lower ones, but not very much more slowly.
And there you have it. For this preset, the programming delivers pretty much what you’d expect. Now, let’s get a little more fanciful.
Case Study Two – Flute 8’ + 4’
This preset set is interesting for two reasons. First, it uses the Noise exciter exclusively. Second, it ably demonstrates Chromaphone’s ability to deliver unexpected timbres with creative (and non-obvious) programming. In it, we find two String resonators delivering a very organ-like sound.
The preset employs the on-board reverb which was left engaged for the demo audio clip.
First let’s examine the Noise settings. The filter is the HP+LP combo, center frequency at 385 Hz with a wide 5 octave spread between the two filter cutoff values. Key position lightly modulates filter frequency. The ADSR envelope lightly influences Density, but again only lightly. We have an exciter that’s delivering a fairly constant stimulation to the resonators, just as you might expect of an organ type sound. Density is set to ~20K, essentially a continuous stimulation.
Now, note the Direct setting is -47 dB, imparting a slightly breathy quality to the final sound, again just like an organ flue stop would deliver in real life.
What’s not expected, given the nature of this timbre, is that it comes from the use of simulated string resonators. But here’s the essential ingredient: the complexity for both resonators is set at minimum. If complexity is increased even one notch, the string-like timbre starts to become evident and the organ-like quality is defeated. The bottom resonator is set an octave up from the base setting of the upper one (thus the “8’ + 4’” qualifier in the preset name), and the two resonators are balanced evenly.
The remaining resonator parameters are not particularly noteworthy. Upper Tone is set to plus 2 dB per octave; low to plus 1 dB per octave. Decay is very long and release is very short – just what is called for in an organ-like preset. Again, this is all in all what you’d expect, apart from the String resonators, for delivering an organ-like sound.
Case Study Three – Chiff Delay
Finally, let’s look at a preset that puts everything to work. The Chiff Delay has the “Delay” qualifier in the name because it employs the onboard Delay effect. For demo purposes, I’ve turned that off as well as the Warm Tube distortion. This preset is a delight, one of my favorite Chromaphone sounds.
The preset uses both the Mallet and Noise exciters. These stimulate the top resonator, a Plate, which in turn stimulates (via the Coupled option) a Marimba Bar. Just as in the previous case, the quality of the resultant sound is, to me at least, most unexpected. The breathy “chiff” quality of the sound – how can that come from either a plate or marimba bar? No doubt, the Noise resonator gets the credit as follows.
First, the Direct setting is at -28 dB, so we get a modest amount of exciter in the final sound. The Mallet Volume parameter is half that of the Noise Volume setting. So, we get a bit of Mallet impact but a whole lot more breathy Noise. Both volume parameters are moderately modulated by velocity, but Noise Density is strongly modulated thus. By now, you should know enough to be able to check out the other resonator parameter settings to see how things are set up.
Here’s where things start to get interesting. The exciters stimulate the Plate resonator. The pitch of both resonators left at base value, and the complexity of both are at maximum. The hit point of the Plate is near the center (or a distance of 40% from an edge) and its vibrations get transferred to the Marimba Bar at a position that’s 5% from its end.
The Balance slider is all the way at the top, but that does not mean we’ll hear only the top resonator. In coupled mode, this control dictates how the two coupled resonators interact. The manual explains how Balance in Coupled mode controls the relative stiffness of the two resonators, which in turn governs how much energy is transferred from one the other (and back again). Having read it, you’ll probably agree with my earlier assessment that this control is best set by ear.
But some further sound examples might be of assistance at this point. First, let’s see what happens when we simply disable the bottom resonator (and disable coupled mode). We’re just hearing the Plate.
In this case, the sound is pretty much what I’d expect, and clearly not at all similar to the coupled preset sound. So next let’s add the Marimba resonator back in and play with Balance. The following clip contains five brief segments with Balance set to 0:100, 25:75, 50:50, 75:25 and finally 100:0 (the setting actually used in the preset). You will note how the 0:100 setting is almost inaudible. By the time we get to 75:25, it sounds very nearly like it sounds at 100:0.
In the end, I find it difficult to understand how this sound actually comes about. The exciter settings make logical sense. The resonator combinations and coupling, however, is far from intuitive. In the end, clearly it works and does so beautifully. But I cannot imagine anyone actually sitting down and designing this sound; it absolutely makes no logical sense – or maybe it’s just that I’m dense.
I hope that this exposition has helped you to learn at least a little about how Chomaphone makes its magic. But I must confess that even after the study it required to write this article, some of what this instrument does remains just that to me – seemingly magic. Simple configurations can produce predicable results, but by coupling unlike resonators in unexpected ways, it becomes all but impossible to predict the sounds that will result.
This is not a review, but I can’t help closing with a very review-like observation. What I’d love to see in a Chromaphone Version2 would be a Serendipity button that would construct random coupled-resonator presets when clicked. In use, probably at least nine in ten of those random creations would be throwaways, but some real treasures might be within reach with a little patience – combinations you wouldn’t ever think to try because they make absolutely no sense intellectually. The Chiff Delay preset is only one of many of the factory preset content that aptly demonstrates this.