Antares's Kantos 1.0 is something new under the sun. Not because it's a 2-oscillator synthesizer with analog-sounding filters and a modulation matrix; there's nothing particularly new about that. Nor because it's a plug-in that works in all the major host environments, as there are plenty of those around.
What sets Kantos apart from other software synthesizers is that instead of being controlled by MIDI, it is controlled by audio. Kantos extracts the pitch and timbral information from an appropriate audio source and uses it to control its many parameters. (Applying an inappropriate source produces less predictable, though possibly still useful, results.) Let's take a closer look at Kantos's facilities and then get into how it works.
Kantos 1.0 runs as an RTAS and VST plug-in on both the Mac and PC, under a MAS host on the Mac, and under a DirectX host on the PC. The RTAS version is compatible with any RTAS version of Pro Tools, including Pro Tools Free. Kantos requires at least a Mac G3 (a G4 is strongly recommended) or a Pentium II — class Windows machine that meets the minimum spec of whatever host you are using. It runs under Mac OS versions 8.6 through 9.x (an OS X version is in the works) and under all current flavors of Windows. I evaluated it on a dual-processor G4/800 MHz using the MAS version with Digital Performer 3.11 as the host.
Because Kantos accepts audio input, you can run it only as an insert plug-in, not as a virtual instrument. When inserted on a mono channel, you can choose mono-to-mono or mono-to-stereo versions; stereo channels get a stereo-in — stereo-out version. Once inserted, the Kantos screen comes up. The interface is an arty, neopsychedelic study in green that looks like a cross between a biological structure and a printed circuit board (see Fig. 1).
At the upper left of the screen is the Input fader, and to its immediate right is a meter showing the input level. Kantos likes a hot signal, and Antares suggests setting the level so that the red clip indicator lights on the highest peaks. To change a parameter's value, you drag the white “blob” on the value's associated slider or fader, or you type a value in the text box below or next to it. You can also use your keyboard's Up and Down Arrow keys to increment or decrement values.
The input signal flows into the Gate Generator, which is a critical element in the chain. The Gate Generator determines how triggers for new notes are extracted from the signal. How well you configure the Gate Generator for a particular input signal determines how well Kantos's output follows the input's note attacks. Setting the On and Off thresholds, Floor (a noise gate for reducing or eliminating the effects of unwanted artifacts in the source, such as hum on a guitar track), and Hold (minimum time before a note is turned off) parameters can require some work.
Pitch and timbral characteristics are extracted automatically; there are no parameters to adjust for those. Antares is reluctant to divulge details of Kantos's analysis methods, but they clearly encompass spectral transfer similar to vocoding, formant analysis, and frequency and envelope tracking.
The output of Kantos's Input section shows up in three ways: the original input signal is available to be mixed with Kantos's synthesizer output, the pitch and timbral characteristics are used to control synthesizer parameters, and an additional pitch-controlled sine wave is available to be mixed into the final output for special effects such as added bass.
The synthesizer itself has a fairly standard configuration. There are two wavetable oscillators, each of which can use any of the roughly 40 factory waveforms that appear in a drop-down menu or any 16-bit AIFF file (on the Mac) or WAV file (on the PC) you put in a dedicated Kantos audio-file directory (see Fig. 2). If that's not enough for you, Antares posts additional waveforms on its Web site. The manual states that there is no restriction on the size of the files that you can use as waveforms; however, files smaller than 400K work “most efficiently.”
An oscillator's pitch can be offset by octaves (±2 octaves), semitones (±12 semitones), and cents (±100 cents), for a total range of just over 3 octaves up or down. A pitch quantizer, shown as a small keyboard whose keys you click on to select pitches for quantizing, is available for each oscillator. Clicking on E, G, and B, for example, quantizes the output of the pitch extractor to the notes of an E-minor arpeggio. Clicking on a single key restricts the oscillator to a constant pitch and effectively disables pitch control by the input.
Each oscillator passes through a chorus algorithm and multimode filter and is then fed into the Articulator, where the timbral information extracted from the input is applied in ways too top secret for Antares to reveal. In addition to the oscillators, there is a noise source with its own multimode filter, whose output is also directed into the Articulator.
You adjust parameters for the filters (cutoff frequency and Q) and the Articulator (amount and Q) by dragging a white dot around an x-y display that is rather grandly referred to as a Biaxial Graphic Adjuster. That is a great way to tweak two parameters at once.
In addition to its basic more-than-a-vocoder processing, the Articulator has a Format Offset parameter and an Emphasis section; the latter is essentially a 3-band tone-control-style EQ (gain is the only parameter that is variable for each band). Although it might appear that the oscillators and the noise source are mixed together before being fed into the Articulator, they are, in fact, processed separately.
The Articulator's output feeds the Sub-mixer and, in parallel, a delay (up to 999 ms of delay with feedback is available). The Sub-mixer has faders for each oscillator, complete with mute and solo buttons, sine waves carrying the pitch of each oscillator, and the noise source. At the end of the signal path, the output of the Sub-mixer feeds the output Mixer, which has faders for the Sub-mixer, delay, and unprocessed original source, as well as master output. In a stereo-output version, the faders in the output Mixer have pan controls.
The basic signal path itself is fairly powerful, but the really wacky results are obtained from Kantos's Modulation Matrix (see Fig. 3). The Modulation Matrix has eight “slots,” each of which maps a source to a destination. Seven sources are available: two LFOs (with a choice of six waveforms for each); two ADSR envelope generators (EGs) labeled Amp Env and Mod Env; and the dynamics, timbre, and pitch extracted from the source. Though the two EGs appear to be identical, they're not. Unlike the Mod Envelope, the Amplitude Envelope has an on/off button that, when activated, uses that EG for the synthesizer in place of the source signal's dynamics.
There are a total of 35 modulation destinations, 27 of which are Kantos parameters. The other eight control the amount of modulation that will be applied by a source. For example, if you had LFO1 controlling filter cutoff, then you might use input dynamics to control how much of LFO1 is being applied to the filter. With that routing, the louder the input is, the more LFO modulation is applied.
Kantos includes a Tempo section, which is an internal clock that can be used to control delay time, LFO1 rate, and LFO2 rate. The tempo can be tapped or typed in.
Pitch extraction is not a simple task, and Antares strongly recommends optimizing your source as much as possible for the job. Ideally, you'll use a monophonic source with no effects and a clean articulation (that is, clear attack transients and pitches). If you want processing on the source (as you often will), then duplicate the track in your multitrack editor and apply Kantos to one version and processing to the other. Another approach is to route the source track to a bus, make two returns that both use that bus as their input, then run Kantos on one return and whatever processing you want on the other return.
Like other synths, Kantos runs into the limitations of the host's plug-in architecture. For example, it's not possible to run Kantos in a sidechain that is fed the clean source and then mixed back in after the source is processed. Nor can you use it on a send that could be “picked off” before processing. Hopefully, these and other limitations will be overcome, allowing soft devices to have more of the flexibility of real audio hardware.
The first source I used with Kantos was one that I knew would be tough: bass. (You can hear these and other examples on the EM Web site.) It was very difficult to get Kantos to track the pitch and dynamics correctly, even after I had learned to operate it fairly well. (Antares intends to put tips aimed at producing good results with various instruments on its Web site.)
The next source I tried, vocals, was one I expected it to do well with. For this test, I used one of the factory vocal presets with no tweaking. By mixing just a little of the Kantos synth with the original, I got a truly trippy effect that sounded like the voice was heavily processed through a Leslie cabinet, yet the overall sound was crystal clear. It was delightfully disorienting and got me excited about the possibilities that Kantos presents.
Continuing with my tests of Kantos's tracking abilities, I next tried vibraphone, another difficult source. The manual cites xylophone as a particularly troublesome source because of the incredible complexity in the attack; it goes on to say that vibes are even worse because they ring. But Kantos didn't do too badly. In fact, even when there were chords in the vibes part, Kantos was able to pick one of the chord tones and track it, an acceptable compromise in my view. Overall, Kantos did a better job on vibes than bass. Not surprisingly, tracking the voice produced the best results.
Kantos also did well with guitar, even when there was a little edge to the sound. When the sound got really crunchy, things fell apart more, but I found that you don't really need a completely clean sound to get decent tracking.
Antares's recommendations for the sources that give “best” results are really for obtaining the most predictable results, and it encourages users to try “inappropriate” sources to get wild and crazy sounds. I tried some polyphonic material, but the most fun was with drums, which gave some wonderful results. Rather than using loops, I used drum tracks. I started out by sending kick and snare to a bus, making an aux strip that grabbed the bus, running Kantos on the aux, then using the timbre of the input to control the Articulator amount. I also used the Tempo feature on the delay and panned the synth and delay to opposite sides. Further tweaking produced a fabulous, rhythmic analog-synthy warped-sample effect.
Kantos sounds great and includes a wide range of factory waveforms. It also enables you to use your own files as waveforms, which opens things up tremendously. Though capable of nastiness, the basic sound of the filters was reminiscent of the warm, round sound of old Moog or Oberheim filters. With the flexibility provided by the wavetable oscillators, Kantos was capable of producing full, rich sounds. It seems analog filter emulations are finally getting good, and Kantos's filters are clearly intended to have the retro sound so popular in electronica and dance music.
With optimized sources, Kantos is quite easy to use, but sources that are even slightly off the mark take more effort. Most of the work involves playing with the Gate Generator parameters to get the triggering right. Though the Gate Generator's graphic display of the input signal's dynamics was helpful, it still often took quite a bit of trial and error to get things dialed in correctly. Some sources never fully yielded good triggering.
Another challenge was the user interface: though pleasing to look at, it doesn't illustrate the signal flow as well as I would like. It would also help to have a block diagram in the manual.
The interface has a few other minor snags that were not seriously debilitating but did slow things down. For example, to change a value, you can drag a fader, type in the text box, or use the Up and Down Arrow keys to move incrementally, but you can't drag in a text box. Another quirk is that dragging in the graphic displays sometimes produces quantized values, which forces you to use text entry to get in-between values. For instance, the tutorial says to set the Gate Generator Hold time to 200 ms at one point, but dragging in the graphic display results in a jump from 194 ms to 211 ms. It would be helpful if you could hold down a modifier key while dragging in the displays to get finer resolution, as you can with the faders. Better yet would be to use ballistics, where dragging faster changes the value by larger increments and dragging slower makes the resolution finer. It would also be nice if the Tab key moved you from parameter to parameter in the interface.
These and some other shortcomings are not showstoppers, though they can be annoying when you're in midflight on a project. No doubt they could easily be addressed by Antares in a future upgrade. The lack of any form of MIDI control is more significant. It would be nice to use MIDI controllers as modulation sources, for example, and also cool to be able to change the pitch quantizing from a MIDI keyboard so that it could follow a chord progression. Kantos's lack of MIDI was a deliberate choice on Antares's part, but one with which I strongly disagree. There are also a few mistakes in the manual; for example, the required location for User Waveforms on the PC is misidentified. Antares claims a new manual will be available on its Web site by the time you read this.
Even with its flaws, Kantos still sparked my imagination and gave me all kinds of wonderful sounds. It's hard to think of two more important attributes for a synthesizer to have, and I think that says a lot. There are many places that Antares could take Kantos in the future (such as a version with surround-panned outputs), but for a version 1.0 (which never once caused Digital Performer to crash, by the way), I am quite thrilled with all the places Kantos takes me now.
Larry the Ois shifting gears and is off to take in the wine, women, song, and food of Italy for a few days. Ciao!
Minimum System Requirements
MAC: G3; Mac OS 8.6; compatible host program
PC: Pentium II; Windows 98/2000/ME/NT/XP; compatible host program
Kantos 1.0 (Mac/Win)
FEATURES4.0EASE OF USE4.5AUDIO QUALITY4.5VALUE4.5RATING PRODUCTS FROM 1 TO 5
PROS: Innovative means of intuitive control. Rich sound. Easy to get useful results from optimal sources.
CONS: Some interface limitations. Gate Generator settings can be tricky. No MIDI control.