Not content to rest on its laurels after it debuted the stunning VoicePrism last year, TC-Helicon has unveiled the next generation of vocal processors. More than simply an upgrade, the VoicePrismPlus uses TC-Helicon's new voice-modeling technology, which could fundamentally change the way many of us record.
Like its predecessor, the VoicePrismPlus is at home both onstage and in the studio. Think of it as a full-featured channel strip — complete with preamp, EQ, effects, and dynamics — coupled with an intelligent four-voice harmony processor. What sets VoicePrismPlus apart is the addition of voice modeling on the lead voice — for the first time, you have control over the physical characteristics that make an individual voice individual. You can alter the apparent size and shape of the singer's mouth and throat, add or subtract breathiness and rasp, create realistic vibrato, and even change gender without a trip to the doctor.
Owners of the original VoicePrism haven't been overlooked — all the new features, including digital I/O, voice modeling, and expanded effects, are available in a card that fits into the expansion slot in the back of the older unit (see the sidebar “Upgrade Path”).
SAME, ONLY DIFFERENT
The VoicePrismPlus user interface is the same as the original: edit and utility pages on the display screen are grouped by function, Menu Tab keys navigate through nested pages, and four assignable soft knobs change parameters. Global functions are on the right of the data wheel, and edit functions are on the left (see Fig. 1). As on the VoicePrism, selecting Harmony or Effects mode limits the data wheel so that it shows only presets related to those modes (the unit ships with 127 factory presets). Each preset can be overwritten and restored. Turning the Lead, Harmony, or Effects-output knobs all the way down effectively mutes their signals.
Next to a ladder-type input meter are LEDs that show MIDI activity and also indicate when the pitch-detection circuitry has locked onto the lead voice, which is vital for harmony functions. LEDs indicate the sample rate (44.1 or 48 kHz) and flash to warn of an unstable digital-clock signal. Context-sensitive Help screens containing useful information that augments the manual are accessed using a single button.
The most obvious change is on VoicePrismPlus's back panel, where the new VoiceCraft card occupies the expansion slot covered by a metal plate in the original (see Fig. 2). The card sports a pair each of AES/EBU and S/PDIF digital ins and outs — handy for processing recorded tracks without leaving the digital domain. The left and right sides of the digital inputs are routed to either the Lead or Aux inputs with the newly added Utility/Routing page. Signals sent to the Aux input can be mixed into the two internal effects generators along with the lead and harmony voices. A new Digital I/O page offers dithering options for the VoiceCraft's digital outs. Internal processing is always at 24 bits.
A second coaxial S/PDIF output, a carryover from the original VoicePrism, now acts as a direct 24-bit digital stereo out for both the mic/line (on the left side of the stereo out) and analog aux inputs (on the right side). That lets you send the processed and dry lead vocals to two different digital targets. I wish this great feature had been mentioned in the manual — I learned about it during a call to TC-Helicon on a related subject.
LIVIN' IN HARMONY
The VoicePrismPlus retains the original's four harmony-generation modes, which cover almost any situation you may encounter. Each mode offers extensive control and as many as four distinct voices (see my review of the VoicePrism in the August 2001 EM).
Shift mode creates harmonies at a fixed interval above or below the lead voice, much like an old-fashioned pitch shifter. Scale mode constrains harmonies to a user-defined scale and key. Though Scale mode works best with simple diatonic material, there are some interesting possibilities using alternative scales. Shift and Scale modes are either Smooth (harmonies follow subtle pitch variations in the original voice) or Stepped (harmonies are pitch corrected).
The impressive Chordal mode creates harmonies from chords entered manually or using MIDI — great for songs with complex chords or key changes. Chord recognition with MIDI works well. The Step function allows you to configure sequences of as many as 50 chord changes and move through them using an optional footswitch for MIDI-free operation. With the ability to store as many as 50 user-defined sequences (called songs), the VoicePrismPlus can handle even the most demanding gig.
The two Manual modes are the ultimate in realism and can generate as many as four independently moving harmonies from incoming MIDI data. Input can be on a single channel (Notes mode) or on four adjacent channels (Notes 4CH) for even greater control. Each harmony voice responds independently to a variety of humanizing parameters: timing, scoop (a vocalist's habit of bending up to a desired note), vibrato, tuning, level, panning, and effects. For even greater authenticity, many of these can be randomized.
A simple twist of a knob modifies a variety of formant-based parameters, creating male and female voices. Icons — from a husky bass singer to an angelic soprano to tiny babies and aliens — let you identify the gender settings at a glance. Preset libraries for gender and other parameters make quick changes a breeze. You can access almost all editable functions on the VoicePrismPlus with MIDI messages. The mic preamp and onboard effects are solid and add considerable value.
One caveat: the VoicePrismPlus needs time to analyze and process the incoming signal. Setting the system latency lower than 20 ms in the Utility menu sacrifices quality somewhat, which is a necessary trade-off onstage where even a slight delay could throw you a curve.
Acoustically, the human voice is simply a bunch of frequency components produced in a resonant space. Features such as the length of your throat and the shape of your mouth contribute fixed-frequency resonances called formants. That's the reason simply pitch shifting a vocal can turn even Barry White into a Munchkin. But formants aren't the whole picture. Your voice is a wind instrument: you can push harder for a bluesy edge, lighten up to get a whisper, and slide up to a note or hit it dead in the middle.
TC-Helicon says voice modeling technology is “real-time reshaping and resynthesis of the human voice.” With it, you can add breath, vibrato, rasp, and growl; change inflection; and simulate the effects of head and chest resonances.
But don't expect to hear your favorite singer's voice modeled here. An individual's style is dependent on more than just the size and shape of the head and throat. No box made can capture phrasing, intonation, and taste, though I'm sure someone is working on it.
As with the original, the VoicePrismPlus's effects presets are designed for solo-voice processing. The presets show off some of the processor's tricks, such as subtle thickening and wild gender transformations. Names like “Wonder?,” “Broadway,” and “Old Bluesman” offer some idea of what's in store. Voice modeling is always available on the lead voice, even in the Harmony presets.
Although fun to use, many of the presets sounded overly processed and artificial; thankfully, editing on the VoicePrismPlus is simplicity itself. The Lead menu lets you set the dry/processed mix level and offers separate pan controls for each as well as a control to detune the modeled voice. The next five screens control various parameters. The options in the Vibrato, Inflection, and Glottal screens are what you'd expect. Inflection lets you imitate the way a singer scoops up or down; Glottal offers nonpitched vocal sounds that create breathiness, rasp, and growl.
Two additional groups of parameters require more definition. Spectral selects various EQ curves to simulate the control a singer has over his or her voice. These curves are separate from the two EQ blocks. Warp modifies formants to model various vocal-tract dimensions. According to the manual, Spectral and Warp are designed to interact; first select a Warp preset, then modify it.
Each of those five screens has two menus: Style and Amount. Style selections in the Glottal menu include Delta Blues, Whisper-Light, and Sweet Voice; Vibrato selections include Ballad, Pop-Diva, Opera Tenor, and the ever-popular Sheep. With names like those, it's easy to dial up just the settings you need.
I wish the Warp styles were as descriptive. Some names, such as “Palate-Thick,” are fine, but “Transmute” doesn't convey much information. Incidentally, onscreen choices differ from those in the abbreviated manual. The Transmute styles create subtle changes that affect each voice differently, sometimes radically so. The nine style variations are not a continuum, nor do they lend themselves to descriptive names.
Armed with that knowledge, I recorded a vocal and looped it. I dialed in about 40 percent of the Warp amount and scrolled through the styles, one after the other. Transmute 2 wrought an immediate improvement. My voice sounded richer and less tired, as if I'd gone back and sung the part again with better technique. In contrast, Transmute 3 exactly simulated the effect of a head cold. The best way to use the VoicePrismPlus is to isolate the modeling parameters to pinpoint the effect, and then just use your ears.
I next set about fixing a track recorded by a female singer. The original performance suffered from a deep, chesty attack and lots of breath — quite a challenge. After a few minutes of tweaking, though, I came up with a subtly sweeter voice that sounded far better for the track. That result alone would make me say that voice modeling is a fantastic addition to any producer's toolkit.
A NEW YOU
What about the manufacturer's claims that the VoicePrismPlus can create entirely new vocal textures? I selected tracks by a number of vocalists and went to work. As before, a style or effect that sounded great on one voice wouldn't cut it on another. Yet in every case I was able to make fundamental changes that sounded natural, turning a pure tenor into a late-night rocker and darkening up an alto to the point that you might be forgiven for thinking she had a beard.
I can't say I'm thrilled with the sound of all of the transformations; even tiny amounts of the Whisper styles overpower the lead voice. The same is true of some of the portamento settings: a little goes a long way. Radical changes almost always sounded synthetic, though that could be a good thing if you are looking for a sound that no one else has.
For the record, the technology used to create gender effects in the harmony voices differs from voice modeling. Just for fun, I recorded a track with a single gender-mutated harmony, then processed it as if it were a lead voice. The result was a much more realistic virtual singer.
WISH UPON A STAR
The VoicePrismPlus has twice as many EQ effects as the original processor: now there are two bands for the lead voice and two for the harmony voices. Likewise, the effects have been upgraded with longer delay times and smoother-sounding reverbs. The compressor lacks fine control, but at least now there's one each for lead and harmony voices.
My wish list is pretty short. Separate outputs for the lead and harmony voices and insert points would be useful, but the increased cost might push the unit out of reach for most users. A tutorial and a printed list of parameters, styles, and presets (with descriptions of what they do) would also be nice.
THIS YEAR'S MODEL
Voice-modeling technology isn't perfect, but few innovations are. Not all the voice-modeling styles sound realistic, and sometimes the effect overwhelms the lead vocalist. Yet it is light years ahead of anything else I've heard.
The VoicePrismPlus is a snap to use, but it isn't a one-size-fits-all effects box. In every case, however, the VoicePrismPlus created effects that were both useful and interesting, which is all anyone can ask for. When you add up all the things that the VoicePrismPlus does exceptionally well, its value is obvious.
Acoustic musicianMark Nelsonlives and works in Southern Oregon's Applegate Valley. A reviewer once described his voice as “not unpleasing.”
FEATURES4.5EASE OF USE4.5AUDIO QUALITY4.0VALUE4.5RATING PRODUCTS FROM 1 TO 5
PROS: Easy to use. Voice modeling allows unprecedented control over lead voice. Realistic harmonies. Extensive MIDI implementation. 24-bit processing. Digital I/O.
CONS: No independent outputs for lead and harmony vocals. No insert points. Only two bands of EQ for lead and harmony voices.
Own one of the original VoicePrisms? TC-Helicon hasn't forgotten about you.
The VoiceCraft Human Voice Modeling Card ($599) plugs in to a slot on VoicePrism's the rear panel. It adds voice-modeling technology as well as 24-bit digital I/O, DSP power, new algorithms, and other enhancements to the VoicePrism, upgrading your unit to the same specifications as the VoicePrismPlus. The upgrade kit includes EPROMs to upgrade your operating system to take advantage of all the new features.
Already upgraded? Check your software version. The current software, version 1.05, adds more processing power to the delays, fixes a few minor bugs, and decreases the lowest latency settings for smoother live use. Download it at www.tcsupport.tc.
Analog Inputs(1) balanced XLR (mic); (2) balanced ¼" TRSAnalog Outputs(2) balanced ¼" TRSDigital Inputs(1) S/PDIF; (1) AES/EBUDigital Outputs(2) S/PDIF; (1) AES/EBUA/D/A Conversion24 bits, 100 dBA dynamic rangeControl PortsMIDI In, Out, Thru; (1) ¼" TRS footswitchSampling Resolution24-bit, 44.1/48 kHzHarmony TypesShift (Smooth, Stepped); Scale (Smooth, Stepped); Chordal; Manual (Notes, Notes 4CH)Vocal Processingthickening; spectral; warp; glottal; inflection; vibrato thickening; humanizing; compression; noise gate; EQEffects Processingchorus; flange; delay; reverbDisplay128 × 64 pixel backlit LCDDimensions2U × 8.2" (D)Weight7.8 lb