I haven't had so much fun with a piece of gear in ages.
In recent years, movie stars have increasingly shared the limelight with computer-generated animation and special effects some so subtle that they go completely unnoticed by the casual viewer. In crowd scenes, for example, virtual extras have to an extent replaced their flesh-and-blood counterparts. Technology is edging closer to creating an artificial actor that can fool the unsuspecting observer. A comparable feat is underway on the audio front. Indeed, it's only a matter of time until you'll be able to dial up virtual singers as easily as some people can generate realistic room simulations.
Every now and then, a product comes along that provides a glimpse of that bold future. TC-Helicon's VoicePrism is such a product. It takes artificial harmony generation to a new level and offers a taste of the direction in which vocal recording and processing seem inevitably to be heading.
Harmony processors have been around for quite a while; most are based on technologies pioneered by Canada's IVL Technology. The VoicePrism is the first offering from a joint venture between IVL and TC Electronic. What sets the unit apart from the crowd are unprecedented “humanizing” control and stunningly realistic results.
Designed for studio and stage, the VoicePrism is a full-function solo voice processor with a mic preamp, equalizer, effects, and dynamics processing in addition to an intelligent four-voice harmony processor. A huge amount of processing power resides inside the attractive 2-rackspace box, yet getting up and running is as simple as plugging in a mic and twisting a knob (see Fig. 1). An onboard demo can save you the embarrassment of singing in front of the sales staff at your local music store.
The knobs and switches feel reassuringly solid. I particularly like the ergonomic slant-front design. Also, kudos to the designer who thought to recess the display panel out of harm's way; it's a small detail, but my road rack is full of gear with scratched faceplates. Moving the screen back a quarter of an inch does make a difference.
Signals enter the VoicePrism through either a front-panel XLR mic jack (with switchable 48V phantom power) or a balanced ¼-inch TRS input on the rear (see Fig. 2). An auxiliary input lets you use your favorite outboard gear along with the VoicePrism's full complement of internal effects. The rear panel also provides a pair of ¼-inch TRS analog outputs; an S/PDIF digital output; MIDI In, Out, and Thru jacks; a jack for an optional footswitch; and an IEC receptacle for the power cord (no wall wart here).
Patience has never been my strong suit; even before the delivery truck had backed out of my driveway, I'd torn open the carton, set the unit on my desk, and grabbed a mic. A tip in the well-written Quick Start Manual alerted me to a nifty feature: the five Browser buttons interact with the large data wheel to constrain selections to a particular mode one of the four harmony types or effects without harmonies. Finding just the right Preset among the 128 choices is a snap.
I began by scrolling through lead-vocal Presets. Most combine some degree of compression and EQ along with reverb and time-based effects such as chorus or delay. They sounded far too processed for my taste, but a few deft tweaks quickly yielded more satisfactory results.
Because I was still in a hurried frame of mind, I decided to forego close inspection and jump right into the harmonies. I learned that the Presets can be overwritten and that factory settings can be recalled at any time.
I have to admit that I'm seriously deficient at singing harmonies. So I grabbed my guitar and dialed up a VoicePrism Preset called Country Tune. Based on a user-defined scale, it synthesizes male and female harmonies voiced a sixth below and a third above the lead voice. For added realism, each voice is delayed slightly relative to the lead, and each has detuning and vibrato settings.
I belted out a couple of country standards, some old-time gospel, and even a close-harmony chestnut by the Carter Family. The VoicePrism tracked every vocal nuance, filling my headphones with three distinct singers. I wouldn't hesitate to use the Preset live or on a demo just as it is, right out of the box. Brothers and sisters, I am convinced.
FINDING YOUR VOICE
I spent the remainder of the afternoon scrolling through Presets, fiddling with parameters, and singing every song I know. Although she doesn't sing, my wife even joined in, realizing a lifelong dream of becoming a Motown backup singer. I haven't had so much fun with a piece of studio gear in ages.
Did I mention the VoicePrism is easy to use? Let me say it again: the designers got it right. The faceplate is split roughly in two, with real-time functions to the right of the data wheel and editing functions to the left.
Four center-detented knobs provide hands-on control for lead and harmony volume, effects level, and input gain. Surprisingly, turning the knobs all the way down doesn't completely cut off the signals. Although a handy button bypasses the harmonies, muting the lead-voice send requires navigating inside the Mix page.
A large bright meter next to the power switch displays input and output levels. Two rows of function buttons run directly above the level knobs. Each button clicks solidly and lights up when pressed, so you won't get lost. I particularly enjoy having the mic/line and phantom-power switches within easy reach.
When you select a Preset, the backlit LCD clearly shows signal routing, harmony type, key, effects type and routing, and various editable parameters (see Fig. 3). Twisting any of the four context-sensitive soft knobs changes relevant parameters in real time; a press brings up a menu with more selections. Clever little cartoonish icons, from a hefty bearded bass singer to tiny babies, identify characteristics for harmony voices. There's even an alien with an antenna the icon for some weird sci-fi effects.
O SOLO MIO
The functions are clearly marked and laid out in a logical manner. If you get lost, everything is explained in the short user manual. The VoicePrism features a dedicated Help button that gives you access to various context-sensitive information screens. There you'll even find the e-mail address for tech support burned into memory.
The VoicePrism is essentially two units: a channel strip/vocal processor and a harmony machine. I tested the channel strip first, beginning with the preamp. For that test, I bypassed everything but the preamp and tracked vocals directly to disk. I then repeated the test with a Mackie CR1604-VLZ mixer and a dbx 386 dual-tube preamp. The VoicePrism certainly held its own; in fact, I preferred it to the dbx for some voices. I also had a chance at another studio to compare the VoicePrism preamp with vintage API preamps and with those in an SSL board. Those tests confirmed, not surprisingly, that the VoicePrism preamp is not as pristine as certain high-end preamps, but it will get the job done and then some.
The VoicePrism's dynamics processing and EQ could be better. The software compressor lacks subtlety and control; except for the 1.5:1 setting, ratios are variable only by whole numbers. In addition, the base settings are not intuitive: patching the compressor into the signal chain calls up a default threshold of -60 dB with no gain reduction. After a couple of rude surprises, I learned to jump ahead a page to check the settings first and then return to assign the lead or harmonies or both to the compressor.
Equalization is limited to two single blocks, each comprising two variable highpass filters (one with variable cut/boost), two variable lowpass filters (one with variable cut/boost), and one bandpass filter (with variable Q). That is sufficient for most applications, but the limiting factor is that the two bands can be assigned to only two places simultaneously. That is, you can assign the two bands to either the lead or harmony, or you can assign only one band to each. (You can also assign either or both EQ bands to the aux input.) Vocal thickening and a noise gate are available on the lead and harmony voices at all times.
SINGING IN THE CHOIR
In general, the VoicePrism's channel-strip aspects are better suited to the stage than to the studio. (An upgrade is available that is supposed to greatly improve the EQ and compression; however, I tested only the stock unit.) Fortunately, the VoicePrism's aux input lets you patch in your own compressor or equalizer, which significantly enhances the unit's value (see the sidebar “Upgrade Path”).
The VoicePrism shines as a harmony processor; it simply blows away anything else that I have heard. Shift, Scale, Chordal, and Notes/Manual modes are four distinct harmony-generation methods that cover just about any situation. Shift and Scale modes are further divided into Smooth mode, in which harmonies follow subtle pitch variations of the lead voice, and Stepped mode, in which harmonies are pitch corrected.
Shift mode creates harmonies at fixed intervals above, below, or simultaneously above and below the lead voice, much like an old-fashioned pitch shifter. Scale mode allows you to specify the key and scale of the harmonies in the Setup menu, so you can create wonderful harmonies in parallel thirds and sixths. Chordal mode ascertains the root and chord type from real-time data the VoicePrism receives when you play a chord on a MIDI instrument. Either of the two Notes/Manual modes lets you control the exact pitch of your harmonies through MIDI.
I first used the Stepped Shift mode to correct the pitch of a wavering harmony vocal; a few minutes of tweaking saved the track. Attempts to transpose a performance by more than a few semitones sometimes yielded telltale warbling and other artifacts, depending on the quality of the vocal input and the amount of transposition. A good source voice provided the most realistic results, possibly because of a skillful singer's unwavering pitch. (I've noticed the same correlation using other pitch-correction hardware and software.) Messing around with vocal formants and effects in Shift mode yielded some outrageous sounds totally unlike anything I have heard before.
Continuing my search for the ultimate in pitch correction, I jumped over to Stepped Scale mode and muted all but a single voice. That mode worked even better than Stepped Shift; rather than transposing the track, only out-of-tune notes were affected. I also used the mode to thicken the chorus for a Scottish singer's demo, blending close-harmony clones with the real singer's voice for a much bigger sound. Although Scale mode works best with simple diatonic material, it yields some interesting possibilities using alternative scales.
In Smooth Scale mode, the harmonies retain the original voice's slight pitch inconsistencies for a more natural-sounding harmony. Chordal mode creates harmonies with as many as four parts from chords entered manually or through MIDI. Most common major, minor, dominant, and diminished chords are supported along with a generous handful of suspensions. The harmonies follow a song's changes in a realistic manner. The unit even recognized my faltering attempts at keyboard playing, quickly generating the appropriate chord type from only one or two notes. Once I had the hang of it, I used my synth workstation to sequence three slightly different choruses for an R&B song. Badda-bing!
Chordal mode features a handy Step function that allows you to progress through a sequence of chord changes (50 songs with 50 steps per song) using an optional footswitch for MIDI-free live or studio use. I picked up my guitar to test the feature and came up with a dandy campfire sing-along right there in my studio.
The Notes/Manual modes provide the ultimate in realism, generating as many as four independently moving harmonies from incoming MIDI data. Data can be received on a single channel or, for even greater control, on four adjacent channels. Setting the Chordal mode to Momentary in the Utility menu's Preferences page constrains harmonies to the duration of MIDI notes. Using that powerful feature, I generated harmonies that came and went at different parts of a song a single voice for some lines and an entire chorus for others. To better mimic real vocals, I could even stagger the releases of individual singers.
Each voice within a harmony Preset responds independently to a variety of parameters, including timing, Scoop (the typical vocalist's habit of bending up to a note), vibrato, tuning, level, panning, and effects level. The humanizing functions are many and deep, and almost every editable function is accessible through MIDI messages.
The VoicePrism lets you adjust gender settings in the various harmony modes. A simple twist of a knob modifies a variety of formant-based parameters, creating male and female voices (see the sidebar “Banish Munchkins”). Results can be startling with no effort at all, I transformed a tenor into a credible bass that fooled some pretty demanding ears. Upon learning he'd been deceived, the producer said, “I wish I'd had that on my last project!”
Sometimes achieving the most realistic effect was counterintuitive. The same tenor's voice sounded unnatural when shifted using any of the female settings, yet I dialed in formants based on a well-known male singer and voilà: instant Barbra Streisand. Who knew?
The two independent effects blocks provide the usual vocal sweeteners. FX 1 is dedicated to chorus, flanger, and a nice selection of mono and stereo delays. FX 2 provides delays and a variety of reverbs ranging from Tight Studio Reverb to Arena. I wish there were algorithms for plates and spring-type reverbs two vocal staples. Both effects blocks accept sends from the lead voice, harmony voices, and the aux input; in addition, the output from FX 1 may be routed to FX 2. Opportunities abound for customizing effects and for saving patches as part of a Preset.
Considering all the VoicePrism does, the quality of the effects is quite high. Reverbs are nice and rich, though just a tad grainy in the tails. I really enjoyed the analog-sounding Mono Tape Delay; with a little slap back, vocal thickening, and compression, you, too, can rule.
VOICE OF APPROVAL
The ability to tie the delays to tempo would be nice, but you can't have everything. Although it isn't the VoicePrism's intended use, the unit could also serve nicely as a second multi-effects box, in a pinch. Just for the record, though, the VoicePrism does not share its effects architecture with TC Electronic's popular G-Force processors.
I like the TC-Helicon VoicePrism a lot. Combining functionality that normally requires a rack full of gear with state-of-the-art harmony generation into a single unit is quite an achievement. The VoicePrism is a ball to use right out of the box, and the more you get inside and tweak, the better it sounds.
Most of my recording work involves acoustic music and relatively natural-sounding vocals, and the VoicePrism proved itself useful in those demanding applications. But given its ability to create unique vocal effects, I recommend it for remixing and other forward-looking applications. If you want a sound that no one else has (yet), the VoicePrism is a good place to start.
I'm not sold on the EQ and dynamics, and I don't think the harmonies always ring 100 percent true. The results depend largely on the source, and great singers definitely produce the most realistic harmonies.
Of course, artificial harmony is not everyone's cup of chai, but if you already use or need a vocal harmony processor, check out the VoicePrism. It's truly one great piece of gear.
Mark Nelson sang in high school until his band asked him to stop. He lives and records in southern Oregon's Applegate Valley, where his voice won't disturb the neighbors. Thanks to A Wing and a Prayer Productions for help with the review.
FEATURES 4.5 EASE OF USE 4.0 AUDIO QUALITY 4.0 VALUE 4.5
RATING PRODUCTS FROM 1 TO 5
PROS: Extremely easy to use. Lots of control. Realistic-sounding harmonies. Two effects blocks. Independently moving harmony lines through MIDI. 24-bit digital output.
CONS: No independent outputs for lead and harmony vocals. Only two bands of EQ. Front-panel gain knobs don't fully mute lead vocal or effects levels.
Audio Inputs (1) balanced XLR (mic); (1) ¼" balanced TRS (line); (1) ¼" balanced TRS (aux) Audio Outputs (2) ¼" TRS; S/PDIF coaxial Control Ports MIDI In, Out, Thru; (1) ¼" TRS footswitch Sampling Resolution 24-bit, 44.1 kHz Frequency Response (analog) ±0.8 dB @ 10 Hz-12 kHz; -2 dB @ 20 kHz Total Harmonic Distortion (analog input) 0.0026% @ 1 kHz (A-weighted) Dynamic Range (analog) 100 dBA Harmony Types Shift (Smooth, Stepped); Scale (Smooth, Stepped); Chordal; Notes/Manual Vocal Processing thickening; compression; noise gate; (2) EQ blocks Effects Processing chorus; flanger; delay; reverb Display 128 × 64 pixel backlit LCD Dimensions 2U × 8.2" (D) Weight 7.8 lbs.
TC-Helicon recently announced two new products for the VoicePrism that incorporate human-voice-modeling technology. The VoicePrismPlus Human Voice Modeling Formant Processor adds head and chest resonance, breathiness, rasp, growl, modeled vibrato, and vocal inflection to the lead-voice input. The VoiceCraft expansion card essentially upgrades the VoicePrism to the same specifications as the VoicePrismPlus and also provides higher-quality compression and EQ, better effects processing, and 24-bit S/PDIF and AES/EBU ins and outs.
What is it that enables listeners to instantly differentiate between singers, even when they're singing the same note? The answer is vocal formants, which are the identifying resonant characteristics unique to each person's vocal tract. A singer's formants are affected by several factors, including size of the head, layout of the vocal cords, and length of the throat; even bone density plays a part. Formants emphasize certain fixed-frequency bands, regardless of a note's pitch; that's why you sound like yourself throughout your vocal range.
Formants come into play any time you pitch-shift a human voice. Simply transposing the voice up or down also transposes the formants, resulting in “munchkinization” or other unnatural effects. Intelligent harmonizers such as the VoicePrism take the behavior of formants into account so as to provide more natural-sounding pitched effects.
Altering formant characteristics is the trick behind gender-bending effects too. Now you can turn a lithe female alto into a cigar-chewing male quartet. Isn't science a wonderful thing?