Combining pitch-correction, multi-part harmony generation, the ability to change vocal character, and multi-effects processing, TC-Helicon's Voice Pro is a complete vocal production suite inside a single box. But does it sound any good?
This new TC-Helicon unit aims to cram into a 2U rack unit practically every vocal production tool currently available: generic processing such as compression, EQ, delay, and reverb; sophisticated pitch-correction; up to four-part harmony generation; and vocal character modelling. It ships with a printed reference manual, a power cable, and an AES-EBU two-in/eight-out breakout cable. The hardware looks very professional, and all the buttons and knobs have a sturdy feel to them. The 2U format means that the various controls have a little space to breathe, and it also leaves room for a generous 320 x 240-pixel display. To the left of the display are the power switch and the four System buttons: Home, Utility, Help, and Tap. Home simply returns the display to a basic mode that shows the input and output levels, how the amount of processing varies with time, which processing blocks are operational in the current program, and which parameters are currently assigned to the four Edit knobs.
The Utility menu provides access to settings for selecting analogue or digital in/out and configuring MIDI assignments. The Help button is, indeed, helpful: pressing it displays context-sensitive tips on whatever function is currently being used — a very nice touch. As might be expected, Tap is used to set the tempo for use with any of the delay-based effects that the Voice Pro offers.
To the right of the display, the controls are organised into a number of groups. The functions of most of these are described more fully below but, essentially, the Effects and Virtualead controls provide access to all the key processing options. The Output controls provide Bypass and Mix buttons, the latter bringing up a useful screen where the levels of the dry vocal, processed lead vocal (called the Virtualead), harmony parts, and effects levels can be adjusted from a single location. The functions of the Preset buttons — Browser and Store — are self-explanatory, although browsing the 250 factory or 250 user presets is made considerably easier because the search can be filtered by processing and application type.
The rear panel features analogue I/O via four balanced XLRs. The Voice input is used to feed the unprocessed mono lead vocal into the Voice Pro from your mixer or DAW. The Aux input provides a separate input to just the µMod (modulation), Delay, Reverb, and Transducer effects — useful if you also want to use the Voice Pro as a standard multi-effects processor on a send-return loop. MIDI In and Out sockets allow MIDI data to be sent between the Voice Pro and a MIDI sequencer/keyboard. A standard BNC word-clock input is also present, and digital I/O is provided via AES-EBU on a 25-pin D-Sub connector. When using the digital I/O, the outputs can be configured to split the dry lead, Virtualead, and four harmony voices to individual outputs for external mixing.
Slightly more odd inclusions on the rear panel are a pair of Ethernet ports and an RS232 port. The manual suggests that the Ethernet ports are to be used for implementing software upgrades and that the RS232 port is 'currently not supported'. I can only assume there is some technical reason for the use of the Ethernet ports rather than either USB or Firewire connectivity.
For those taking advantage of MIDI to connect the Voice Pro to a sequencer, one additional element not included in the box is very much worth acquiring. A free software editor (produced by Psicraft) is available for as a 13MB download from the TC-Helicon web site. The editor can run as a stand-alone application or as an MFX or VST plug-in.
For the purposes of testing, I hooked up the Voice Pro via its analogue I/O to my mixer and fed the unit with some vocal parts via Cubase SX. The manual makes some comments about the latency induced by the processing of the Voice Pro, and a Utility page allows this latency to be adjusted between 'minimum' and 'high' settings, with various stops between. This setting only affects the latency of the Virtualead output (between 15ms and 34ms), whereas the latency of the dry lead and harmony voices stays at 2ms and 34ms respectively regardless. While the manual does not make it explicit, I assume the lower-latency settings are achieved by some compromise on the processing quality. This seemed to be confirmed in auditioning the Virtualead vocal in isolation, although the differences were only really noticeable on the very lowest-latency setting and on heavily processed sounds. In a studio context, the latency can, of course, be compensated for by sliding track timings within your sequencer.
Although most operations can be carried out using just the Navigate cursor keys and main Select/Scroll wheel, things are made considerably easier by the four Edit knobs, the functions of which are defined by labels on the currently active screen. Given the multitude of processing elements, editing is a little daunting at first. However, even with only a little experimentation the process soon becomes familiar, and the unit is by no means difficult to use. Processing blocks are accessed via the appropriate button (Harmony, Multi-FX, Pitch, or Character, for example) and pressing the OK button will toggle a block on and off. Editing the parameters for each block simply requires a combination of the cursor keys and the Soft Knobs and, given the excellent display, it's actually quite speedy.
While the Voice Pro's generic effects are not the key selling point of the unit, they are worthy of some comment. Under the Multi-FX group are the µMod, Delay, and Reverb effects, and these operate as send-return loops within the Voice Pro — a separate send control is available for the dry lead, Virtualead, and harmony parts to this effects chain. The µMod effect provides a variety of chorus, flange, panning, and other detuning-style effects, with some weird and wonderful special effects thrown in for good measure. There is plenty of user control on the effects, with over a dozen editable parameters for those people that like to tweak. Each of the other generic effects has a similar degree of editing, so you can see that this is a serious multi-effects processor.
The Delay and Reverb blocks both offer flexibility and, to my ears at least, very good audio quality. For example, the reverbs — which I suspect borrow heavily from parent company TC Electronic's other processors — are very nice indeed. The Dynamics block, which provides compression, limiting, and de-essing, and the EQ block are equally well specified. For example, the compressor features threshold, ratio, knee, release, and make-up gain controls, and can be positioned before or after the EQ. The display also includes a graphic showing the amount of gain reduction. The EQ block has low and high shelving filters with variable frequency, and two fully parametric mid-bands. A further low-cut filter is provided with presets to roll off below 60, 80, or 120Hz.
The Transducer section provides a range of more specialised effects. Here, you can simulate the sound of the voice being passed through a radio, a megaphone, a telephone, and various amplifiers, and there are some additional sound-mangling options, including a nice tube emulation that is capable of adding just a little warmth.
Three modes of pitch-correction are provided by the Voice Pro. The scale-based automatic mode is, in essence, similar to the automatic mode of Auto-Tune. The user specifies the key, scale, attack rate, and amount of pitch-correction to be applied, plus a 'window' around each scale note where correction will occur. As with Auto-Tune, custom scales can be defined if required. Provided that you do not expect it to rescue a very poor singer, the results are transparent and comparable to those obtained with Auto-Tune. That said, if you do want to abuse the pitch-correction for creative effect, then there's plenty of scope to do that.
The second mode of pitch-correction also uses scale-based automatic adjustment, but allows the user to override this on a temporary basis via MIDI. Any MIDI note played will become the target note until released. This can be used to fix the occasional note that lies outside the specified scale. The final pitch-correction mode simply takes this MIDI control further for full 'manual' playing of the pitch, although this requires some MIDI editing within your sequencer to create a natural result.
The largest number of the Voice Pro's 250 factory presets focus on automatic double-tracking and harmony generation. For double-tracking, under the Harmony block's Humanise page, the user can control the degree of pitch and timing tightness to the original dry vocal. In addition, the style and degree of both vibrato and scoop (the 'scoop' that most singers apply at the starts of notes as they slide into the required pitch) to be added to the double-tracked vocal can be adjusted. As suggested by a range of the preset names (for example, Subtle Double, Thicken, Loose Double, Octave Double), almost any combination of super-tight through to downright sloppy double-tracking can be created.
Further variety can be added by including elements from the Character block to give the double-tracked voice a different character to the dry lead vocal — more on this in a moment. While some of the Character-based presets do require careful use, the double-tracking can be very convincing even when heard in isolation. Within a full mix, even some of the more extreme presets are totally acceptable. Given the degree of control available, the ability to simply create the 'ideal' double-tracked feel from your original vocal line is hugely impressive.
The various harmony presets allow up to four-part harmonies to be generated in exactly the same fashion as with the double-tracking — and with the same degree of control over the relative timing and pitch tightness for each of the four voices. The Voice Pro can construct the harmony parts in a number of different ways, but the three most useful modes are Scale, Chord, and Note, each of which will produce a different style of harmony. In Scale mode, the key and scale need to be set (custom scales can be created if required) and the pitch of each harmony part is set relative to the pitch of the dry lead vocal. For example, harmonies might be set at +3rd or -6th relative to the lead voice, and a range of ±2 octaves is available. This mode tends to produce fairly dynamic harmony parts, as every change in pitch in the lead vocal also produces a change in pitch in each harmony part.
In Chord mode, the harmonies are created intelligently from chords received via the MIDI In (most easily done via a sequencer, although you could play them in live). This tends to produce slightly less pitch variation within the harmony voices, and would certainly suit some kinds of material better than Scale mode. For total control over the harmony construction, Notes mode uses the pitch of MIDI notes to define the harmonies. Harmonies are only generated when notes are being played, and this can work very well if you only require certain words within a vocal line to be harmonised.
There are all sorts of additional settings within the Harmony block that offer further variety and help create a natural end result — volume, pan, and gender can also be adjusted. Many of the presets also add in some further elements from the Character block to help differentiate the various harmony voices even further. As with the double-tracking, the end results of the automated harmony parts can be very, very good indeed. The more subtle ones work well enough a cappella, but within a mix even some of the more adventurous presets can work well enough to be convincing. What's more, the Voice Pro provides a means of experimenting with harmony parts that simply wouldn't be possible with a live singing group — unless they happened to be extremely talented, patient, and have the stamina to perform for hours on end!
We last looked at TC-Helicon's voice-modelling technology in the review of the TC-Helicon Voice Modeller plug-in back in SOS February 2004. The Character processing block contains six sub-sections: Resonance, Spectral, Inflection, Vibrato, Breath, and Growl. The Resonance settings produce changes in the position of the harmonic content of the voice. For example, the Length parameter simulates a change in the length of the vocal tract and can almost be considered a 'gender' control. Subtle changes work extremely well, but if you try to do anything more extreme the processing does become apparent, at least when heard in isolation. The Spectral settings attempt to model the natural EQ control a singer has over their own voice. The controls are similar to standard parametric EQ, with three bands provided, but the results are intended to be more natural and dynamic.
The Inflection settings allow the pitch, timing, and level of a voice to be adjusted, and this includes the scooped pitch inflections most singers generate at the start of a note. Obviously, changing these settings is one way in which harmony voices can be modified to sound less like the original part upon which they are based, but they can also change the performance character of a lead vocal. Again, more extreme settings enter special-effects territory and there are even several Inflection presets that create an instant 'bad singer' vibe! The Vibrato settings do pretty much what you might expect, and the results can be very realistic indeed, generating a completely different vibrato character.
The settings within the final two parts of the Character block — Breath and Growl — do require some care in use if a natural result is required. The Breath settings can enhance the breathy or raspy sound of a voice. The various styles included range from 'Natural' through to 'Heavy Rasp', with a couple of additional special effects thrown in. In contrast, the Growl settings simulate the situation where the throat is constricted. Growl styles such as Grainy Voice or Rough Edges give a clear idea of what is intended! Usefully, the amount of Growl can be related to the input level, with more added as the voice gets louder. Both Breath and Growl can add a little 'whisky and cigarettes' edge, but only a little, as over-cooking either of these blocks soon causes the processing to reveal itself.
While the hardware user interface of the Voice Pro is very easy to use, the free downloadable editor is well worth using if you have a studio computer. On my test PC, both the stand-alone and VST versions ran without a hitch with Steinberg Cubase SX v3.1.1. In stand-alone mode, you can resize the window so that all the controls for a particular block can be seen at once, whereas the VST plug-in has a fixed window size which requires a certain amount of scrolling if you want to get at all the settings for some blocks — I felt that this was worth the extra effort, though, in order to retain my settings within each Cubase project.
In testing, both versions seemed reliable and it is possible to upload the factory or user presets stored on the Voice Pro to your computer for backup. Provided that the appropriate processing block is displayed on the Voice Pro's screen, any edits made on the computer are reflected on the Voice Pro display. Perhaps the only quirk is that, in moving between blocks (for example, from the Character block to the Harmony block) in the software, the new block is not recalled to the Voice Pro screen.
In some form or other, all the functions that the Voice Pro provides are available in other products on the market. As a straight multi-effects processor for functions such as delay and reverb, the Voice Pro is very capable, and the quality of the effects, including the reverb, is very good indeed, but there are many other hardware and software versions of these processes that also produce high-quality results and come without such a weighty UK price tag.
For automated pitch-correction, I think there is little to choose between the results that can be obtained with the Voice Pro and something like Auto-Tune. Either product will be able to tighten up the pitching of a vocal in a fairly transparent fashion, provided that you have a half-decent performance in the first place. Therefore most users would not, if budget were an issue, need to have both products in their armoury — either would get this particular job done. The Voice Pro can be used in its 'manual' MIDI pitch-correction mode to totally alter a vocal melody and, in addition, make some adjustment to the vocal phrasing. However, if this is the function of most importance to you then Celemony Melodyne 2 (as reviewed back in the SOS January 2004) is probably a better choice.
Where the Voice Pro does, in my view, have a significant lead over the competition, is in the automatic harmony generation (including the double-tracking), and in the voice character modelling. Admittedly, the character options are a mixed bag, but it must be remembered just how complex a thing the human voice is, and that what TC-Helicon are attempting is extremely ambitious. Despite this sort of technology still being in its infancy, the Voice Pro's Character processing can, if used with care, produce some very convincing results that will make a vocalist appear bigger or smaller, and certainly transform them far enough to sound like a different singer. The Breath and Growl effects certainly need to be used sparingly, but they can add a little edge to a voice. The only obvious competitor to the Character processing is the Throat plug-in in the Antares Avox suite but, in my view, the Voice Pro has a clear edge here — both in terms of sound quality and flexibility.
Personally, the excellent automatic doubling-tracking and harmony generation are the Voice Pro's main asset. I haven't had the opportunity to hear the Voice Pro alongside TC-Helicon's own Voice Works (a considerably cheaper unit dedicated to just these tasks), but the results that can be obtained with the Voice Pro surpass those of any other product I'm currently aware of, including my own (aging) Digitech Vocalist and the more recent Duo and Choir plug-ins within Avox. The ability to experiment with harmony generation in this way opens up all sorts of creative possibilities and would enable producers to experiment with harmony vocals in a way that would leave a group of session singers both very tired and writing out a big session bill.
The Voice Pro will not (thankfully!) turn Will Young into Robert Plant, but it is nonetheless a very impressive piece of equipment. While I've emphasised its uses with sung vocals, the range of special-effect voices that can be created for spoken-word applications is also considerable.
The main sticking point for many potential users is likely to be the serious UK price. Whether this is worth paying will depend on whether the easy double-tracking, harmony generation, and vocal character alteration appeal, and whether you're going to find it convenient having pitch-correction and multi-effect processing in the same box.
Personally, though, if I didn't already own a Voice One I'd certainly be very tempted to flex my credit card and add the Voice Pro to my own rack.