What is it supposed to do?

Steinberg Spectral Designs has recently released this real-time voice pitch transformer. According to product literature, VoiceMachine Processor allows you to change melodies or correct intonation by "changing the pitch of the voice while maintaining its natural character". VoiceMachine purports to pitch-process vocal recordings in a precise manner "without affecting the tone of the voice", as well as allowing the user to "adjust the voice resonances freely, changing the tone of the voice independently from the pitch". There are two included plug-ins: VoiceMachine Processor and VoiceMachine Generator.

Where can I use it and do I need other software to run it?

VoiceMachine is specifically designed to work as VST plugins on both Mac and PC, and requires another host program to run. VoiceMachine Processor works in "static" mode in any VST host program, and both it and VoiceMachine Generator have full functionality in any application that supports the VST 2.0 standard. Judging from the accompanying documentation, a major strength of this software is that is can be automated using MIDI within a VST 2.0 host environment.

What can I use it on?

The VoiceMachine manual recommends using discretion in choosing the source material for processing. Users are advised to keep the source as "clean" as possible, because extraneous background noise on the track can cause "errors in the pitch algorithm or formant misinterpretation". Creative use of non-vocal, "monophonic sound that isn't too harmonically rich" is also suggested. The clearer the source, the more precise the result.

What is the purpose of this review?

I'll give the software a basic test drive, using both female and male pre-recorded vocal tracks, and determine the general range in functionality of the plugins. I want to find out if they are as artifact-free and clean as I need a pitch and formant-based vocal processor to be for realistic recording; how far I can push the parameters and still get a reasonable result. I also want to hear what happens to a non-vocal monophonic sound processed with the plugins. It occurs to me this might be a wacky processor for my analog synth tracks!

VoiceMachine Inside the VST Host Program

Let's get started.

Installation on my Mac was a snap. The VoiceMachine installer located all the VST Plugins folders on my machine and asked me which I wanted to use. (After installation, the plugins can be copied to any other VST Plugins folders desired.) Recently I have been using Emagic Logic Audio Platinum as my multitrack recording/sequencing program, so I chose that plugins folder. VoiceMachine installs two VST Plugins: VoiceMachine Processor and VoiceMachine Generator.

VoiceMachine Processor

So what controls do I get?

VoiceMachine Processor has a Mix control: 100% is full processing, 0% is no processing, and all the points in between allow you to mix varying proportions of processed and source material (which can give you a very cool chorus effect). The GUI (Graphical User Interface) is laid out kinda like a maze, but after a couple mouse-taps of the "link" buttons, I saw it's simple: the main two parameters are Formant and Pitch. These two can be linked, or operate independently. Pitch has a Fine tuning control, which can also be linked to it, or not. You also get a Vibrato section, which seems to be a group of LFO's (Low Frequency Oscillators) that modulate the source material. You can choose among Sine, Sawtooth, or Triangle waveforms as your modulators. There's a Tempo Sync that allows you to sync the modulation to note values in your host sequencer, and you can adjust the Rate (0.1 - 10Hz), Depth (0-100%), and Delay (0-10 seconds) of the LFO's.

Add MIDI control

Each of the previously mentioned controls can be automated via a variety of MIDI controllers; hit the Setup button and you can assign to MIDI control most of the plugin parameters (thought not the LFO waveform type). A Key Control button allows you to send MIDI notes to control the plugin; in this mode, if no MIDI notes are sent, no sound will be heard. The Key Control mode also gives you additional Vibrato settings: Portamento (0-5,000ms), Tolerance (0-100%), Tolerance Delay (0-1,000ms). Under Key Control mode, there is a De-Tune meter, which shows you how much the pitch of the source material differs from the incoming MIDI notes.

Test it immediately under conditions of extreme duress!

Well, first thing I did to push the limits was disregard the manual's suggestion to use only "clean" vocal tracks. I've been working on some tunes featuring a solo vocalist playing acoustic guitar. Had a vocal track (male, raspy baritone) miked simultaneously with his guitar playing, so there's guitar on 25-30dB under the tube-pre'd, compressed vocal track. Sounds like a "dirty" track to me! And yes, as predicted, the quiet guitar in the background confuses VoiceMachine Processor. The resulting sound is flanging and weird morphings of the guitar under the voice, which does get processed and ends up not quite right, but much less unnatural than the guitar leftovers. This was surprising to me, considering the confused focus of the Processor. Under these conditions the plugin seems to give better results lowering the pitch of the voice than raising it. I suppose it's plausible that someone might be able to use VoiceProcessor on such a "dirty" vocal track if it were to end up buried somewhere in a muddy mix, or perhaps in a non-realistic setting where wacky sound is the goal.

Use it on a "clean" vocal track. . . .

I tracked in some female vocals to test the realism of the plugin using a "clean" source. Overall, I think VoiceMachine will be most effective for those users who want an exaggerated effect. Even using a "clean" vocal track, it was difficult to do any pitch or timbre manipulation that didn't to my ear sound tweaked. Formant shifting, changing the vocal timbre, consistently resulted in processor artifacts related to the upper partials/vocal harmonics. Using only slight setting changes produces effects that are plausibly usable in a realistic context; though I should point out this is not a high-end tuning plugin. Where VoiceMachine Processor will get the most useful results in a realistic context is as a chorus generator tool. The Vibrato section alone, applied to a vocal track, could certainly be used for chorus voices in a mix.

And in case you want ready-made effects. . . .

The included presets demonstrate mostly an exaggerated use of the plugins: Little Kid, Sore Throat, Harmonic Duet, Female Unisono, Sick Tenor, Doomsday Echo, Chainsaw, Busy Bee, and Out of Tune. Some of these effects are right of out the elecronica genre; anyone fond of vocoders will enjoy these presets.

VoiceMachine Generator

So what controls do I get with this one?

According to the Spectral Design manual, "VoiceMachine Generator can be viewed as a 'multiple' version of the VoiceMachine Processor in Key Control mode". You can process your source audio in Generator and produce four MIDI-controlled voices. Each one of these voices has a Vibrato control unit (as described in the previous section about Processor), plus a stereo Balance control (panning left-right), Gain, Formant, and Fine tuning controls for each voice. As with the Processor GUI, there is a main Mix level for processed and unprocessed signal mix, and an additional Humanizer button (hmmm . . . gotta check that out . . . random delay for different voices) as well as the Setup button that allows you to set MIDI controllers for the most of the parameters.

Generator in use

I ran some monophonic analog bass synth tracks from a MAM MB33 MkII (think 303 basslines) through Generator and was impressed by my newly polyphonic synth! Many possibilities here for live synth performance, with MIDI-controlled parameters. Formant shifting also produced some interesting creative results on the synth. The male and female vocal tracks I had processed in Processor produced comparable results in Generator.

At the end of my session with VoiceMachine, what do I think?

For realistic work, the clearer the source, the more precise the result. This is not necessarily the best choice for extremely realistic, high-quality tuning manipulation. For sound design and inventive vocal effects, these plugins offer many possibilities. As a live performance tool, VoiceMachine has amazing potential, as you can automate most of the controls via MIDI. I haven’t tested the plug-in latency in for live manipulation; this might present some limitations, though with pre-recorded tracks it was not problematic.

For the price (list $199) it is an amazing tool for certain realistic and countless experimental sound design effects. Functionally, the hippest stuff it does is 1) instant chorus and 2) MIDI control, which gives you nifty performance as well as studio applications.

For more information about this or any other product on our site, please write us at:
sales@audiomidi.com

Or to be kept informed of all the latest news, reviews, articles, and more, click here to subscribe to the audioMIDI.com newsletter.