Review: Mu Technologies Mu Voice 1.2 (Mac/Win)

Sep 1, 2008 12:00 PM, By David E. Weiss

A GREAT-SOUNDING PLUG-IN FOR INSTANT HARMONY

Those hoping for new developments in the vocal-processing market have reason to rejoice, for a new voice has entered the choir. Mu Voice 1.2 is a plug-in that tunes and harmonizes your vocal tracks. It is the flagship product of Belgium-based Mu Technologies and is delivered in VST and RTAS formats for the Mac and PC and AU format for the Mac.

Mu Voice uses a proprietary algorithm recently developed at Antwerp University in Belgium. This algorithm is distinctive in that it analyzes audio in extremely small units, which reduces latency and gives higher resolution in the time domain. Because there is a lot of variation in pitch in every note that even the most accomplished singer produces, large analysis windows create artifacts, called smearing, when the same frequency is assumed over too long a time. This is analogous to the way that a high-resolution digital photograph shows fewer artifacts than a low-resolution photograph because the low-resolution photograph uses fewer pixels, which smears the same color over a larger area.

Lift Up Your Voice

Mu Voice is easy to use, and I found it fun, inspiring, and instructive — fun because it's immediate, inspiring because the harmony voices are very realistic, and instructive because when you slide off-key, so do your virtual backup singers, greatly enhancing the tiniest error. Enable it on a vocal track, pick up a mic, and choose a preset (I recommend starting with a trio), and you'll hear several clones of yourself singing along in harmony in different ranges (see Web Clips 1 and 2). You can try this out in the free, time-limited version available from the Mu Technologies Web site.

As with any vocal-processing software, Mu Voice performs best when it's processing a clean signal, so put it first in the chain if possible, and certainly in front of any reverb or delay. Occasionally the harmonizing voices were garbled beyond recognition, but clicking on the Inversion button in the Analysis settings quickly fixed that.

Mu Voice has a 4-channel mixer with which to control the source voice and three harmony voices (see Fig. 1). By default, the second voice sings a third above the first, the third voice sings a fifth, and the last voice sings an octave. But it's easy to change the pitch of any voice within up to two octaves above or below the default pitch. With a bit more tweaking, you can change the pitch to anything you desire (more on that later).

Each channel strip has Volume, Pan, and Mute controls, as well as a Humanize slider to introduce subtle variations in pitch and vibrato and a Formants slider to adjust the timbre and character of a voice. Two drop-down menus (Filter and H.EQ) offer a variety of effects. Mu Voice's Pan slider delivers a bit more depth than left-right balancing; the manual says that the slider's parameters are based on measurements relative to the listener's nose, and to my ears it has a full, three-dimensional quality.

The Formants slider sets the contour of frequencies that characterize the timbre of a voice. You can tweak this curve without affecting a note's pitch. Moving the Formants slider to the left expands the formant spectrum, producing a more nasal, squeezed-down effect (similar to the e in ear). Moving the slider to the right narrows the formant spectrum, producing a more open sound (similar to the o in olive). The middle position preserves the formant spectrum of the lead voice.

I found Mu Voice's controls very useful for smoothing out harmony parts. As noted in the review of Antares Audio Technologies Harmony Engine and Zplane.Development Vielklang (see the April 2008 issue of EM, available at emusician.com), synthetic harmonies can sound a little unnatural when soloed. Though harmonization software is not intended for this, soloing the parts helps you make subtle improvements (see Web Clips 3 and 4).

Strike a Chord

Beyond creating harmony voices at regular intervals, you can make Mu Voice follow specific chords: major, minor, diminished, sixths, sevenths, ninths, and suspended fourths. You use the two Chord menus to define chords for the source melody. One cycles through the available key signatures, and the other lets you choose a chord from your chosen signature. I found that awkward and a bit mind-numbing. Fortunately, Mu Voice also lets you select the key from your MIDI keyboard. Better yet, play a chord and Mu Voice will capture it for you.

Once you have the chords, you play the piece through in Write mode, indicating chord changes with your mouse or your MIDI keyboard. Then, in Read mode, Mu Voice will dutifully follow the song, moving from chord to chord at the appointed place.

Calling the Shots

The chord scheme is where Mu Voice gets interesting: a cell does more in Mu Voice than define the key and chord for a certain passage; it is a snapshot of all of Mu Voice's settings at that point. For example, if you've designated that a certain passage be linked with a C major chord, you can also tell Mu Voice to apply a certain preset. (A preset defines a wide variety of characteristics, from the presence or absence of an EQ filter to how many voices are singing and in which octaves.) Mu Voice comes with 36 presets, which you can overwrite without erasing the defaults, and it lets you save an additional 24 of your own. Projects, which contain both Chord Schemes and settings, are saved as XML files, and they are compatible with any version of Mu Voice on any platform.