Harmony Voice music app update – VirSyn add Audiobus support

Download from iTunes App Storeharmony voice 2 logoI reviewed VirSyn’s Harmony Voice music app a while back (you can read that review here) and was impressed to see this type of vocal processing – familiar to desktop musicians for some time – starting to appear for iOS. As the app’s name suggests, Harmony Voice produces vocal harmonies based upon your audio input and the pitch of these four-part harmonies can be produced either automatically by the app or triggered via MIDI. And while the app didn’t perhaps compete with the very best that is on offer on the desktop (Antares’ Harmony Engine, for example), with its pocket money price tag, it did a pretty good job.

Harmony Voice in action.

Harmony Voice in action.

I noted two practical limitations in my original review; first, that it was tricky to just output the harmony voices from the app if you wanted to use those harmonies in another app (perhaps a DAW) and, second, the app only worked with a live audio input – it couldn’t take a pre-recorded vocal and process that.

This week, Virsyn have released v.2.0 of Harmony Voice. Graphical adjustments aside (to accommodate the iPhone 5 display), there are two key features in the new update, although these are both pretty fundamental and therefore ought to have a big impact on the app and its operation and, potentially, address the limitations/comments that I made in my original review; Audiobus support and a complete reworking of the underlying harmony algorithm.

On the Audiobus

Harmony Voice being used with JamUp Pro and Cubasis via Audiobus.

Harmony Voice being used with JamUp Pro and Cubasis via Audiobus.

Harmony Voice can now be used in any of the Audiobus input, effect or output slots. The most obvious choice for me would be as an effect slot processor. In this position, you can feed either a live audio signal or audio from another app into Harmony Voice and then ‘capture’ the harmonies produced in your DAW running in the Audiobus output slot.

This worked a treat. I was able to generate harmonies from a live vocal input but the fun isn’t limited to vocals as Harmony Voice also did the business with a melody guitar line. In this case, I had JamUp Pro in the input slot and, using a mixture of Harmony Voice’s soprano and alto voices, I was able to created a nice harmony guitar effect and capture the end results in Cubasis (sat in the Audiobus output slot).

Processing a pre-recorded audio signal requires an extra step but is still pretty straightforward. While I suspect there may be other ways to configure this, I found the easiest thing to do was open Cubasis in the Audiobus input slot as well as the Output slot, solo the track I wanted to process through Harmony Voice in the effect slot and then capture the processed signal back to a new track within Cubasis. Once I’d made the new recording, before playing it back, I then had to remove the Cubasis instance from the Audiobus input slot (otherwise, the new track gets sent back through Harmony Voice and harmonies are generated on top of the harmonies – which obviously gets a bit messy).

I was able to capture harmony Voice's output within Cubasis and even route an existing track from Cubasis through Harmony Voice and back to Cubasis - very neat.

I was able to capture Harmony Voice’s output within Cubasis and even route an existing track from Cubasis through Harmony Voice and back to Cubasis – very neat.

In essence, this routing allows you to use Harmony Voice as a sort of ‘send’ effect loop within your DAW. It requires a bit of messing about with switching apps in and out of Audiobus but it does seem to work – very cool. The most obvious application here is so that you can generate automatic vocal harmony parts based upon your recorded lead vocal. The ability to do this from a pre-recorded vocal – rather than just a live audio source – means that you can experiment with the harmony generation in a more controlled fashion to get the best result possible. From a workflow perspective, this is a big help.

Overall, I had no problems with the Audiobus implementation in Harmony Voice. It seems solidly built and it means VisSyn’s app is now much easier to integrate into a recording workflow.

Perfect harmony?

In experimenting with the Audiobus support and the options it opens up, I have to say I was also struck by the quality of the harmony generation. While most of my listening was done working with a guitar input (audio artefacts are perhaps less obvious with an overdriven guitar than with a pristine voice?), used with care and some musical intelligence (a bit of music theory helps even though Harmony Voice can generate harmonies in an totally automatic fashion), the results can be very good.

As well as up to four part harmonies, Harmony Voice includes a useful effects section.

As well as up to four part harmonies, Harmony Voice includes a useful effects section.

We are all so attuned to the sound of the human voice that anything ‘odd’ about its sound is immediately apparent. When applied to vocals rather than an instrument, a little more care is needed. In truth, this is still also true of even the best auto harmony generation available in the desktop computer environment – push it too hard or expect miracles and you will be soon be given a very audible reality check – and Harmony Voice is not different in that respect. That said, v.2.0 sounds like a significant step forward to me over the earlier version.

In summary

VirSyn have produced an excellent update to Harmony Voice. The Audiobus support is a big plus and means the app moves from being ‘interesting and with potential’ into a practical and useful tool for the iOS music producer. The quality of the harmony generation also seems to have been improved.

Harmonies can also be controlled via MIDI notes using the virtual keyboard or a hardware keyboard connected to your iPad.

Harmonies can also be controlled via MIDI notes using the virtual keyboard or a hardware keyboard connected to your iPad.

If you have ever used an ‘auto’ harmony product before you will know what I mean when I say the Harmony Voice is not a ‘set and forget’ effect. This is not a criticism of the app itself or VirSyn; it’s simply a realistic statement that reflects just how complex the audio processing task is if you are trying to automatically generate believable harmonies of something as intricate as the human voice. In audio processing terms, this kind of algorithm is about as tough as it gets.

Harmony Voice takes a pretty good crack at it. Perfect? No, but then none of these sorts of tools – desktop or mobile – currently are. Getting the best from this kind of processing requires a realistic acknowledgment of what is possible, a touch of music theory, and some due care and attention to your input signals.

All that said, Harmony Voice, with its Audiobus support, is now a very useable tool for those building an iOS-based recording system. And at £6.99 (or the equivalent $/€ price), it both a bargain and a lot of fun. I’m already looking forward to the next update :-)

Harmony Voice


Be Sociable; share this post....

    Speak Your Mind

    *