One of my favorite new products at the massive NAMM musical-instrument show was the DigiTech Vocalist Live, a stompbox that listens to your guitar playing and automatically generates vocal harmonies. There's no complex configuration or hexaphonic hardware; you simply plug a guitar and a mic into the Vocalist's audio inputs and then strum and sing.
The Vocalist just started shipping last week, but O'Reilly Digital Media blogger George "the Fat Man" Sanger has had a prototype since March. He was so impressed that he wrote, "Do you sing and play guitar? You must buy this."
I called the Fat Man at his Texas studio via Skype to hear what he likes so much about the Vocalist Live. He played a bunch of fun examples and revealed what happened when he unleashed this bold new technology in a jam session. (DMI 06-28-2007: 19 minutes 26 seconds)
Control-click to download this MP3 file. You can also subscribe to the Digital Media Insider podcast or add our O'Reilly Network podcast feed to your podcasting application and get the files automatically.
To overcome Skype's dodgy audio quality, the Fat Man and I recorded the interview using a radio-broadcast technique called a "two-ender": We used Skype for communication but simultaneously recorded each side of the conversation locally with better mics in a higher resolution. The Fat Man then sent me his recording as a 160Kbps MP3 file, which I aligned with my side in Ableton Live 6.
He used Steinberg Nuendo to record. I used Ecamm Call Recorder and my Logitech USB headset. I considered using the much better-sounding Rode Podcaster mic I typically use for the show's voiceover, but couldn't get it positioned comfortably on my desktop. I also wanted to record at a higher level than it puts out. Call Recorder captured both sides of the conversation to a QuickTime file, although my side sounded far better, because it was local.
Of course, that's the point of the two-ender technique: The Fat Man's Skype vocal was for reference only. Once I had his recording, I simply dragged it sideways on Live's timeline until it matched up with mine, and then I muted his Skype track.
The Vocalist Live 2 generates two harmony notes for your voice based on what your guitar is playing. It can also compress your voice for smoother dynamics and add reverberation. (Click to enlarge.)
Mr. Fat threw me a challenge, though, by recording his entire vocal through the Vocalist Live 2, bypassing the harmony effect (usually!) when he wasn't demonstrating the stompbox. The result was a dual-mono file, with his reverberant voice on the left channel and his guitar signal on the right. I used BIAS Peak 5 to split it into two mono files I could load into Live and edit independently.
To see what would happen, I dragged the Skype recording I made (blue track) and dual-mono MP3 the Fat Man made (pink track) into Ableton Live. I was pleasantly surprised to see that Live displayed the text markers I'd typed into the recording. But Live was unable to recognize the file's weird multitrack format, so I extracted the two sides with QuickTime Pro and started over.
With previous interviews, I've done the bulk of editing and cleanup in Peak, but in this case, I had three synchronized tracks: my voice, the Fat Man's voice (both straight and harmonized), and his guitar. Because Peak handles only stereo files, I turned to Live instead, using volume envelopes to mute our respective vocals when the other person wasn't talking. (There was an annoying microphone bleed between the tracks.) Another challenge was that the Fat Man's vocal level jumped drastically when he turned on the harmony effect, so I did a lot of sculpting with volume envelopes to even it out. You can see the level difference in the screenshot above.
I then used BIAS's SoundSoap plug-in to remove the hiss from my vocal and the guitar track. To restore some of the spaciousness of the Vocalist's stereo output (which we'd recorded in mono), I added a touch of reverb with Izotope Ozone. (The bulk of the reverb is from the Vocalist.)
Pages: 1, 2