VirSyn have announced one of the nicest effects I’ve seen yet on a mobile platform in the form of iVoxel, a vocoder for iPhone/iPod touch and iPad. On the handhelds, it looks like something you could easily hold up to your mouth and rest atop a synth; on iPad, you get a nicely-sized touchable keyboard for control. And they’re clearly hinting at more work to come.

http://www.virsyn.net/mobileapp/

iVoxel isn’t available yet, but it is coming. So, how does this compare to your laptop? You can get VirSyn’s full-blown MATRIX 2 vocoder for EUR119 on sale now, run it on a computer, and get all the advantages of hosting in your existing production software along with more advanced features (which you have to give up on the iPad). That laptop rig is the same one Kraftwerk used, in conjunction with Cubase 4 and a Sony laptop. iVoxel uses the same, superb sound engine – an impressive feat. My guess is – irrespective of iOS, Android, or whatever OS – these devices will really excel once they’re able to provide the same sorts of workflows that the laptop do. The appeal of touch input and the form factor is clear.

Open source pitch correction?

Poor man’s Android AutoTune, anyone?

MicDroid by Ethan Chen is a pitch-correction app for Android. Aside from enabling I Am T-Pain-style effects without iOS, pitch matching could lead to lots of other creative applications – pitch to MIDI, vocally-controlled synths, and more. The code is mostly available under an MIT license.

You can also grab MicDroid from the Android Market, with features like a shiny UI, sharing, recording to SMS, and encoding – and it’s still only in “alpha” state.

The underlying library on MicDroid is called AutoTalent by Tom Baran, and if you’re not terribly (or exclusively) interested in Android, that library is available under GPL2 on any platform. That gives you the ability to create all sorts of creative voice input applications on various platforms. Check out both from GitHub, and if you’ve got an Android, you can give MicDroid a try whether you’re a developer or not:

MicDroid @ github

http://web.mit.edu/tbaran/www/autotalent.html [MIT research page]

AutoTalent free plug-in downloads for LADSPA, Mac AU and VST, and Windows VST – stability may vary

Tom describes Autotalent “as the result of a week of recreational signal processing.” And yes, if you’re a PhD candidate at MIT’s DSP group, “recreational” and “signal processing” can actually go together.

I’m excited to mess with these plug-ins, though, because to me the generic pop music effects you hear all the time are just the tip of the iceberg. Experimental vocals, ho!

Updated: Via comments, Paul Davis points us to a more up-to-date “fork” of AutoTalent, Talented Hack. It features improved detection, MIDI input (for greater control of pitch correction), MIDI output (so you can use this as a pitch-to-MIDI plug), and generally cleaned-up code. It also runs as an LV2 plug-in, which is far more convenient on Linux. But even just having cleaner code alone means it’s probably a smart candidate for swapping into the Android tool (cough, cough).

The Internet, having satisfied itself yesterday with video that faked a Beyonce who couldn’t sing, now imagines news that can. And Steve Reich is proven ahead of his time — again. (Congrats on the Pullitzer – it took them just five decades to notice!)

Yes, Antares’ Auto-Tune plug-in – now so ubiquitous in mainstream, non-audio-engineer knowledge that it’s become a generic description like “Kleenex” – can be applied to everything. (We, um, can only hope these industrious YouTubers are using legally-licensed copies – that is, until Antares releases a 99-cent iPhone app.) And so, hilariously, we imagine a world of news sung hip-hop style.

As it happens, this digital foolery does reveal something deeper. One of the joys of language in general, certainly true of English, is the degree to which musical-like inflection turns our spoken words into songs. In English, these inflections are more decorative than syntactical – good news, as unlike a language like Mandarin, the wrong inflection won’t get you in trouble. But I think a lot of the texture of the music of English-speakers – native and non-native alike – is influenced by the rhythms and melodic contours of our speech. Would Jazz have happened in a country without American English and its regional dialects? Given the sounds of “talking” trumpet mutes, my guess is it would have sounded quite different.

Poor video, but gives you the idea (where’s the official Steve Reich YouTube channel?):

The Auto-Tune News is intentionally silly, of course. But even without digital aids, people have been finding songs in recorded speech. Take composer Steve Reich: without the aid of Auto-Tune, he found surprisingly in-tune sounding melodic fragments in interview recordings for pieces like Different Trains, and later built an entire opera around the technique. (The Cave, with its accompanying video, below.)

Antares, for their part, is keeping a good sense of humor about all of this – and laughing all the way to the bank. There news stream has followed the pop culture references to their product, and even jokingly suggested they would introduce Direct Mind Access Composition Technology on April Fool’s Day. (Don’t laugh too much: I heard composer Jon Appleton, sitting alongside fellow luminaries Bob Moog, Laurie Spiegel, Morton Subotnik, and others, suggest a musical brain hat at a panel on the future of music. I’m happy to actually shut down my mind occasionally, so I don’t entirely understand the appeal.)

Previously:

AutoTune: The Song, a $99 Version (Hide!), and Some History

And here’s part I of Auto-Tuning the news. Daily Show, eat your heart out:

Thanks to decrepitude in comments. (Erm – that being the dude/dudette’s alias, not actual decrepitude in comments. That we’re not thankful for.)

Updated, for all time:

Readers are nearly 100% for judging this one. It was a fake. And the site with a really stupid name (hellohomo??) admits that it was faux.

Howard Stern Hoaxed! Beyoncé "Outtakes" Are Fake, Creator Admits [E! Online]

Wow, that may be the last time CDM links to E!

Lesson learned: yes, the Internet has the power to spread rumors at new speeds. It can also debunk them even faster. That’s something to pass along to the “get off my lawn!” crowd.

Okay, sound engineers and audiophile experts out there: it’s time to play “is this YouTube video real?” (And, heck, even if it’s not, it’s oddly hilarious.)

A recent episode of Howard Stern’s Sirius radio show claimed to expose a “leaked” feed of the raw vocals for Beyonce Knowles as she sang live on the Today Show. I could try to describe just what they sound like, but it’s really best to hear for yourself.

So, what’s the deal? You can hear the in-tune vocals in the background, as though they actually are bleeding into the mic. Does Beyonce have access to some super-secret, military grade version of Auto-Tune? Is the difference between her processed voice and her actual voice as comically radical as depicted in the ‘N Sync episode of The Simpsons?

I mean, I know – YouTube? Blogs? Howard Stern? The Internet? How much more credibility could you possibly ask for?

I’ll let you be the judge. For the record, the Today Show audience actually heard this.

Found via the brilliantly-domained Yes But No But Yes and Matt Ganucheau. YBNBY or whatever you want to call it describes the singing talentes here as “a parrot being sawed in half.” Hatebeak, an actual parrot trying to sound as though he’s being sawed in half and unofficial mascot of CDM, I’m sure you’re jealous.

Previously: the Great Van Halen Detuned Keyboard Incident.

Oh yeah, and you may want to download this video before someone has it removed.

And yes, hey, if it’s just a fake, it goes nicely with the "shreds" video meme. So bring it on.

Updated: Consensus is that it’s a beautifully-executed fake. (I’m inclined to agree – sorry, Mr. Stern.) And to think, we thought all this time the advantage of tools like Melodyne and Auto-Tune would be taking out-of-tune things and making them in-tune – ignoring the expressive potential of doing the reverse.

Britney, meanwhile? That’s another story.

Now, what I’m curious about:

Who faked this?

How did they do it?

Speculation?

This week, the Web is buzzing over the music video of AutoTune, the (parody) song.

Sadly, this video could have been so much more – not even so much as a Cher reference, really? (Cher’s producers: AutoTuning way before Kanye West, and then lying about it! Brilliant!)

For a bit of AutoTune reflection and history:

Read the 1999 Sound on Sound article in which the producers tried to fool people into thinking they used a Digitech Talker vocoder, which, come to think of it, sounds like it would have actually been a pretty decent idea, anyway. That story is now updated with the correction. I’m sure the producers are relatively sorry about it certain they can’t get away with it any more / it’s hardly a trade secret.

Sasha Frere-Jones wrote a thoughtful article on AutoTune for The New Yorker earlier this year. Best bit:

Someone once asked Hildebrand if Auto-Tune was evil. He responded, “Well, my wife wears makeup. Is that evil?” Evil may be overstating the case, but makeup is an apt analogy: there is nothing natural about recorded music.

That much is true. Of course, it begs the question: does his wife smear lipstick randomly over her forehead? Can you actually see her face? You see my point.

Perhaps feeling the pressure of free tuning and vocal plug-ins now shipping with many audio apps and DAWs, Antares have introduced Auto-Tune efx, a US$99, simplified version of the plug-in for Mac and Windows now available exclusively at Guitar Center. Oddly, a selling point is that it currently comes with a free iLok; given that it’s targeted at beginning users who likely would be shocked that they have to pay extra to use DRM added to a program, that seems like not something one would advertise. (Wow! Thanks!)

In Antares’ defense, though, no, I don’t think AutoTune is evil. In fact, I think ironically, it’s drawn attention to some of the potential fictions of recording – and, through the magic of reverse psychology, made a great case for making changes to the actual vocals and using the computer for more creative tasks rather than seeing it as a panacea for fixing human beings.

Antares also does produce software that can be used to creative effect, like the AVOX2 toolkit and its mutating effects.

To me, the most interesting (and overlooked) thing about AutoTune is its roots in seismology and geophysical data. Yep, that’s right: founder Andy Hildebrand got his start at Exxon doing things like looking for failure points in pipelines. He went on to study composition at Rice’s Shepherd School of Music, and used his smarts in seismology to solve musical problems.

For more on that history, read the 1999 awards citations in the newsletter of the Society of Exploration Geophysicists [PDF], recognizing Hildebrand. (I love search engines.)

So, knock AutoTune if you like: what it demonstrates is the flexibility of digital algorithms. In fact, the beauty of computers is that they don’t worry about issues like taste or the difference between music and underground oil. And that means you can take a tool and apply it to a radically different job – giving us human beings near endless potential in how we interpret digital tools.

And that suggests that you ought to be able to use AutoTune and your voice and do something that isn’t awful at all.