1. Autotune My Heart

    I’m at the ESOMAR Congress in Atlanta where I’ve seen some good presentations and some less good ones. One of the good ones was by Rana El Kallouby of Affectiva, talking the audience through Affectiva’s emotional recognition technology and its application at scale on mobiles. Not having played with Affectiva output I don’t know how useful it all is but the tech is developing fast and it’s easy to believe that accurate, nuanced reading of facial emotion is just around the corner, even if you don’t think it’s already here.

    What struck me about the presentation is that I’ve got emotion-testing tech wrong, in terms of its use. I’ve been thinking about it as a way to get an overall read on the emotions a piece of content - an ad, most commonly - elicits. But as the presentation pointed out its real value is in easy optimisation of ads, split-testing of individual moments to create maximum emotional effectiveness (and I loved the work on repeat viewings of adds to measure the tail-off of that effectiveness, too). One client tweeted how effective he’d found Affectiva for making “frame-by-frame” changes to creative work.

    This kind of talk might well send a shudder through creatives, though. It’s not as if traditional ad testing methods haven’t been rough on them, constantly railroading their work in search of clunky brand linkage and crassly “persuasive” messaging. Now, just as emotion starts to move to the centre of the testing process, along come tools which allow intense client micromanagement of emotion.

    If you believe that the holistic emotional take-out of an ad is more important than the second-by-second reaction you might also roll your eyes at this. But these things are testable and will be tested. And besides, doesn’t it follow that tweaking the individual moments in an ad to maximise emotion will lead to a better experience overall?

    Here’s where analogy bells started ringing deep within my rock critic lizard brain. I’ve come across second-by-second optimisation before, after all. It’s what pitch correction and autotune techniques are built on, and it’s also what lies at the root of the “loudness wars” - the advance of compression techniques to make radio hits seem ever louder and brasher, to the point where if you look at a waveform of a current song (in almost any genre) you’ll see everything maxed out, redlined to a point where no subtlety is allowed.

    Are the loudness wars a good analogy to emotional micromanagement? The desire to achieve cut-through in a noisy marketplace is the same, after all. But perhaps autotune is a better one - its become shorthand for a variety of pitch correction techniques designed to iron out flaws in recorded music and allow singers to deliver perfect performances on record (and increasingly live) every time. It’s easy to map such idealisation onto advertising creative: imagine a world where every emotion was perfectly pitched, every gag exquisitely timed, every musical cue or cut dropped to precision for maximum emotional effect.

    The only question is, would people notice? With loudness comes flatness, with autotune comes suspicion of fakery. The emotional palette of adland is often already very familiar to audiences - uplifting music, quirky casting, a series of real-life interactions (flirting, banter, parent-child relationships) rendered exhaustingly familiar by constant use and predictable outcomes. Would emotional optimisation accelerate that exhaustion, airbrushing emotion to create an onscreen world that’s subtly alienating rather than richly recognisable?

    But there’s another, more positive way to look at it. When voice manipulation effects are easily noticed it’s because their users - from Cher to T-Pain - want you to notice it. They’re deliberately accentuating the artificial, drawing your attention to their tweaks and manipulations. Looked at in this light, emotional optimisation tools become rather more exciting - giving creatives the opportunity to engineer wholly unnatural emotional effects: maxed out peaks of happiness, switchback oscillations between, say, surprise and disgust from one second to the other. Whether anyone would use Affectiva’s fascinating technology in this way I have no idea, but for me it’s a more tantalising prospect than micromanaged, idealised, emotional work.

  1. davebloom likes this
  2. minimoonstar likes this
  3. cureforbedbugs reblogged this from blackbeardblog and added:
    Here is an excellent post that requires at least 100% more Farrah Abraham. Wait, what’s times 100?
  4. hautepop likes this
  5. hardcorefornerds likes this
  6. blackbeardblog posted this