Have you ever experienced the uncanny valley with VFX? But more importantly, have you experienced it with sound? Or do we experience every day working in post, with every sound? And it is just an issue of plausability/sustaining/reinforcing suspension of disbelief? More info here: http://en.wikipedia.org/wiki/Uncanny_valley
I find that my own suspension of disbelief tends to arise from synthesized sounds, as opposed to recontextualized field recordings. This is true of a lot of genre films (Escape from New York, Solaris, and untold hundreds of other horror and sci-fi films), but even recent films such as Terminator Salvation, where the sound design of some of the 'bots was so synthetic sounding that I found that it pulled me out of the moment and narrative. Some films where the sound is just super weird, like Jacques Tati, you can somehow just roll with and accept the oddities. But when a film's played mostly for realism, that's when it hits ya.
I think people experience something like that when they hear themselves talk. Your brain knows what you're hearing is the truth, yet what you're hearing does not
quite sound like match the real you (from inside the skull that is). By then the brain simply refuses to accept the artefact based on this distinction malfunctioning.
I think with sound that has little to do with us, humans, we're happy to go for the abstract and exercise interpretation and association. But take a speech synthesizer, my, those things sound ugly mostly due to how advanced they are nowadays, and how wrong they get the intonation.
This, in my view, is much more closely bound to sight than it is to hearing. We're largely used to the idea that one thing can sound like another (maybe as long as it isn't us). But to look very much like another, that's largely the cause of the phenomenon....... Maybe if we attempted to bark at a dog, the dog would have some sort of similar feeling?
The first thing that popped into my mind was a scene from Twin Peaks I watched recently. The character The Man from Another Place (I think that was his name) was speaking but every word he uttered was reversed and twisted and quite unnatural. At first hearing it shocked me, then it got progressively more and more spooky and hard to deal with.
When doing research on how horror games utilize sound, I came across this conference paper by Mark Grimshaw titled 'The audio Uncanny Valley: Sound, fear and the horror game': http://vbn.aau.dk/ws/files/61573698/audioUncannyValley_MG.pdf
An interesting read.
Well, with CGI I'd say it's far from rare, though of course there are a lot of really convincing CGI's too! Just look at Gollum! But with sound, I'd say you have something there with the suspension of disbelief
This here is NOT a rant about professionalism, it's about my opinion on keeping withing diegesis and nothing else, ie giving a natural impression of the final sound but not the source contra a more exaggerated and expressive sound:
With sound effects and ambiances, and this is a philosophy I'm convinced of right now and work by, I'd say the more one can maintain a full transparency and a natural feel in the soundtrack (not to be confused with a truly natural sound/kitchen sink realism...not a fan of that...), excluding the music and non-diegetics room-wise of course but not in characteristics, the less the risk might be that the audience will get distracted and breaking the illusion. Wall-E, Dark Knight Rises, Delicatessen and Bourne Supremacy is in my opinion great examples of that. But frankly, and I guess most people here in all likelihood have very much seen the effect of this regularly, sound is rarely even remotely as prioritized as picture in most movies, making it difficult to always take it that extra distance we really wanted. And I dare say people are pretty used to that, making small errors and omissions in insignificant sounds more forgivable and out of the valley. I haven't read the article from cover to cover yet though I think it's very interesting, I'm just too tired right now, but it seems to focus on body-language, and that's pretty much my point. We're heavily trained from birth to read and interpret even the tiniest nuances of an individual, be it intonation of the voice, to the way someone wrinkles his or her nose, to how healthy-looking the skin is, most of us are extremely susceptible to even very subliminal changes.
With everything else on the other hand, we can be fooled very easily. Take for example the shotgun in Terminator 2. I've shot 12 gauge both pellets and slugs many times, and the first time all I could think was "What the deuce!? It sounds more like a really loud Champagne-pop than a gun!?"...frankly, I've even sweetened a Champagne-pop with said shotgun. Yet, even though I know full well that it's fake, I even know what it's made of, and I actually work with these things myself, still I feel in my heart that that cool blast from the humongous cyborg is real! As a contrast: "realistic" speech-synthesis gives me the creeps. I very much prefer the Amiga/Atari/Hawking's kind or none at all...
Sound design-wise I do however often strive for the Uncanny Valley when doing horror and dystopian movies. As it's much easier to accept sound effects as they are, and one must find what makes the individual sound fit even though sounding unnatural, it's not an easy thing to pull off. But as long as the scene doesn't depend on it, I consider a successful application of it a sweet bonus, and a failure not much more than a day-long dent in my self esteem :-)
Of course one might argue if there really are any clear humanlike emotions readable from un-antropomorphic stuff, but therein lies the challenge :-)
I don't think there's any sound equivalent yet, because the curve from positive to negative is so specific:
"Mori's original hypothesis states that as the appearance of a robot is made more human, a human observer's emotional response to the robot will become increasingly positive and empathic, until a point is reached beyond which the response quickly becomes that of strong revulsion"
It seems like the effect is specific to something mimicking us, so perhaps the obvious audio equivalent for this is speech synthesis. I imagine that we're not so far from creating basic systems using voice recognition, basic AI and speech synthesis for roles such as handling queries in call centres or other similar tasks. When these become so realistic that you're really not sure whether it's a person or a machine on the other end, that could be very strange indeed, and possibly incite the same revulsion that characterises the uncanny valley.