6 ms·
Quasi-pop psychology: The human brain is a remarkable machine. It is fantastic at extracting data from any stream of data it can find. It is aided in this by th
by jerf 14y ago
Quasi-pop psychology: The human brain is a remarkable machine. It is fantastic at extracting data from any stream of data it can find. It is aided in this by the neural ability to accommodate and model constant elements of the stream. In the simplest form, you can see this effect with the simple persistence of vision phenomena [1], but the effect can also be applied to fantastically complicated phenomena by our overpowered brains; for instance, can you see your blind spots? (And that's still comparatively simple.) The brain builds a neural net that adapts to the deficiencies of the medium (whatever it may be, movie, audio, etc), streams the data "through" this net, and undoes the medium-induced damage until we are no longer seeing a stuttering 24-fps image, but instead we are seeing John's desparate and heart-tugging last ditch effort to win the affections of Marsha before she finally ties the knot with another guy. And I mean that literally, not merely poetically, the brain is processing the image all the way to its semantic content, which is in some sense the "real" movie.
But the brain, for all its truly amazing capabilities, still has some trivial limitations that any computer would laugh at, and one of them is that the choice of model is not consciously driven, and it can not be changed on a dime. So you go to a 48-fps movie for the first time, and now your brain, sitting in a theatre, smelling the popcorn, watching the corny premovie trivia stuff, flicks in its standard "movie filter", but, alas, oh no, now a 24-fps filter is being fed a 48-fps signal! Now the amplifiers are overamplifying and the net is miscalibrated and what works its way up to the semantic part of the brain is not about John and Marsha but about "Hey! This signal is wrong! The inputs are out of spec! Hey! Hey! The output isn't coming out right! Hey! Hey!" In the audio case, "Hey, after running the automatic sound unwarmers the sound doesn't sound right! Hey!" You "need" the warm distortion of analog audio, you expect it, because you've basically got a hard-coded compensator that won't shut off right away undoing distortion that is no longer present, which is itself a distortion.
The semantic part of the brain, which is still not consciously driven, does not truly understand why it is getting this signal, because it's not like it did anything to create the original net or anything, so it just reports that it is unhappy, and starts casting around for the most likely reason that may be. Not being sufficiently introspective, it decides that rather than its own processing algorithms being miscalibrated for the incoming signal, it decides that the fault must lie in the signal, for which the conscious mind readily provides the seemingly-salient detail that "this is in 48fps", and, bam, explanation achieved. Man Was Not Meant To Know Movies In 48 Frames Per Second. Some people seem to act like this is some sort of Obvious Truth without asking how on Earth such a... specific law could have come to be encoded in our genes, or, yea verily, the laws of the universe itself or something.
Of course, as everyone grows a new neural net to handle the filtering of this new input into semantic content, the lower layers will stop screaming "Hey! Hey!", and the brain, being the magnificent seeker of Information that it is, when it has fully formed 24fps and 48fps nets available to it, will notice that 48fps has a much higher information content, and that it likes it better when given the choice, and indeed may very well decide that 24fps totally sucks.
You will someday stop noticing 48fps. You will someday specifically notice 24 fps and say, wow, how did I ever watch that.
But not today.
If you don't spend much time with 3D, you can observe this effect in yourself usually within the span of one 3D movie. At the beginning, the 3D jumps out at you and you can't help but catch yourself repeatedly "noticing" it. But you'll adjust by the end and not really "notice" it per se; it has faded to the background and you are once again only experiencing the semantic content of the movie. Your neural nets have adjusted. (They're very good at it.) And once you observe the "feeling" of an "unhappy" neural model, you can start noticing it almost any time you're doing something new, new language, new sport, new hobby, almost anything, really.
[1]: http://en.wikipedia.org/wiki/Persistence_of_vision http://en.wikipedia.org/wiki/Persistence_of_vision
- anigbrowl 14y agoHmm, no. You don't necessarily want endlessly increasing information content because that's more work for your brain to filter. As I pointed out below, 48 FPS has been well within our technical capabilities for a long time - far longer than the current polarized-3d technology, for example. Yet there isn't a body of 48 FPS films out there, in contrast to 3d, super-widescreen, movies with vibrating chairs, 'smell-o-vision' and various other technical innovations. TV frame rates (at least in the US) are ~25% faster than film, and have been for years. Standard definition video doesn't have the same resolution of film, but it's still quite viewable on a medium-size projection screen. HD is ~4x better and can be projected on large screens in excellent quality. So why hasn't higher-frame rate video become popular for film (by which I mean moveis, not celluloid)? There's no problem in projecting it (most commercial theaters have a HD projector next to the main one); there's no magic about projecting at 48 FPS, and indeed many affordable pro and semi-pro HD cameras can shoot at 60 FPS progressive. You might as well argue that dance music should approach ever-higher BPMs because increasingly fast tracks incorporate more musical information. Now I do (sometimes) like very fast dance music of 200+ BPM but that's so high above normal human heart rates that it has an extremely limited audience because most people just don't want beats coming at them that fast.
- jerf 14y agoInformation theory is a tricky and subtle thing. A dance song at 300BPM may not actually contain any more "information" than one at 150BPM. And the idea that the brain will get too tired processing that much data explains too much... you just explained why reality is too real for the brain to handle. Also, I'm not claiming More Is Always Better, but in this case we've got a pretty decent bound being put in place by the bandwidth of the eyes and ears themselves, both impressive, but both quite a bit less than our brain convinces us that they are.
- anigbrowl 14y agoIt may not, but then again it may. when I work on a fast tune I certainly don't feel that I can safely reduce the number of musical events therein and have it work well. Reality is of course not too real for the brain to handle, but when we are being entertained we do not necessarily wish work at it in the same way as we do when processing reality - we accept a somewhat limited sensorium in exchange for enhanced semantic complexity. Consider that in things like sports or stage performances (two areas where high-framerate video has proved especially popular) the audience is dealing with a narrowly-tailored field of interest. You're ignoring the elephant in the room: 48 FPS and a variety of other frame rates been affordably available for a long time on both acquisition and playback hardware, both analog and digital. What hasn't it caught on?