Phi and The Human Face

I’ve long harboured a suspicion that with art (and probably nature, but especially art) there’s about a 3% margin of error when it comes to things “said to conform to the Golden Mean” actually conforming… they’re close, but often fall within a 3% range rather than being spot-on.

I’ve been trying to put together a new movie explaining what the calipers are for recently… and one of the things that I wanted to talk about was how the Golden Mean can be useful when forming the dimensions of the human face – because a lot of the calipers are bought by cosmetic surgeons.

So… I thought I’d show the proportions on a face…

Obviously I didn’t have a face to hand :), so I made one. I did this by averaging two other faces that I found on the internet (hint, one is a Brazilian Supermodel). Averaging faces is interesting because for some reason you wind up with something incredibly beautiful – which becomes more soft-focused and angelic, the more the average is applied.

I didn’t want the picture to look toooo unrealistic, so I only used two faces. For what it’s worth, Adriana Lima is (so she claims) a Swiss, African, Indian and possibly Japanese mix – so maybe the face-mixing thing is operating at a genetic level as well.

Anyway, when I superimposed golden mean measurements over it, instead of being 3% they were… spot on… a pixel here or there either way.

Amazing.

There’s an idealised face-mask that is also (apparently) based on the Golden Mean… which I think originates here… I’m not sure how alive that website is. Anyway… this is the mask

Which seems to work pretty well – although I think the fashion is for bigger lips these days – which is handy for Adriana. I also think the fashion is for people to look elfier… if that makes sense. Maybe it’s just neotony. The scourge of the age. However… this is how it pans out on the face I made…

I’m not sure about that mask though – it reminds me of G’Kar of Babylon 5

Even though now I come to look at them side by side, the similarities are tenuous. Still reminds me of it a bit though.

Joking aside… this has really surprised me. Not that I didn’t think there was something to it – I just didn’t expect it to be this accurate – a mathematical averaging of faces produces a result that follows the lines very closely.

Which begs the question I suppose – is this something that we’re sexually (natural) selecting for, or is it somehow encoded at a genetic level?

I think it’s the latter… I think that the equivalent of the Arithmetic/Logic Unit of DNA somehow symbolically represents numbers so Phi is an easy measurement for it to make. Computers use base-2… humans extrapolate that up to base-16 (hexadecimal) to make it easier for us to handle… but natively we use base-10. I don’t think DNA uses numeric symbolic representation at all… it uses something else – geometric maybe. Phi is a very complicated number to represent numerically, but very simple (like Pi) geometrically.

Phi is one of those numbers that arises naturally from simple geometry – representing it numerically isn’t being true to it’s nature. It’s like the difference between analogue and digital. Perfection is an analogue condition.