I want it: better 3D in films, not superimposed layers

3D films and programmes often look a little like paper theatre sets. Objects look as if they're layered on top of each other, each itself in 2D. When will we get better 3D?

For now, images are analysed so as to pick out the objects in a scene (by comparing contrasts apparently). Imagine there's a character, a tree, a field and a building: often the work stops here, once each has been detected and singled out. The silhouettes are then placed one in front of the other to create a depth of field but the effect isn't always particularly convincing.

Many more scans in fact are necessary and each object should itself be redefined so that our character's nose sticks out properly, or that his/her eyes are properly sunk into the head, their throat is proportionately in relief...

We recently looked into this with a Korean engineer from Samsung who is in charge of on-the-fly 2D to 3D conversion for Samsung TVs. He confirmed that the problem is that current chips being used to scan the 2D image aren't powerful enough to carry out the additional necessary scans. There are better processors on the market but they would take costs up even more! He did however confirm to us that in the future things would be moving in this direction: bigger processors + more detailed scanning = more realistic 3D.

> 3D TV reviews

 

---

> I want it: the keyboard of the future
> I want it: sort out the brightness on 3D TVs

> Buyer's Guides: Our Pick of The Best Products