| 
| Posted: Thursday
1st of July 2010
at 9:58:54 PM | 
|
|
| 
| 
| 
| 
|  his is a topic which has been bugging me since they started doing 3D at the cinema and has really started to annoy me since they started trying to push for 3D at home...
Firstly lets start with an overview of 3D technology:
There have been several types of 3D through the ages, starting back in the 1800's with stereoscopic viewers:
Stereoscopic: A frame is used to hold a stereo image (two paired images taken from slightly different perspectives). Each eye looks at a a different image and the effect of depth is formed. Probably the most common example of this type of 3D are those red binocular-like ViewMaster picture viewers with the circular picture discs that we all had when we were kids.
Anaglyph: A pair of stereo images (monochrome) are printed on top of each other in two chromatically opposed colours (red and green or blue) and then paper glasses with coloured filters are worn so that you can only see one of the images through each eye. You can't really have colour images using this system.
Lenticular: This method involves taking our stereo images and cutting them into thin vertical strips. You then create a new image by taking one strip from each of the separate images in turn. (The first slice is slice one from picture one. Then slice two from picture two. Three from one. Four from two. etc...) Finally, you place a sheet of plastic over the image. This sheet of plastic has ridges in it which act like lenses and mean that, from a given angle, you will only see one of the images at once. If done correctly then each eye will be looking at a different image and you have your 3D again. (It is also possible to embed more than two images using this method). The main problem with this approach which is that there are certain 'sweet spots' and when not positioned correctly you can see both images or only part of one.
One of the most recent (non-3D) application of this technoogy has been in cars so that you can have a screen in the center of the dashboard and while the driver is seeing the navigation computer, the passenger can watch a film.
Random Dot Stereogram: This is really no different to the old stereoscopic vision. You have two separate images that appear to be nothing more than noise, but when viewed as a stereo pair you see depth in the image. The easiest way to explain this one is to imagine a floating dolphin. Now cover both it and the background with a random pattern. If you look at it through one eye (and remain perfectly stationary) then the dolphin would probably vanish into the background, but with both eyes, the depth of the object defines it for you.
Magic Eye (Single Image Random Dot Stereograms): Have you ever looked at patterned wallpaper and crossed your eyes so that it appears to be further away? That is how Magic Eye images work except that the pattern is much smaller and the intervals between the parts of the pattern change in order to alter the depth at which you see something. You should be able to see the effect with the following text examples. Crossing your eyes should make the following asterisks pop out of (or into) the screen.
* * * * * *
* * * * * * *
* * * * * * * *
* * * * * * *
* * * * * *
* * * * * * * *
* * * * * * * *
* * * * * * * *
* * * * * * * *
* * * * * * * *
More recently, it has been possible to create a 3D image by projecting two polarized pictures on top of each other and then using polarized glasses to separate each image out to the correct eye. (This is another variation on stereoscopic vision.) Everyone will be familiar with the push for 3D now available in cinemas. The most common implementation of this technology is to use an LCD screen to switch the polarization of alternate frames coming from a single projector. A silver viewing surface is used to maintain the polarization once it has been reflected off the screen. (Circular polarization is now used over earlier implementations using vertical or horizontal polarization so that you don't get strange ghosting or darkening effects when you tilt your head.)
Another stereoscopic effect recently introduced uses something called a Parallax Barrier to get an effect very much like the Lenticular method outlined above. (This is how the new Nintendo 3DS works.) As long as your head is positioned in the 'sweet spot' your left eye will see only the even numbered pixels across the screen and your right eye will see only the odd.
OK. That's all well and good but why am I writing all this? Am I just showing off what I've found out about 3D?
NO!
...
Maybe a little.
...
Mostly I just think that there are too many problems with 3D for them to be forcing it down our throats so hard. They have to be working on a next generation of 3D, in which case early adopters are going to get boned hard! The main problem I have is that it's designed to be watched sat up straight with your head upright. This kind of works in a cinema where you're generally trying not to encroach on the personal space of the person sat next to you. If I'm watching a film at home then I actually want to collapse and most of the time I'm actually laying on my side. (One of the advantages of being single: Not having to share the sofa.)
This may not seem like an issue. The current glasses ensure that the left image always goes to the left eye and the right image always to the right eye, but my problem is this: As stated before, the two images you are shown are filmed from the upright position and you will always see them that way, even when you head is tilted. Unfortunately the images you should see when your head is tilted are different from those when upright. The more you tilt your head, the more "wrong" the images are. This is a rough example. Here is an object:
This is how that object would appear to me if I looked at it, and if I saw it through a 3D TV:
All good! Now if I am in my normal viewing position it looks like this:
That's just not right! How about if I'm feeling mad:
Now the object is kind of back to front...
The differences may not seem like much but they are enough to screw with your head. Next time you're at a 3D film, try tilting your head 90 degrees and see what you think. I know some people will probably still be able to see the 3D effect with their head tilted, but in this situation, your brain is doing a lot of compensating for the inconsistencies in what you are seeing. (The brain is very good at doing stuff like this.) I can't imagine you'd be able to do it for too long though before it gave you a serious headache.
There is also another wackyness associated with 3D films and 3D TV (and this is why all 3D comes with an age warning): We spend years learning how to focus and see things. The brain does an amazing job comparing the two pictures it gets in order to give us depth information. We know that if you want to look at something close you need to cross your eyes and focus close. For things at a distance, our eyes go straight and we focus far away.
With 3D we have to override these automatic behaviours and always focus at the screen distance no matter how close or far away our brain is telling us that we are looking.
Which leads me onto another point:
Even if you do buy into 3D, nobody knows how to use it properly yet! For example, setting the focus in a shot so that the main characters are sharp but the things in the background and foreground are blurry. This is great for a normal film as it gives the viewer a simple cue as to where they should be looking. But in 3D when your eyes are already doing things that are abnormal. Putting a blurry object up close to the viewer is just distracting. A couple of times in 3D films now I've found myself straining to bring a blurry foreground object into focus when I should be looking elsewhere. This seems to be especially bad in films which were 2D and then converted. Clash of the Titans was bloody awful for this. That film gave me a headache. There were even sections where they had black flickering around forground objects where you should have been able to see a little further around an object but they didn't have the footage to put back there!
I should also point out another of the limitations of 3D which is that you are fairly restricted in how far out of the screen an object can pop. (Don't believe the BS that the marketting folks tell you with things flying out fot he screen and around the room!)
You need to see two copies of an item for it to appear correctly (one for each eye), that means that very close objects can only appear steadily in the center of the screen (with your left eye looking to the far right and your right eye looking to the far left). The further left or right your object moves, the less it can comfortably float forward.
In other words: Film makers need to stop doing fast panning shots in 3D! It is very difficult to see close objects travelling past the camera as you only have a moment (when the objects cross the center of the screen) to actually let your eyes focus on them!
Personally I'm hoping they they can figure out a way to improve the parallax barrier technology to a point where they can actually direct light in a fixed direction rather than the abitrary positioning we have right now. Combine that with the head-tracking technology which we already have and you could have glasses-free 3D which could actually take advantage of another mechanism humans use to perceive depth: Parallax. It'd be difficult to take advantage of that in film, but it'd be awesome for anything generated in real-time.
Of course, we could just skip this whole thing and wait for Holography to progress enough for general use! Can you imagine the problems that film makers are going to have to run into once faced with a medium where (in theory) your audience could be seated all around your "stage"? Perhaps cinemas will become more like theatres with pre-recorded stage shows.
Who knows... | 
| 
| | 
|
Modified: Thursday 1st of July 2010 at 10:04:27 PM |
t's not cooked yet!
