Questions about Stereoscopic 3-D

[tade]

When we watch 3-D films in the cinema, some objects appear to protrude, others recess. How would a technician control which objects protrude and which recess?

Also, why do the 3-D images sometimes appear like cardboard cutouts?

 

[NascentOxygen]

When we watch 3-D films in the cinema, some objects appear to protrude, others recess. How would a technician control which objects protrude and which recess?

I guess you are talking about computer generated images, rather than those taken by a pair of cameras?

Take a photo of any scene and make a second copy of that photo. Carefully cut around a few objects in that second photo and slide some of them slightly to the right, and others to the left, relative to the rest of what's in the photo. Now, superimpose that pair of photos using your stereoscopic vision. The objects whose image you moved to one side appear recessed in 3D, the objects you moved in the other direction will appear to have come forward off the photo's background.

Also, why do the 3-D images sometimes appear like cardboard cutouts?

If you want a model's nose to appear closer to the viewer than that model's face, then you must slide the nose a tiny bit further to the side than you slide the face itself.

You can see that without the assistance of a computer the task of creating a realistically good 3-D effect will be painstaking.

 

[tade]

I guess you are talking about computer generated images, rather than those taken by a pair of cameras?

Take a photo of any scene and make a second copy of that photo. Carefully cut around a few objects in that second photo and slide some of them slightly to the right, and others to the left, relative to the rest of what's in the photo. Now, superimpose that pair of photos using your stereoscopic vision. The objects whose image you moved to one side appear recessed in 3D, the objects you moved in the other direction will appear to have come forward off the photo's background.


If you want a model's nose to appear closer to the viewer than that model's face, then you must slide the nose a tiny bit further to the side than you slide the face itself.

You can see that without the assistance of a computer the task of creating a realistically good 3-D effect will be painstaking.

So, if I was converting a 2D video into 3D, I might have to do it "manually", and this will result in the cardboard-cutout effect?

 

[NascentOxygen]

Whether "manually" or by using the computerised equivalent basic technique, that outcome will be as I described. So a crude conversion of a football give it the appearance of a flat lollypop. You'd need to shift different parts of its surface by different amounts to give it a surface with realistic 3D depth.

 

[tade]

Whether "manually" or by using the computerised equivalent basic technique, that outcome will be as I described. So a crude conversion of a football give it the appearance of a flat lollypop. You'd need to shift different parts of its surface by different amounts to give it a surface with realistic 3D depth.

Will this problem exist with the footage from 3D Cameras? i.e. Cameras with two lenses

 

[NascentOxygen]

Will this problem exist with the footage from 3D Cameras? i.e. Cameras with two lenses

I don't expect it would ― the cameras record precisely what the eyes would see. However, while I've dabbled in computer-generated 3D, I know nothing about proper 3D photography.

Did you see the 3D effect in the pic I attached?

 

[tade]

I don't expect it would ― the cameras record precisely what the eyes would see. However, while I've dabbled in computer-generated 3D, I know nothing about proper 3D photography.

Did you see the 3D effect in the pic I attached?

Samsung released a phone with 3D Camera and a lenticular 3D display.


I don't know how to make it work.

 

[NascentOxygen]

I don't know how to make it work.

You have been short-changed if your science education overlooked an introduction to viewing random dot stereograms.

Practise on these. http://www.eyecanlearn.com/random_dot_stereogram.htm

Disclaimer: persons without sight in both eyes will be unable to view the 3D effect

 

[tade]

You have been short-changed if your science education overlooked an introduction to viewing random dot stereograms.

Practise on these. http://www.eyecanlearn.com/random_dot_stereogram.htm

Disclaimer: persons without sight in both eyes will be unable to view the 3D effect

No ****, schools wouldn't give a damn about this.

Those are awesome, it's as though I'm looking at a wall that's behind my monitor.

 

[NascentOxygen]

Now go back to the simple pic I generated. It works the same way, and you can easily see how I modified the first image.

 

[tade]

Now go back to the simple pic I generated. It works the same way, and you can easily see how I modified the first image.

Ok, I managed to see some floaty letters

Yeah, you shifted the d and the g.

 

[NascentOxygen]

g is part of the background. I shifted d and f.

 

[tade]

g is part of the background. I shifted d and f.

oops. funny thing is, I believe I still have depth perception when I cover one eye and look around my room.

 

[tade]

oops. funny thing is, I believe I still have depth perception when I cover one eye and look around my room.

though I shouldn't have depth perception, right?

 

[NascentOxygen]

Using only one eye you lack depth perception. But I think we acquire much depth information by a number of alternative techniques. You know that a drink can is smaller than a chair, so if the drink can looms larger then we infer that it is closer. (But you could be tricked by someone fabricating a gigantic can and placing it more distant than the chair.) Also, by moving your head to one side you can see which objects appear to undergo the greater shift relative to the background, indicating they are closer. In addition, you can discern objects which are close by the effort the eye muscle must exert to focus the object's image.

 

[tade]

Using only one eye you lack depth perception. But I think we acquire much depth information by a number of alternative techniques. You know that a drink can is smaller than a chair, so if the drink can looms larger then we infer that it is closer. (But you could be tricked by someone fabricating a gigantic can and placing it more distant than the chair.) Also, by moving your head to one side you can see which objects appear to undergo the greater shift relative to the background, indicating they are closer. In addition, you can discern objects which are close by the effort the eye muscle must exert to focus the object's image.

But my depth perception doesn't disappear when I cover one eye. At least, I have this strong belief that it doesn't. It's probably psychological.

 

[NascentOxygen]

With only one eye, you have no stereoscopic vision, so lack the depth perception due to stereo vision. But you still have other means for perceiving depth, the ones I outlined.

 

[tade]

I see.

So that's why I don't notice any immediate loss in my depth capability