[Lab] Depth Perception

Darcy Whyte darcy at siteware.com
Mon Aug 22 16:47:07 EDT 2011


http://ignacio-torres.com/projects/stellar/

here's another cool 3d that's based on the observers' motion...


On Mon, Aug 22, 2011 at 4:29 PM, Darcy Whyte <darcy at siteware.com> wrote:

>
> Also, some of the joy of depth perception comes from when you move your
> head and the image changes. With a camera system the cameras can only change
> direction but not location.
>
> I suppose some sort of RC aircraft could carry each camera. That's be cool.
> Now we're getting expensive. You'd need two quad copters, some nice
> stabilized cameras and lots of time. :)
>
>
>
>
>
> On Mon, Aug 22, 2011 at 4:24 PM, Paul & Andrea Mumby <themumbys at gmail.com>wrote:
>
>> Too bad stars are so far away that even with several kilometers of
>> stereo separation you likely won't get depth out of the starfield. Because
>> that would be damn cool...
>>
>> - Paul
>>
>>
>> On Mon, Aug 22, 2011 at 4:21 PM, Darcy Whyte <darcy at siteware.com> wrote:
>>
>>> Ah, I see, the eyes can can change their alignment to snap on to the 3d
>>> experience automatically.
>>>
>>> So the easiest solution is to just have the cameras stationary, perhaps
>>> converged a couple km out.
>>>
>>> Once adjusted you can see the clouds go by in 3d.
>>>
>>> I guess time lapse would be a great way to capture the material. Then you
>>> could play it back whenever you want and the motion would make it cooler.
>>>
>>> But if you wanted to go real time an pan, you'd need to make sure the pan
>>> rates are well match or it would "un-snap" you from the 3d experience.
>>>
>>> I bet you could get a nice 3d effect even if the cameras were only like a
>>> few meters apart.
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>> On Mon, Aug 22, 2011 at 4:01 PM, Jason Cobill <jason.cobill at gmail.com>wrote:
>>>
>>>>
>>>>    Alignment's not too tricky - generally, 3D movies aim the cameras
>>>> parallel to eachother and let your brain do the tricky work of converging
>>>> the images. They don't need to be exactly parallel for this to work, and
>>>> this is the reason some people get wicked migraines watching Avatar.
>>>>
>>>>    You can get the same effect as the XKCD experiment (and Avatar) with
>>>> a highly simplified process:
>>>>    - Find a point on the horizon and take a picture of it.
>>>>    - Walk a few feet.
>>>>    - Take a picture aimed at the same point on the horizon
>>>>    - Bring the two images together side-by-side on the computer
>>>> (preferably without border lines between them)
>>>>    - Cross your eyes really hard like you're looking at a "magic eye"
>>>> image, and they'll converge and you can see your clouds in 3d.
>>>>
>>>>    Or save yourself the effort of going outside and look at these:
>>>>    http://phereo.com/   (Be sure to click "Mode" and select "Crossed"
>>>> if you don't want to use special glasses)
>>>>
>>>>    The technique is really old - I have a number of friends with
>>>> collections of stereoscopic images from the mid 1800's! It was a fun party
>>>> toy for Victorians.
>>>>    You can find out a bit more here:
>>>>    http://en.wikipedia.org/wiki/Stereoscopy
>>>>
>>>>    Also worth noting that stereo pairs are extensively used in aerial
>>>> surveys - a plane takes two photos in succession and the pair can be
>>>> converged to get a 3d view. I've often spotted clouds, boats, and even other
>>>> planes in high-altitude aerial surveys. I should mention that the effect is
>>>> a little less exciting than the XKCD comic paints it to be. Having a pair of
>>>> stereo glasses definitely relieves the stress of staring crosswise and lets
>>>> you feel a bit more immersed in the image.
>>>>
>>>>    -Jason Cobill
>>>>
>>>> _______________________________________________
>>>> Lab mailing list
>>>> Lab at artengine.ca
>>>> http://artengine.ca/mailman/listinfo/lab
>>>>
>>>>
>>>
>>> _______________________________________________
>>> Lab mailing list
>>> Lab at artengine.ca
>>> http://artengine.ca/mailman/listinfo/lab
>>>
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://artengine.ca/pipermail/lab/attachments/20110822/726168c7/attachment-0001.htm>


More information about the Lab mailing list