Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
How to implement horizontal 360 degree view using OpenGL?
73 views
Skip to first unread message
Linda Li
Feb 14, 2013, 2:47:11 PM2/14/13
to
Like the one:
http://www.gopano.com/video/MjI3
I can think of two ways:
The first method: use cube map. But the top and bottom faces are black.
The second method: map the horizontal 360 degree picture on a cylinder. Then put the camera at the center of the cylinder, and use lookat to control the eye direction.
Which is faster on mobile phones, especially for videos?
http://www.gopano.com/video/MjI3
I can think of two ways:
The first method: use cube map. But the top and bottom faces are black.
The second method: map the horizontal 360 degree picture on a cylinder. Then put the camera at the center of the cylinder, and use lookat to control the eye direction.
Which is faster on mobile phones, especially for videos?
Linda Li
Feb 14, 2013, 3:41:30 PM2/14/13
to
Personally, I think cylinder is faster, since the picture can be directly mapped on the cylinder.
For the cube map method, conversion is needed to convert the picture to 4 cubes.
For the cube map method, conversion is needed to convert the picture to 4 cubes.
Nobody
Feb 15, 2013, 2:25:50 AM2/15/13
to
On Thu, 14 Feb 2013 11:47:11 -0800, Linda Li wrote:
> The first method: use cube map. But the top and bottom faces are black.
> The second method: map the horizontal 360 degree picture on a cylinder.
> Then put the camera at the center of the cylinder, and use lookat to
> control the eye direction.
>
> Which is faster on mobile phones, especially for videos?
It depends upon how the video is mapped, and how much distortion you can
> The first method: use cube map. But the top and bottom faces are black.
> The second method: map the horizontal 360 degree picture on a cylinder.
> Then put the camera at the center of the cylinder, and use lookat to
> control the eye direction.
>
> Which is faster on mobile phones, especially for videos?
tolerate.
If the video was shot using 4 cameras, the data will be in the correct
projection for a cube map from the outset.
Linda Li
Feb 21, 2013, 6:04:01 PM2/21/13
to
Thanks. The input image after processing is a horizontal 360 degree image, a flat image.
So I used cylinder to do texture mapping.
And it worked very well.
So I used cylinder to do texture mapping.
And it worked very well.
0 new messages