I am rendering a bunch of cg images but now I need to make an applet that will view them. More specifically these images are very much like the first generation of google street view where you have a picture and you can get a camera to crop a section of the image and as you pan left, right, up or down the cropping moves left, right, up or down. However this presents one problem. I can only render as a spherical image and that is the equivelant of setting the texture of a sphere where the camera is inside the sphere. But as I know Java doesn't support 3d properly for applets. This means I need an algorithm to fake the allusion of the sphere object by re-arranging the pixels and removing some pixels. An example of the resulting image can be found at the below link.
So my question is how can I turn an image such as the one above to a Java applet where you can pan around like street view without 3d Java?


hmmm why reinvent the weeel,

this one has flickering issues too -> Route 66 Follow Me (I think that's withOut public API)

sure Google street (GNU or some version of public API)

- 3D is allways only 2D plus knowHow, be sure "Java Animations" reqiured healtly dvivers for GPU
- for a.m. descriptions you'd needed think and impements 3D
- you have to create DrawableImage, ZoomIn/out, some woodoo for Switch between Images
- flickering you have to calculate with increase/decrease acceleration instead of linear, that's your job, correctly handle floating acceleration and change to the quasi linear ...
- <fw> pentium4/AMD64 is best of the HW platfom for testing </fw> :-) if still exist ....

only about JApplet

please carrefully read todays JApplet restriction(by Andrew Thompson, follows link to his web) http://forums.oracle.com/forums/thread.jspa?threadID=2189526&tstart=0

I don't know another person with endless JApplet knowledge (plus Swing) and endless EGO too (who's want to hepl and publish own knowHow)

Although I haven't got much code from your post I liked liked the first bullet point.

Basically to achieve my goal, I need an algorithm for creating a textured semi-sphere which is always at the same position where the inside center of the semi is pointing flat towards the camera where the camera is the display area of the applet. And then the sphere gets cropped in the middle as big as it can to get a square. Then this square that is cropped is the resulting image for the current view.

Also to pan left right up and down, the input image is patterned horizontally and vertically twice so the image reappears 4 times in the new source image although the vertical image will have the top part cropped in half and the bottom pattern cropped in half with an entire middle so it is like a grid of 2x3 where the 3 is contained of two half images (top and bottom) and a whole image in the middle. All of these images in the parent image are the same input image.
Then the image gets croped from specified positions and the width and height of the crop is equal to x=width/4 && y=height/2. So that new cropped image will then be processed in the sphere. But different positions to crop from (eg moving mouse changes crop position) will then result in different panning.

So my main question is how can I draw a hollow semi-sphere which is pointed to the screen like the inside of a bowl pointed towards a camera and crop the circular picture to make a square where each corner of the square meets the boundary of the circular semi-sphere result. The position or angle of the semi-sphere never changes. It is only the texture which is constantly changing. Any advice on how to program such graphics.

Thank you

Does anybody know if OpenGL has been written in Java instead of it's normal C language. If it hasn't then I might have some fun converting OpenGL from C to Java so that any Java application can run OpenGL.

avoid and don't complicate simple things, because you are knew these variables
(are strict and overRodeAble),

1/ Assume therefore that we have no pictures only enjoy with plain vanilla Colors
2/ Color must be changes if you change any of 6th. directions
<forward, backward, up, down, left, righ>
3/ color must be changed (f.e.brighter) if accelerate, then darker on decelerate event
4/ that nothing differents as your Images snipped, just you have to tune the transition between each images

there I didn'd give you any advice, because I think that there .... could be lots of isssues on EDT by using Timer and Alpha for stepped transparency == switching betweens ImageSnipped,

with Colors its very easilly and confortable, with Images to hard ...

much luck