3D images

Hi,

Im not sure where to go with the 3D image functionality required by the Bazaar website.

- Ive tried a variety of freeware 3d head/face utilites, and end up with a very nice 3d model of a head, with a ceramic cup bitmap across the face.

- Ive tried using Swift 3D, but it seems u must use primitives, and boolean operations, to generate the 3d model. Then u apply a material / texture to the model. Im very new to swift 3d, and even if i could do it, i would imagine it would take 1-2 weeks per model. At 12 models theres a few months work there. (3d max as far as i know is basically the same )

- Ive tried flash, and currently the best i can do is to add reflections. Im using some coding i wrote a few years ago, basically the ground control for a mario kart engine. I added reflections on all 4 sides, and it gives the impression of 3d, although its not. For anyone thats interested, heres 12 sample methods i wrote in Action Script previously (http://www.gooism.com/docs/RHForum.txt). Basically they are in increasing difficulties, trying to store values in an array, so that it can be converted to a ray tracing engine at later stage. Ray tracing engines are very complicated, but there a few about. The best one a wolfenstein engine. The best tutorial ive been able to find is (http://www.student.kuleuven.be/~m0216922/CG/raycasting.html).

I dont really want to get into ray tracing engines, although if it worked, it would take 5 minutes per 3d model.

Just wondering what others are going to do for 3d images. I cant find a simple free utility that converts an image to 3d.

Craig

I forgot to add,.. - or stereoscopic images ( or anaglyphs )

Usability isnt good, cause u need to wear glasses. But it might be a simple solution. On a product page, u could click to change the image to 3D, with a note requiring glasses. Not as good as a flash solution, but in true 3D. So far i found, http://www.stereoscopy.com/downloads/ , where the best one is called AnaBuilder. Works well and freeware. Another one ive had for a while, but only works fully for a trial period is http://www.softwarelode.com/18768/details-digital-3d-photos.html . Might be the way to go due to time restrictions. Just a matter of generating thumbnails, and setting the 3d depth.

Also, one called oneeye at stereoscopy. It converts images to 3d, and doesnt need glasses, but does need the software. Doesnt do a very good job.

I have a rather comical issue, and its spread to my bazaar slideshow. On a plus side, it has the effect of bumping my hit count up. The issue is creating 3D Reflections with Flash Action Script 2.0. I have to use the duplicate movieclip command. This command duplicates the clip, but unloads the loaded bitmap image. Ive tried copying it to an object variable, then swapping it back, but this doesnt work. Flash Action Script 3.0, includes object.bitmap data, and the adobe website, gives a general function to load an image into it, then clone the image. I understand this function, but would have difficulty converting my slideshow actionscipt coding to AS3, its very different, although as standard practice i script all my clipevents and mouse functions.

So, the result of my slideshow reflections, having to attach, and duplicate clips, is that it loads the image, and 10 copies of it, and if theres 7 images on the stage, it loads 77 images. And each time it generates a new image, it loads a further 11 images.

Searched everywhere, and the best i could do was the adobe sites code. The basic bitmap command of AS3.0 references the library linkage's, and not url files. But the adobe method works around this.

(The only way around this issue in AS2.0 is to have all the required images on the screen at the same time, or paste all the images into flash, making the flash file very large. My images are read from an xml file, so i would prefer to keep it in a dynamic state.)

After looking for a suitable 3D programme, of which I can recommend blender.org if anyone is wanting to do 3d rendering and animation, I've concluded the simplest(and probably best) way to do the 3d is with a video camera filming while a pot is turning. I found a site that would turn a 2d picture into a 3d scene, but it's nothing like what we require. 3d progs are not made for this purpose.

Hi Martin,

Ive been busy,.. saw ur prototype bazaar site, great stuff,.. saw ur cube image changer, and experimented with 2 methods of my own. Did it in javascript, but needed to cut the images into strips, only ie has a mask filter. So, made it easy in flash. Refresh this page twice, to avoid the welcome screen. http://gooism.com/bazaar/products/bcpot002.html . The code is on www.gooism.com/bazaar/js/bazaar.js . Action script code, and the page is a bit slow, 3 images instead of 1. The middle one is me mucking about with cos to fudge a good look. Silly idea, i joined the 2 image changer images together, in 2 halves. Stretches them out too much at top and bottom.

I Found a Flash AS2.0 3D class, that is easy to make pologons. AS2.0 cant copy pixels, or work with bitmaps, it just draws the lines of a polygon and colors it. Too early to go to AS3.0 for me. Trying to remember my uni maths, a circular integral would just give the area of a 3D shape, which u can calculate, but can u do a circle integral with respect to angle , and what might it imply ? for every point in x,y or z. area = dxdydz (1/2 f1 + f2..fn-1 1/2 fn) can change it to in terms of angle(xy,yz,xz) and yz, angle(xy,yz,xz) xz or angle(xy,yz,xz) xy. A script loop. The derivative gives the change of the variables with respect to the above arranged functions. Changing angles in the 3 planes will rotate the 3d object.

But how do integrals and derivatives help in the modeling process and 'application to programming'.

If i set a function for the entire model, ... x = angle(xy,yz,xz),y,z, ...y = angle(xy,yz,xz)x,y ... z = angle(xy,yz,xz),x,y, then what might happen to the 3d shape. Might be interesting. I think stokes theorm is for identical f1 f2 and f3 functions, in rearranged terms, but perhaps ray tracing engines work diffferently. Trying to make a coord system in flash, that i can spin, and test some functions out in.

Ive found a AS2.0 ray tracing engine. But they work by applying ittle images on to boxes in 3d. They are not really 3d at all. Just gives a floor , wall and ceiling bitmap fill.

Perhaps in flash AS2.0, i can calculate an object z depth, as 1/2 the distance from any 2 xcorrd points on the cut out bitmap. Say, vase, then id need to somehow convert my image into a xml file storing the rgb vales per pixel. AS2.0 cant do this , AS3 can. Maybe u know of a program that can do that ???. Then id need to apply a burring function that averages the color. So that i can draw it in polygons,.. then fill with a color. It might work. Good stuff. Going to give it a go.

Ill check out that blender one.

Thanks

.. That all sounds a bit too complicated. Ive forgotten far too much maths, did hons engineering, but 10 yrs ago, trying to simpify everything. I made functions, where i enter in a point on the screen, (0-650),(0-400), and it returns a number in cartesian coords, (-1 0 1), (1 0 -1). And for scaling, i can just ask for the screen corrd at (-1 0 1), (1 0 -1). Helps, cause flash and javascript all rely on 0,0 at the top left. And maths is not like that. Then i can just use y = f(x) in an eqn. Gotta make little return angle and radius, polar to cartesian, and cartesian to screen. Then somehow, add the z field. But at its simpliest it will be a 10 line code for mouse trailers, that rotate on their own centerpoint, and u just apply a cos rule to it. Kinda of like the ships circle and flyin in galaga.

I want to use 3d to draw a lazer special effect around an image/text outline from a vanishing point. I made http://gooism.com/bazaar/products.html where in the code i just enter a y=f(x) eqn, and it helps to remind me what the hell cos is anyway (lol). But the idea was,

If i draw a line from every point on the screen, to a vanishing point, when it hits an object, it can , say remove the line. I would be left with a great special effect. Unfortunately the idea is not practical. I needed an array of (1280x800), huge, then a found array set to max 10% of screen, 10% of huge. I divide the screen into boxes, equal, and worked out how to move the center point,. a nightmare. In flash, u draw a line, and it centers itself on the top left point. The scale is always 100 in both directions, and the height and width are always positive. Its very hard, to even keep track of the 2 end points. So i just redrew the lines in the example. Not a bad special effect, and the code is so simple.