« Return to Thread: Tutorial for taking photos for 3D?
Re: Tutorial for taking photos for 3D?
by Ashley Sheridan-3
::
Rate this Message:
I'd recommend taking shots at less than every 36°, for the simple fact that it will only give you 10 images, which is slightly under a seconds worth of animation (assuming the typical 12fps used in computer screen animation.) Ideally you should decide how long a rotation animation should take, assume 12fps to work out how many shots you need and then divide 360° by that amount. Obviously, the quicker the animation, the less room you have for erroneous jerks in the animation, and the smoother it will appear.
For completely free movement though, I'd go for Lukes method and make a low-polygon model. Applying the images you took as textures, you can make even a low-poly model look realistic enough. The best shots will probably come from the underside of the board, as you can take them without all the components getting in the way, although you'll have none of the text that you get on the topside of boards (they only print it on the tops,) but lots more of the pathways.
Thanks,
_______________________________________________
Papervision3D mailing list
Papervision3D@...
http://osflash.org/mailman/listinfo/papervision3d_osflash.org
_______________________________________________
Papervision3D mailing list
Papervision3D@...
http://osflash.org/mailman/listinfo/papervision3d_osflash.org
For completely free movement though, I'd go for Lukes method and make a low-polygon model. Applying the images you took as textures, you can make even a low-poly model look realistic enough. The best shots will probably come from the underside of the board, as you can take them without all the components getting in the way, although you'll have none of the text that you get on the topside of boards (they only print it on the tops,) but lots more of the pathways.
Thanks,
|
Ash www.ashleysheridan.co.uk |
[Long post]
Edward,
It sounds more like you want to do is simulate 3D by showing different pictures of an object from different angles. In that case you wouldn't need PV3D as it can be done simply with flash movie and a series of images.
You would start out by taking a picture of the object from successive angles (say every 36 degrees) for a full 360 degrees. Import the pictures into flash - placing one picture on a frame by itself. You can make it interactive by writing a script to gotoAndStop() on each frame corresponding to an angle based on the user's mouse position. That will give you a view around a single axis. For a double-axis view, take pictures as you would, but for each revolution around the object, rotate your camera on the perpendicular axis by a set angle (also 36 degrees) and then do another revolution. You would then combine the images in a 2D grid, and show a cell from that grid depending on an x and y coordinate take from the user. The advantage of this technique is that you can get photographic 3D views. The disadvantage is that your angular resolution is dependent on the number of images you take. If you take photos at 36-degree intervals, you will only have a resolution of 10 degrees, which may or may not be of use to you.
However, if you do need a real-time 3D rendering that you can fly around and rotate and zoom to your heart's content, your other option is to scrap the rotary table and rather manually create low-poly models of the objects you want to portray (in 3DS MAX for example) - then use orthogonal photographs as texture maps to add detail to the components. By low-poly I mean your ram module that you mentioned might be a plane for the memory board (with texture mapping on either side). The ram chips themselves would then be flat rectangles (also texture mapped) stuck onto the ram module. You definitely wouldn't model things like the legs on the chips - you just want the minimal number of vertices in order to map textures and give you some 3D properties.
I hope this introduction helps. Please let me know if you need any explanations or clarifications on the concepts mentioned here ;)
- Luke
Edward,
It sounds more like you want to do is simulate 3D by showing different pictures of an object from different angles. In that case you wouldn't need PV3D as it can be done simply with flash movie and a series of images.
You would start out by taking a picture of the object from successive angles (say every 36 degrees) for a full 360 degrees. Import the pictures into flash - placing one picture on a frame by itself. You can make it interactive by writing a script to gotoAndStop() on each frame corresponding to an angle based on the user's mouse position. That will give you a view around a single axis. For a double-axis view, take pictures as you would, but for each revolution around the object, rotate your camera on the perpendicular axis by a set angle (also 36 degrees) and then do another revolution. You would then combine the images in a 2D grid, and show a cell from that grid depending on an x and y coordinate take from the user. The advantage of this technique is that you can get photographic 3D views. The disadvantage is that your angular resolution is dependent on the number of images you take. If you take photos at 36-degree intervals, you will only have a resolution of 10 degrees, which may or may not be of use to you.
However, if you do need a real-time 3D rendering that you can fly around and rotate and zoom to your heart's content, your other option is to scrap the rotary table and rather manually create low-poly models of the objects you want to portray (in 3DS MAX for example) - then use orthogonal photographs as texture maps to add detail to the components. By low-poly I mean your ram module that you mentioned might be a plane for the memory board (with texture mapping on either side). The ram chips themselves would then be flat rectangles (also texture mapped) stuck onto the ram module. You definitely wouldn't model things like the legs on the chips - you just want the minimal number of vertices in order to map textures and give you some 3D properties.
I hope this introduction helps. Please let me know if you need any explanations or clarifications on the concepts mentioned here ;)
- Luke
On Wed, Jul 16, 2008 at 6:13 PM, Edward P. Sager <edsager@...> wrote:
Hello All,
I have a few small computer items (processors, RAM, etc.)of which I need to take photos and transfer such to a 3D representation of the items. I am using Swift3D and the Export to Papervision 3D and Photoshop for image manipulation/cleaning.
Does anyone know of any tutorials which show how to take photos for later 3D representation?
I have a green screen background and have fashioned a turntable to rotate the items to be photographed, but I am unsure of the best methods to go about this project.
Thanks for any help,
edsager
_______________________________________________
Papervision3D mailing list
Papervision3D@...
http://osflash.org/mailman/listinfo/papervision3d_osflash.org
_______________________________________________
Papervision3D mailing list
Papervision3D@...
http://osflash.org/mailman/listinfo/papervision3d_osflash.org
_______________________________________________
Papervision3D mailing list
Papervision3D@...
http://osflash.org/mailman/listinfo/papervision3d_osflash.org
« Return to Thread: Tutorial for taking photos for 3D?
| Free embeddable forum powered by Nabble | Forum Help |
