front fixed projection textures

[Gmanx]

You could edit the camera and apply a texture to it?

[Samir]

If I'm not mistaken, applying a texture to the camera would make the texture a sort of light filter for that camera. It would be like projecting (as in, light projection) a slide of your texture image on to your scene geometry. That is not quite what that Lightwave tutorial is trying to achive. Also, the camera image projection would follow camera movement, destroying the illusion of a properly textured 3D scene.

What you need is a way to use the original camera location as the coordinates to project the texture location on to the geometry in the scene (not just a light projection). This can be done in Blender using the "Object" texture projection mode.

Basically you create an empty at the exact location and rotation of the base position of your camera. You then use this empty for texture projection as described in this classic tutorial:
http://www.ingiebee.com/tutorials/Decal%20Mirror%20Modelling/theeth%20decal.htm

This way, even when the camera moves, the texture coordinates are fixed by the Empty and you get the required illusion of a 3-dimensionally textured scene.

I've never tried this myself to mimic camera projection, but it should work.
Best of luck.

Samir

[Joat]

You had a thread with an example movie of cameramapped scenery? For what I understood, the technique used in that tut of yours is exactly the same.
Or is there something I'm not getting?

[RedSquirrel]

The problem is if you read that lightwave tutorial it says that you can rotate, the camera slightly, now if you do this using "win" in the "map to" settings it will place your background image/movie on your mesh and will move it with the camera which is not what your want because you want to make the scene a little more 3D, so instead of clicking on the "win" click on "stick", if you render it though nothing will happen... well there is another button that you have to press before it works go to the edit button "F9" and see that button that says "sticky : make" click on it an render(it now will save the sticky coordinates), render it and you see the image on the mesh. There you go and you can rotate that camera and render it again and it becomes more 3D just like the lightwave tutorial, if your not happy with the view all you do it click the "sticky: delete" button which will clear the texture sticky coordinates from memory.

:thumbsup: have fun.

[jendrzych]

Nice technique, regular photogrammetry btw, but has several limitations (sticky coordinates; texture covers only area which photo does).
If I were You, RyanVII, I'd use UV mapping (planar UV mapping from windowpowiela). It'd let You simulate other sides of those boxes, which are not on Your picture, by simply cloning part of texture with 2D app :)
Moreover You can use two photos (front and back), split Your geometry and map both of them with those pics, receiving full 3D copy of part of reality :P (Blender doesn't support multi UV maps AFAIK).

[RyanVII]

joat: my demostration of camera mapping has it's limitations. you have to use photoshop to clone scenary that will be revealed by camera movements. not to mention every element in the photograph has to be outlined perfectly and made into a layer. the layer is then applied to 3D geometry by "sticky texture".

the light wave tutorial seems to allow you to apply a photograph to 3D geometry that resembles the photograph without the need of 2D application (photoshop).

i was wondering if that is possible in blender.

[jendrzych]

RyanVII, I don't get, what You need photoshop for?
It's easy, just as RedSquirrel wrote. I attached zipped pdf with quick test.
Maybe didn't understand sth?

[RyanVII]

I still do not understand how you accomplished your quick test.

First you contructed the 3D geometry of your picture. Next you applied a shadeless material to the 3D geometry. Then you mapped the photograph to the 3D geometry with "Win". Next you erased the "Win" and mapped it with "Sticky".

Here is the picture I am trying to camera map.

I'd really appreciate your help in getting me to understand how you accomplished what you did.

Here is my blend file.

[Joat]

You use the mapping method "win" to get your image to show when you create the geometry. When you have done with creating it, you apply the image on top by changing the texture's mapping method to sticky. Before that you should make your geometry sticky by pressing "make sticky" on editing screens of the object(s) (F9).
It pays to subdivide your geometry a few times before making it sticky. That way there is a lot more points to attach the cameramapped image to, and less distortion when you move the camera
You don't need to use photoshop at all, only if you want to reveal geometry that is not visible in the original image you of course have to create textures for that. That is where you could need photoshop or gimp. As you can see from the Lightwavetutorial, thy also created new textures for the hidden geometry. I think there's no software in existence that can produce textures by quessing from the original image. For that you need human. In your example it would obviously mean creating a second image where the boxes are removed from the picture. And using that image as a projectionmap for the walls and the floor. Separate textures could be made for the boxes too, and you could quite easily even create uvmapped textures for them if you wish.

Check out Gordon and Alex Levy's exellent videotutorial. (http://www.peerlessproductions.com/alex/CM.html) on the subject.

[jendrzych]

Joat was faster...
Here's *.blend to play with the geometry and mapping translated from lightwave, (made very rough).

http://blender.isoserv.com/uploads/sticky%20test.blend
http://blender.isoserv.com/uploads/sticky%20test.blend

[RyanVII]

jendrzych and joat. thanks. i think i understand how to accomplish this now. at least for the time being ;)

do you two use this effect at all?

[jendrzych]

At work, sometimes use two photos to recovery 3D geometry of surrounding environment, to measure things that are impossible to measure live, or when there's no time to measure things, or when just forgot about sth. So by the way I'm getting 3D background for previsualisations my designs. Nothing really spectacular. I love to make renderings, but there's no place for such activity at my job. Only few of my clients really needs renderings of things, which will be made in short time. On the other hand previsualisation of a conception of architecture placed in real scenery, even rough one, makes design drawing more understandable for common mortals :)

[members only hunter]

i am in the process of making a movie and will be using this process along with motion tracking and pure 3d.

the project is a superhero movie and will be taking place in a fictionalized real world that will be filmed and photgraphed but will be utilizing purely 3d characters. i will go out and film/photgraph the background plates and drop the characters in as 3d objects... there is a nice little free motion tracking program out there for linux that you might also want to look into... the only thing is that it only exports to blender but i guess that isn't so much of a problem now with all of the formats that blender exports to...

chris

[TroutMaskReplica]

there is a nice little free motion tracking program out there for linux that you might also want to look into

what is the name of the motion tracker?

[2d23d]

do you mean Voodoo?
http://www.digilab.uni-hannover.de/docs/manual.html

[CGTalk Moderation]

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.