karl krach

i'm pretty sure this is quite easy but i still need some help with this: what i want to do is, in a gpu/3d-environment, make a plane which works like a (physically correct) mirror. so that when i point a 3d-camera at it, it shows what is behind the camera (and also behaves like a real mirror when moving the camera).

i made a stripped down, documented demopatch explaining the problem, see below. the setup is: a 3d-camera inside a textured environment sphere. the "mirror" is a plane cutting the sphere in half - the camera then films the mirror and should see what is behind itself.

i had a look at different shaders with "mirror" in their name but it seems these are more for kaleidoscope effects, not real mirrors. also i found this thread 

https://cycling74.com/articles/mirror-mirror-creating-a-glassy-texture-U

 which comes close but this works (i think) with some "environment_texture" i did not find documentation on.

anyone there to help me out? any help is appreciated a lot!

hey there.
i'm pretty sure this is quite easy but i still need some help with this: what i want to do is, in a gpu/3d-environment, make a plane which works like a (physically correct) mirror. so that when i point a 3d-camera at it, it shows what is behind the camera (and also behaves like a real mirror when moving the camera).
i made a stripped down, documented demopatch explaining the problem, see below. the setup is: a 3d-camera inside a textured environment sphere. the "mirror" is a plane cutting the sphere in half - the camera then films the mirror and should see what is behind itself.

i had a look at different shaders with "mirror" in their name but it seems these are more for kaleidoscope effects, not real mirrors. also i found this thread https://cycling74.com/articles/mirror-mirror-creating-a-glassy-texture-U which comes close but this works (i think) with some "environment_texture" i did not find documentation on.

anyone there to help me out? any help is appreciated a lot!
k


how-to-make-a-physically-correct-mirror-plane

add the camera and objects you want mirrored to a jit.gl.node and capture that node to a texture.

send that texture to jit.gl.slab and use the @rect attribute to flip the texture horizontally.

send the slab output to your plane shape.

add the camera and objects you want mirrored to a jit.gl.node and capture that node to a texture.
send that texture to jit.gl.slab and use the @rect attribute to flip the texture horizontally.
send the slab output to your plane shape.

i just examined what you did and though i still need some time to completely understand it, unfortunately it does not seem to do what i need: eg. if i move the camera closer to the mirror plane (camera facing the mirror), i do get to see less of the surrounding room behind it (which is the behaviour of a lense) not more (which is what happens when you move closer to a real mirror).

the background is, i try to make a monitor behave like a real mirror: in my full patch this 3d camera is controlled by the tracking data of the user in front of a screen and the texture mapped to the sphere is the actual camera feed...

hi rob!
thank you very much for your quick help!
i just examined what you did and though i still need some time to  completely understand it, unfortunately it does not seem to do what i need: eg. if i move the camera closer to the mirror plane (camera facing the mirror), i do get to see less of the surrounding room  behind it (which is the behaviour of a lense) not more (which is what happens when you move closer to a real mirror).

again, thanks so much for your help!
karl


sorry, another update: i thought some more about my whole setup and finally figured out that this "mirror plane" is not going to solve my problem in the first place, my whole approach about this was just wrong. at least right now, i think the effect i am looking for has to do with the lens angle of the 3d camera which has to be calculated from the distance of the viewer and the physical measures of my screen...

it keeps to amaze me how complicated it really is, to understand what is happening exactly in a simple mirror.

anyway, sorry for wasting your time but regardless thanks again for your help! all the best!

sorry, another update: i thought some more about my whole setup and finally figured out that this "mirror plane" is not going to solve my problem in the first place, my whole approach about this was just wrong. at least right now, i think the effect i am looking for has to do with the lens angle of the 3d camera which has to be calculated from the distance of the viewer and the physical measures of my screen...
it keeps to amaze me how complicated it really is, to understand what is happening exactly in a simple mirror. 

anyway, sorry for wasting your time but regardless thanks again for your help! all the best!
k


how to make a physically correct mirror plane?