how Render Camera View To Texture(Interactive) ?

[Joker386]

hello

how i can create interactive texture that show my camera view and use it for material or composite network ?

maybe i need a expression that when time changed , render current frame and save it in tmp file , then load this tmp file in composite network or apply to material,s color map !

but if this way is correct , how i can access to mantra node with HScript or Python ?

thanks

[Annon]

hello

how i can create interactive texture that show my camera view and use it for material or composite network ?

maybe i need a expression that when time changed , render current frame and save it in tmp file , then load this tmp file in composite network or apply to material,s color map !

but if this way is correct , how i can access to mantra node with HScript or Python ?

thanks

I might be wrong on this, but you won't be able to have what you want totally live.

I think you're going to have to write out a sequence as you say, but might have to force a node to re-cook every frame etc and render your mantra node.

I'd just separate the two. Get your first camera working how you want it, render out the sequence, then use that sequence in your second render.

[gaurav]

I think its very much doable with UV unwrapping with a bit of twist

1 . Delete all back facing prims so uv's dont overlap (Quick vop sop would help.

I dont know if there is already a sop for something like this.)

2 . Lay down UVtexture sop and project from camera you want, Apply to vertex .

Do a quick check for uv's in uv layout window to make sure uvs are not overlapping.

3. Create your mantra rop and point it to your obj in the following location.

properties>render> UV render obj> [your scene level obj]

4. Now in Cops use Render node and point it to the rop just created.

and finally

5. Use op: expression to fetch the image in desired texture node.

6. last step : you need to test it all out.

Cheers,

p.s. - I wish opengl rop had unwrapping option

[rafaelfs]

There's a Render COP, you can pipe a mantra node through it and use the output as a texture map elsewhere with the op: command.

That way you can render the scene in one single step. Although setting up render with dependency is not that difficult either...

Cheers

[rafaelfs]

There's a Render COP, you can pipe a mantra node through it and use the output as a texture map elsewhere with the op: command.

That way you can render the scene in one single step. Although setting up render with dependency is not that difficult either...

Also, I just came across this:

As usual, in houdini, there's a number of solutions to the same problem...

[Joker386]

realy thanks all my friends

i can create it

tnx * 100000

[papsphilip]

old thread to revive , but there should be an easier way to just grab the viewport as an image realtime and have it as a texture on a grid.

The raytracing experiments are really interesting but they are a bit involved and allow for a lot of tweaking if you want something specific.

The viewport gets rendered realtime and it would be nice to have an option to grab that live openGL somehow.

i suppose you could create a TOP network and trigger an openGL ROP every time there is a change in your hip file or when you save or when the camera moves, but that would be slow and certainly not realtime.

i am working on project where i am making a short video for a screen inside a room, i have setup the room in 3d and i would like to check the layout of my scene on that screen and not in the 3d space i am building it.

could there be a hack to steal the viewport stream?