Hi my friends.. i’ve a photo where need to insert a new building near one that actually exist.. the problem is that the original picture is far from my new build because need to view all the other buildings..
So what is the problem? is that i need to do the render with a large resolution because my new project is too small in the render..
There is a simple method to do the render of a restricted area of my screen without loose the prospettive of my object?? In this case.. i can do a lower resolution render and lose less time..
Re: How render a part of screen to insert in a photo?
paerhaps.. i’m not explain so well…
so.. this is the original picture:
This is the render:
and this.. the merge:
So.. you see that i need to do a render with a highest resolution because the point of view is far from the object.. then.. i need to spend too much time to render..
So.. how can render only a small part of the scene (the part that contain the new build??) to decrease the render time??
Re: How render a part of screen to insert in a photo?
Orie
I would only render at a resolution to suit the output. So if you original pic is 3264pix then render at this resolution.
As you will be deleting virtually all the foreground, I would put it on another layer and turn it off.
Then surely we are only talking about a few minutes to render the rest.
Re: How render a part of screen to insert in a photo?
That’s a good idea.
I never used Photomatch but, another way you can do it is by rendering a wide image, so the hight is cut down to the needed part of the model.
You can use Thom’s V-Ray Toys for accuracy.
Re: How render a part of screen to insert in a photo?
Orei,
1 ) DCauldwell’s tip is spot on- it’s often best to render at exactly the same resolution as the original photo anyway.
2 ) Regarding high-resolution images of distant subjects, something a lot of people forget is that changing the focal length of your camera will bring your subject closer to the camera without distorting the image like zoom (which is effectively moving the camera) would.
So set up your photomatch (as you already have), with the subject in the distance just as it is in the original photo, then simply type a longer focal length into the VCB/Measurements box and your subject will “magically” fill more of your viewport without any zoom-related distortion.
I think the confusion arises from the fact that “zooming” in photography means not moving the camera, but changing the focal length of the lens whereas “zooming” in 3D modelling programs is actually moving the camera and leaving the focal length unaltered.
3) If you still need to create a “render region” at any time I just use the workaround of modelling a vertical plane right in front of the camera with a hole in it and painted 0,0,0 black. I just move it or the hole around until the region I want is framed in the viewport. The black frame part of the plane will render, but VRay calculates it extremely quickly. It’s clumsy, but it works.
