|
|
Thanks Le!
I just posted the message in the other group and your advice was quite helpful
in tracing down the variables I needed to dump camera parameters!
Thanks a lot!
Ankur.
Le_Forgeron <jgr### [at] freefr> wrote:
> Le 11/07/2011 20:21, handos nous fit lire :
> > Hi everyone,
> >
> > I wanted to obtain depth-map from POVRay 3.7 and I figured out that I could get
> > the output from TraceRay function defined in (trace.cpp) and store it in array
> > and then write it into a file. I was doing it very conveniently after the
> > renderTask.AppendTask( new TraceTask(..)) function call defined in view.cpp's
> > StartRender() function. However when the image takes a bit longer time to
> > render, the function returns from StartRender() while already running the
> > threads in TraceTask. I tried to go through the code to understand at what point
> > can I write the depth-map array but I totally got confused with backend frontend
> > thing. If it's not very difficult to figure out could anyone please tell me
> > where exactly can I write the depth-array (already stored) into a file at which
> > point where all threads have finished.
>
> Find the moment the metadata are written in the output image.
> Who call Image::Write in base/image/image.cpp ?
> (hint, it's in ImageProcessing::WriteImage, in frontend/imageprocessing.cpp)
>
> >
> > Also, I'd like to write the camera parameters into a file too, once you specify
> > the parameters into the camera{ } object in povray file, where exactly it
> > applies all the various transformation defined within the camera{} Object and
> > latches the final camera_position, camera_up and camera_direction? As an
> > instructive example
> >
> > camera{
> > location <x,y,z>
> > direction <a,b,c>
> > up <ux,uy,uz>
> > matrix < val00, val01, val02,
> > val10, val11, val12,
> > val20, val21, val22,
> > val30, val31, val32>
> > };
> >
> > After applying the necessary transformation, where exactly does it store the
> > final direction, up vector and position vector in the code?
>
> I have that information (used for a camera_... access patch, not in 3.7)
> The main trick is that you might have one (usual) camera, or more
> cameras (for a special mode of rendering: same scene, many camera)
>
> So, it's either sceneData->parsedCamera or sceneData->cameras[idx]
> + .Location,
> + .Direction,
> + .Right,
> + .Up
>
> I guess you can match the ends.
Post a reply to this message
|
|