Le 05/07/2017 à 17:08, qingshu a écrit :
> Hello everyone:
> I have read the online Docs about Version 3.7 (current). But there is little
> about how the source code organized and how POV-Ray works. So is there any help
> to understand the workflow of the POV-Ray and the function of each code file. I
> want to use part of source code to simulation the moving camera.
> 
> Thanks!
> 
> 
My hints, hoping they can help:

the data structure is more and more complex as version increases.

There is a virtual front end (abstract) which is OS specific (name vfe),
making the whole thing into movement.

There is at least 2 steps:
* parsing
* rendering

Parsing is single threaded, and transform text input into internal data
structure.

Then comes the rendering time, section (blocks) of the final picture are
distributed from the thread in the front end to the multiple threads in
the back end. Once a block is done, the internal messaging transfer the
data from the back end thread to the front end thread which (might, it
can be disabled) store the message in the file to continue a rendering,
and store the actual data into either RAM or buffer file (there is a
setting to choose between RAM or buffer file). Once all blocks have been
rendered, the image file is computed from the data stored in RAM or in
the buffer file.

The camera(s), as they are described anywhere in the scene, get a
post-processing stage before the start of the rendering phase.

That was for the short and easy steps. You might have to add a photon
distribution step (more internal data) and a radiosity step (even more
internal data).

Post a reply to this message