Posted by Wade Walker on Mar 30, 2014; 6:56pm
URL: https://forum.jogamp.org/Rendering-one-scene-multiple-times-per-frame-to-a-frame-buffer-tp4032079p4032080.html

My advice would be to start with a working example code from the web that uses FBOs, then gradually change it to do what you want, while keeping it working at every step. OpenGL codes can be very painful to debug if you've just got a bunch of non-working code and no idea why it doesn't work

Is your processing of the rendered image going to happen on the GPU or CPU? If on the CPU, it might be faster to only create one FB to render into each time, and just copy it back to the CPU for processing before overwriting it. If you're going to process it on the GPU (via OpenCL-OpenGL interoperation, for example), you still might just want to create only one FB for the sake of speed (or maybe two if you're going so fast you need to double-buffer).

It sounds like you're working on the pose estimation problem (http://en.wikipedia.org/wiki/Pose_(computer_vision)) -- are you a computer vision researcher perhaps?