Here’s the thing: I wrote a fisheye camera and created 5 scenecapturecomponent2d of the same size. For each frame in the rendering thread, I will read the render targets of these 5 cameras into FColor arrays into the memory, and then construct a new FColor array of the same size based on my mathematical model, and finally send it out or save it to local.
I did this and found that it took a bit long. So I have the following questions
- Can I put those pixels calculate in the pixel shader?
- If I place those pixels calculate in the pixel shader, is that mean I save the time of copying 5 camera data from the GPU to the memory?
- Will the GPU do the calculation work of these pixels faster?
- If I want to write these tasks as a pixel shader, how should I do it? I have never written a shader. How can I take the render targets of 5 cameras as input and get a new render target based on special calculations? Because I briefly looked at the shader course titles, they all seem to be related to post-processing. I didn’t find the function for render target as input to pixel shader.
- All I want to do is to put pixel calculate work on GPU, and see will get a higher frame rate than before.
HELP ME PLEASE AND THANKS A LOT!!!