Hi though I’m not advanced developer the question seems is quite advanced
We have a big depth texture covering 360 of view field.
as well we have big background texture of the same size and
also camera which has a limited FOV and rotaes in our world and some geometry objects rendered over background texture wich are depth sorted with the use of depth texture values.
for the small static scene ( no camera rotation in world…) the following sample works great see bufferregion from http://www.delphi3d.net/listfiles.php?category=6
( actually we use C and not delphi but this sample on the net is on delphi)
but when it comes to change depth buffer
completely on frequent basis ( for ex fast rotation of camera which observes 360 field of view with it’s limited FOV) the approach fails.
as complete update of depth buffer slows down the system significantly …
the next possible approach to try is that to use our depth texture with shadow texture extension
from camera view and com-
pare depth at pixels with
value in map using ‘features’ of shadow map sorting …
but before the try the question is if this approach will give a requireded speed
or accuracy ( if shadow maps and depth textures are good for precise depth sorting for as they might have just only 8 bits accuracy ? )has anyone expereince?
are there subtle points here?
are there other approches which could be used such as using texm3x2depth or texdepth 1.4 pixel shaders instruction ( with the precision higher than 8 bits)
are there samples how to use pixel shaders them for such tasks ( C/C++) ?
So I would be grateful for tips to approach
to develop fast rendering app with the use of hight resolution pre computed depth textures
which cover entire 360 of scene
for depth sorting of local camera view?