There is no .blend attached…
What is the resolution of that environment map? Remember that this texture has to span the entire “world” - which is pretty big. So unless that environment map is huge, it will of course look fuzzy, as it is stretched the full 360/180 degrees around your scene.
That texture is only 736 x 368 pixels aka tiny.
Get something in the region of 16.000 x 8.000 pixels and we can start talking about environment sharpness…
This environment map e. g. is 10.636 x 5.318 pixels and still is not entirely sharp:
The jpg image is SO small and has SO many artifacts that even in the very unsharp enlargement, you can STILL make out the jpg artifacts ( the blurry “squares”)
Use a huge environment texture. This takes a lot of memory.
Use a wider field of view camera. Zoomed in makes it worse than zoomed out. Ortho would five a single pixel from image.
Use a background plate, often found together with the environment map, but tonemapped and unwarped.
Use a blurred version for camera only and call it “camera depth of field”. Can also “blur” the original by noising up coords. This is my favorite as the image is already in memory, and help separate the model from the background.
Don’t use/show background for camera at all.
On CUDA, memory usage is a big deal, and I have many big environment maps I have stored in “too big to be useful” folder :). Use small ones if you can for lighting and reflections (lighting isn’t crucial, but for reflections find a relatively flat surface and use the highest camera zoom you think you will use in the various shots), sun lamp for sun if needed (pain to match, but you don’t have to locally fix the sun brightness which is kinda messy), and point #4 below for the camera version.
CarlG, I used a background plane and unwrapped it using the texture image. I am new to using background imagery, so I will research what tone mapping is. Also, I am going to try an HDRI image.