[michaelplogue]

I started doing some tests to see if the perceived "zooming" effect we have been noting is in fact true. I did a simple test by comparing the use of HDR's as the environmental setting from the plugin with the HDR converted to an MXI and mapped to a sphere. In order to compare them better, I set the environment intensities to the default (1), and did not modify the hdr when converting to MXI - so the images are a bit dark. I then brought them into photoshop and compared them using the difference modifier.

I've noted three differences:

1. The environment version is slighly brighter than the sphere-mapped version, though it appears that the reflection intensities are about the same. This could be attributed to the size of the sphere I used to map the MXI (50 ft radius)

2. The sphere mapped version was much more efficient in rendering.

Sphere: Benchmark - 264, Rendertime - 3m25s

HDR Environment: Benchmark - 105, Rendertime 8m35s


3. The HDR environment version does appear to come out a bit larger than the sphere mapped one - though not as much as I was expecting. (see difference comparison in the third image)

Sphere Mapped version



HDR Environment version



Difference (auto-contrast used to display better - black shows exact match, the brighter the pixel, the greatest difference)

[jespi]

Thanks for the test michael . Is very interesting the different between time of rendering . Maybe HDR with sphere maping is what maxwell need to competitive(speaking about render time of course) .
But still the different between background and reflexion channel are too much differents.

[michaelplogue]

It just occured to me that my test may not be accurate. Both the reflecting sphere and the mapped hdr/mxi sphere are centered at 0,0,0 coordinates. The camera is a few feet away. It's possible that MR centers the HDR environment always around the camera. So with my test, the mapped sphere is not centered around the camera (plus, the camera is also raised a few inches above along the z axis). These slight variations could account for the differences in the renderings.

Can't really test this rignt now since I'm in the process of breaking down my rendering computer in order to move it to a different room. As soon as I get it re-settled, I'll check it out.

The time and benchmarks were surprising to me as well. However, this may also be a result of a poorly conceived test on my part. Using the HDR environment settings, I had checked all options - to include refraction. However I didn't have an object in my scene that used refractions.

The new test will use a refraction object as well, in order to ensure the test is 'fair'......

[3dtrialpractice]

ya I am HOping that the env is alwasy placed around the camera thats how i designed my scenes here..

BUT SOO strange that render times are soo OFF.. I would TOTally expect the Sphere MApped version to be WAY slowwer.. What was the Sphere's POLY count?

MAybe the maxwell Env scene is Slower because it loads the same Mxi/hdr 4 times (cuz they can all be different maps so maxwell always loads 4 times instead of just loading the map once for all channels?)

I'm gonna look into thismyself.. IF the sphere mapped versionis better im gonna change my scenes over.. (the downside is that you cant create seprate files for reflections vs lighting...like using a ¼ rez lighting hdri w/ a high res reflection )