Hey,
We are doing some internal tests for checking the performance of Corona 6 and another 2 engines with the exact same ArchViz interior scene. We will also try to repeat tis kind of exercise with one of our exteriors. We have no big complains about Corona, we think it is very fast and consistent with most of our interiors, but we had too much curiosity about how it compares to another 2 engines, both CPU and GPU.
We have succesfully managed to covert the scene into these 3 engines, it will never be 100% exact, but they look really similar and it is good enough for us. This is not a scientific experiment, so no need to have super precise results. However we have faced some issues when trying to do comparable renders that stop at the aprox. same quality, so we could compare render times. In these 3 engines we have the option to fix a noise threshold that automatically stops the render at a given noise %. The surprise is that when we visually compare the results, Corona render looks much cleaner than the other renderer image. For example, we rendered one of the interiors in Corona at 2k, with a noise limit of 2%, and the coverted scene in Vray CPU, with a noise threshold of 0.02 (I believe this is the same % as in Corona?). VRay CPU render finished faster than Corona, a noticeable difference, something like 30%, but when looking close at them, Corona result has clearly less noise overal, and imho the noise is more even across the image, maybe causing it to be less noticeable to my eye in the worst areas?
Anyone could help me to be able to tweak my settings in both engines so that the time comparison is as fair as possible without eyeballing too much? Like using passes or subdivisions in VRay, or whatever...
Thanks in advance
