The renderers integrated in Cinema 4D always render internally using 32 bits. These color values must be interpreted in order to convert these to 8 or 16 bits for commonly-used output devices. A series of algorithms that use different methods are used to do this.
It’s difficult to recommend which operator should be used when. It’s best to test which one works best for the scene at hand.
Generally speaking, the Rheinhard method is the most flexible method. It is based on photographical principles.
Note that Tone Mapping permanently modifies the image. If you are planning to perform a color correction in another application (e.g., Photoshop, After Effects, Nuke, etc.) you don’t have to perform tone mapping in Cinema 4D. Remember that images need to be saved with 32 bits if you want to get the full dynamic extent of the image.
Note also that Tone Mapping does not work correctly with the ProRenderer’s bucket rendering because the entire image must be available for Tone Mapping to work correctly. For the same reason, the first iterations can deviate greatly because Tone Mapping can react strongly to large changes in noise in the beginning.