Independent blind A-B tests have showed that about 30% of listeners are albe to distinguish 16bit sound from 24 bit sound, and 0% of listeners were able to hear any difference between 24bit and higher bit depth audio. There you have it.
As for 16 vs 24bit, the larger dynamic range is a major benefit, quiet sounds becoming audible and / or much more detailed (due to DAC computation bias towards the high end of amplitude range). Remember, this is not only for render/mixdown and playback but all processes (recording / use of samples, plugin processing, outboard gear, converters) throughout the entire production. Main disadvantages use to include size, but actually less so in regards to file storage, more so related to lower realtime file transfer rates (from your drive to your DAW through USB, FW400, etc.) and realtime file processing (playing back 50 tracks in your DAW at 24 bits were more demanding, so was 24 bit processing in your session). Which are not real concerns anymore, as long as you have a decent multi-core processor and FW800, USB3 or TB connections and 7.2K rpm or faster HD.