I've set up a test for a client that involves testing individual clips, comparing 8bit vs 10bit quality. I'm using Premiere CC to set up a timeline with a single clip and encoding out of Premiere using: H.264 and customizing the Bitrate settings to Target=8 Max=8 for the 8bit version of the clip. Then repeating with 10bit in the settings, getting two MP4s, one with 8bit & and one with 10bit quality.
As a secondary test, I'm also exporting out of Premiere a 4K UHD Mezzanine of the clip, using ProRes422, and putting it through an Elemental transcoder to 1920x1080. I've checked all the settings and I'm almost positive it's all the same. The Elemental transcoded MP4 looks way better than the MP4s out of Premiere: less banding which is what the test is supposed to show.
Apples to oranges - Premiere uses Apple's native (MOV, TERRIBLE quality) or MainConcepts(MP4, decent quality/speed) while Elemental uses in-house developed codecs (my bet is a fork of x264). There are a lot of variables that can be fine tuned in terms of quality over speed which makes comparing 8bit and 10bit a little bit muddy.
You may find x264 through the use of the command line app FFMPEG to be of better quality than MainConcepts if you want something to deploy to your client on the cheap. If you're on Windows then XMedia Recode is an excellent front end (site is in german, app is in english) that will expose you to a lot of those fine tune settings so you don't have to go through pages of command line options.