6-bit = 2^6 = 2x2x2x2x2x2 = 64 = adjacent shades are 1/64th brightness steps apart
8-bit = 2^8 = 2x2x2x2x2x2x2x2 = 256 = adjacent shades are 1/256th brightness steps apart
FRC dithering only uses adjacent colors, so at most, you're getting 1/64th luminance changes (1.5% flicker) or 1/256th luminance changes (0.4% flicker). Incandescent bulbs have a bigger flickerdepth than that due to the AC current, as the filament partially cools off during the zero-voltage crossings on the alternating current.
So generating 10-bit (1024 colors) via 8-bit plus FRC, it's pretty gentle all things considered.
While 6-bit FRC flicker is just about detectable, even flicker from inversion artifacts is even a bigger flickerdepth than that. That said, panels that have 6bit+FRC also often have inversion artifacts. Many people give the wrong wild goose chase thinking (incorrectly) flicker from inversion artifacts is caused from the FRC. Though, there are indeed situations where FRC flicker and inversion artifact flicker can interact with each other if the FRC pattern is poorly chosen. For an example of flicker from inversion artifacts, see
http://www.lagom.nl/lcd-test/inversion.php
PWM dimming is the thing you need to worry about more, because THAT is a 100% full flicker from total darkness to total brightness (essentially a squarewave flicker) rather than a gentle flicker.