Yes, shot-by-shot metadata will have a bigger impact when the content is brighter than what your TV can do and Inception constantly peaks over 1000nits and it's quite a saturated color grade (too much in my taste).
When the colors are brighter than your TV capabilities, they can look slightly desaturated in HDR10. The generated metadata does not change the contrast and colors, they are just compressed better in DV because the TV knows the shot brightness in advance, it can compress the highlights and gamut much more accurately than a static tone mapping curve which is what you are noticing.
For content within your TV dynamic range, you wont see much difference with HDR10 (assuming your TV is correctly calibrated)