There's certainly a indiescene, but it's more a bunch of individuals sharing their work and inspiration on the internet and on gaming conventions (pax, for example). I don't see how you can in any way compare that to Crytek's Crysis games (EA just bought the studio, like they did with many other studios).
Unfortunate enough, the "new, fresh and great" isn't entirely true. Since they don't have as much marketing budget (if they have any at all), they rely on word of mouth and being picked up by youtube let'splayers. That filters things, because not all games make it to the spotlight (and to be fair: some absolutely don't deserve it either). Result: games that have an unique approach or interesting mechanic are more likely to surface. Interestingly enough, this is somewhat the opposite if kickstarter is involved: money usually goes to throwbacks of old game genres or reboots of forgotten franchises (I already mentioned parkitect, right?).
In comparison, AAA-titles simply cost too much money to make for the creators to take risks. That's why there's emphasis on sequels and known and proven-to-work mechanics (especially ubisoft gets overboard on this last one). Shooters and 'action-adventure' games sell the best, so it gets made more. Horror and survival games are somewhat of a niche, so they are rarely ever the main component of the game.
But you're right on one thing: indies sometimes tend to look somewhat similar due to probably three things:
-pixelated or 16-bit art style. This is both easier on computing power, easier to make look good and sparks the nostalgia card. While this sometimes really adds to the atmosphere (eg shovel knight), sometimes it's just done for convenience (3D modeling is pretty time-consuming to do well).
-similar editors. Unity, game maker, unrealED and perhaps a handful other game building engines are powerful, easy-to-use and not expensive or even free (I remember getting a few in humble bundles

). So the majority of indie titles are done within those. I don't really think it's noticeable, but more experienced gamers who care to look into this (like totalbiscuit) disagree on this
-game designer is a profession, now. There are classes in which the do's and don't do's are actually taught. That, and the fact that most designers of today grew up playing video games (unlike the AAA-directors, I sometimes think), mean they have a more refined version of what makes "a good game". And I've got the idea that maintaining that similar idea is easier on a small group than in a large studio.