Building an Aircraft Carrier... First
Where the above music analogy breaks down is that composers are often able to write music primarily alone. While that is true for a subset of videogames, it is not the case for the size and scope of videogames that you've identified as your target.
If someone is interested in making a boat and decides that the kind of boat they wish to design is an aircraft carrier, that's necessarily going to require a massive scale of teamwork, money, experience, earned trust/credentials, compromises, and a lifetime of work leading up to that point. It either won't be the first ship they work on, or the rushed result will involve so many compromises that it will not come anywhere close to what is typically expected of this ship type.
While it's true that we don't have many of the same material constraints and labor challenges as ship builders, what we are unavoidably faced with are literally millions of design, artistic, and technical decisions, from the very large to very small, which we have to make about everything in the game.
Faced with all of those decisions, making them well at every step either means having the experience from enough variety of past projects to make informed tradeoffs between alternatives (which is why teams generally want professionals experienced in videogame making, not just people who "can program" or "know Photoshop"), prototyping like crazy to find even a few small innovations that work, or adhering fairly strictly to conventions.
However, note that even companies filled with experienced, full-time professionals creating massive games that do adhere to most conventions of their genre still manage to go through $20 million just to get their game done and complete. These companies are in the business of maximizing profits and minimizing costs, with both internal and external pressures to figure out ways to do more and better work with the lowest possible expenses.
It still costs them $20 million or more to make what may appear to be a formulaic and relatively straightforward followup to an existing franchise (note that the actual changes and additions involved are often quite complex, but the differences may be difficult for someone who is not on the team to fully appreciate).
An Idea is Only as Good as Its Execution
One of the other common themes that arise in advice from experienced developers for others just starting out is along the lines of "ideas are a dime a dozen" or "no one cares about your great game idea." This is in regard to what you're calling your "secret" that you suggest will, alongside world building, set your first game apart from Call of Duty and Uncharted.
My entries presenting my take on this include Stop Arguing About What Makes a Better Game andStart Before You Have an Idea, which covers an idea I often explain to others as:
"I'm going to make the world's best painting. I have an idea way better than the Mona Lisa. That one's just a picture of some girl."
Most successful games aren't a particularly genius idea. Conceptually they're usually pretty blatantly derivative but are polished in their execution to an amazing degree. When the idea is at least unusual, perhaps one that has been done before but not well enough yet to achieve widespread commercial significance and recognition, at best that concept serves as a marketing point to get people talking about it. But people generally don't keep playing a game, nor recommend a game to their friends, based primarily on whether a game has an interesting idea, unless it also has world-class execution and refinement. Provided it's got world-class execution and refinement, an interesting or original idea is often not even needed.
Pick A Fight You Can Win
Sometimes there's a confusion that because one or two indie games succeed financially in a genre previously thought dead (point and click adventure, for example), or with a very weird type of game (much of what's found in IndieCade or IGF), those games have been proven to be economically viable after all.
The underlying misunderstanding though is that what may be a tremendous amount of money when split only 1-8 directions among a small team of indie developers working out of their homes or a tiny shared office does not even show up on the radar of the kind of money that AAA companies burn through in only a few months of salaries for massive teams comprised entirely of experienced developers, legal/business personnel, and support staff along with paying for sprawling office complexes, retail distribution arrangements, television/billboard advertising budgets, etc.
Part of the way this all holds together is that even though a few individuals (out of the many, many more trying) might make enough money to pay their own costs and then some, it's still often far from being something that the big studios can take seriously as a business opportunity.
In 2010 I made a strange iPad entertainment app that paid my rent for 18 months. That registers as a medium success - weirdly putting me somewhere in the upper percentile in terms of return. While it clearly didn't make me wealthy (my rent at the time wasn't much), many more developers lose considerable sums of money or generate very tiny revenues in the range of fewer than dozens of dollars from ads, virtually non-existent sales, etc. Also, as a fair warning: the App Store was definitely a lot less of an over-saturated mess in 2010 than it is now. (Speaking of which, back in 2008 I developed a game for the app store that earned considerably more than that... for the publisher, which is how I learned a lot about being the weaker party financially in a negotiation.)
Anyhow, for perspective, a massive publisher pays their CEO more money than that 2010 app's total earnings in what calculates to about 24 hours of their time. That's not including the salaries of 9,000 other people, the costs associated with massive facilities, or the remaining mountains of money expected leftover for annual profits, etc. They have truly staggering ongoing costs to keep up with, which means they can't afford to take risky chances on little weird stuff, or on genres that are no longer desired by mainstream players. They deliver on what they have solid evidence to think will make a sizable return on their investment. Projects that don't meet that goal get eliminated. They're a business.
This is why when we see advanced indie developers succeeding commercially, it's usually not by doing better at what the big companies are in position to do well, but instead from some combination of strange projects (with regard to their execution, often not in the main idea; i.e. tons of personality in an otherwise relatively simple experience), retro genres that the AAA companies haven't been able to justify for 10+ years due to the shrinking level of market demand for such games, or the intersection of both (meaning tons of personality in a retro genre that has otherwise lost most of its market significance). With the rare exception of some mod teams that quickly get swallowed up into the machine and vanish for years doing polish work, indies stick to guerrilla tactics in niche domains that the huge companies can't justify meddling in.
It's not a matter of purely creative vision that Minecraft is built out of 1-meter cubes textured with pixel art, instead of having the look and feel of Uncharted or Call of Duty. It's small team resourcefulness.
(For more along these lines, read up about Blue Ocean / Red Ocean Strategy.)
Minecraft is basically 3D pixel art.
Funding Will Affect Design
Getting funded does not mean that you have the freedom to design your perfect game. When money gets involved, it tends to influence what gets made, even when sometimes only out of some crude evolutionary feedback loop where the games that do well in a given market space tend to soon be followed by variations from competitors.
But what about something like KickStarter, in which customers prove that they're there before development even begins? The space is filled with people that, because they've never had to work with a real budget before, tend to not know how to schedule and plan a commercial project.
If KickStarter sounds like freedom, it's worth looking up Code Hero, which raised $170k, well past their $100k target. I'll spare you the web search, and share this update from last year. Effectively, they ran out of money long ago and are now an entirely volunteer effort. Here are their KickStarter updates.
Where I'm Coming From
Although I started as a hobbyist, returned to being a hobbyist, and help other people get into videogame making as hobbyists, what I'm suggesting about the nature of huge commercial-scale videogames is not just guesswork.
I've worked professionally as a game designer on the size of games and teams that you're talking about: games that take years, 80-200 full-time professionals, and tens of millions of dollars. I've also worked professionally as a game designer at other commercial scales: at a casual games start-up, in a gameplay R&D company, and as an iPhone game developer both independently and through publishers. I am familiar with at least a cross-section of commercial videogame development.
More importantly to this exchange I have also, with zero budget, served as a lead or solo developer on more than 85 completed freeware games with 3-18 month production schedules, and hundreds of same-day playable prototypes. Some of those were for class projects, some were created in clubs that I established, and many were just developed on the side.
Being in a position to compare those professional and nonprofessional experiences, I can say with total certainty that I have found infinitely more creative freedom in the projects that have no budget or funding than I've seen in working on games with a big budget.
That applies no matter whether I was being paid steady salary, hiring/paying/leading a team of others, funded by outside investment, or even when completely self-funding projects and theoretically having total directorial control. Games that cost money to develop have a certain obligation to prioritize earning (at least) that money back, which places certain inflexible outside demands on the work.
Make Finished Games. Without Funding.
There are ways to make videogames that don't require getting anyone else's approval, taking anyone else's money, or handing over veto control to your design ideas.
It takes time and a lot of work. There's no easy shortcut to it. But it is a way.
That way is to start small, building your way up from modest beginning principles based in time-tested classic designs, toward the point where you're prepared to do your more grand and elaborate game ideas justice, so as to have to ability to realize them without making compromises in service of financial obligations.
In other words, the way is to make videogames for free for awhile, incrementally advancing your skills and design understanding with each new project.
If however you choose to still push on the route of seeking substantial external funding, especially without any track record yet of completing any simpler games, I wish you only the best. Good luck.