“How have game file sizes become so big? Is it just the graphics? Just the high-fidelity audio?”
Essentially, yes. 100 KB of OPL instructions can play hours and hours of music, but the same amount of space can only hold 5 seconds of MP3 audio.
The same sort of thing can be seen in graphics, where a single texture for a modern high resolution 3D model can be larger than all of the graphics in an old low res 2D game combined.
The code itself is larger as well, but I don’t think that’s nearly as significant a factor as the audio and visual data.
Here’s a nice article about this very subject with some interesting quotes from game developers: https://www.digitaltrends.com/gaming/why-are-video-games-so-big/