Ok, it took me a while but I have finally figured out what's going on with the slow initialization of your game. Two things: first, you should decode the sound effects into raw byte slices, but not the whole BGM. Loading the whole BGM at once is overkill. Second, the ogg file's sample rate is 44100kHz, as most music distributed in the universe right now. You have 44000 in the project instead, which leads to having to resample all the audio you load. For the long BGM this is a massive pain.
When you see such a slow initial load you should definitely suspect that something is off, there's no reason why a game like this should struggle so much or even have an initial load screen.