Quick question about Mixing/Normalizing

[vince94]

Over the past few years, I've been making a SNES-style soundtrack mod for Cave Story+ with OpenMPT in my free time. Since all of the songs are finished now, I was able to successfully convert the exported FLACs into seamlessly-looped Ogg Vorbis files, which the game engine can play back. However, when you hear it change from song to song in-game, it sometimes feels like some of the songs are too loud or too quiet in comparison to each other.
I'm surprised that this is happening, since I made sure to click Normalize Output on every export. I don't know much about mixing and mastering, or if those terms are even appropriate in the context of tracker music, but does anyone have any experience with this kind of issue, or know what I should do to further even things out?

[Saga Musix]

I'm surprised that this is happening, since I made sure to click Normalize Output on every export.

Normalizing is not the same as dynamics processing. If one of your songs has one very large peak somewhere but apart from that is rather quiet, then this peak will remain the loudest part even after normalization, so the rest of the song will stay quiet too. You can counteract this by manual or automatic dynamics processing. Manual processing requires to adjust the global volume of your songs to make the loudest parts more quiet. This is a good idea in general, but it's also a bit of work. The alternative is to use a dynamics processor, e.g. a master limiter or compressor effect. There are thousands of such effects available as VST plugins, but you shouldn't overdo it (or you will be part of the loudness war).