When I hear a song I like, I can pretty much repeat it note for note when I whistle, and it conveys really the same feeling and meaning, but it's also a bit blurry on account of the details getting fuzzier like lyrics, specific instruments etc... but it feels a lot more authentic to the original memory/experience than anything I can ever visualize.
What I'm saying, is that I wonder if there's something special about how spoken/vocal memory sticks vs other kinds of memory.