I do this with all my projects just to see both my progress in learning UX/UI and also my thoughts. I've learned a lot by taking a look at what I thought was "amazing" then the next "amazing" and so on until I got the "perfect" design. For me, the #1 thing I've learned is to leave trying to perfect the design until the real data is available. It makes such a huge difference that I always, without exception, change some aspects of the design as soon as the data is displayed. Just my experiences over the years.
Thanks! It's was interesting to see how a 3rd party view made different connections, especially as some of your correlations never crossed my mind. Notably, the red header was a short-term random color test (I tried tons). The timeline was actually inspired by a dribbble shot I ran across for mac app. Souncloud certainly had minor influences, I do love waveforms :P (not sure what the "moment" was though).
EDIT: You did forget the Rdio phase :). I do love their UI. Unfortunately I couldn't find a way to get the cover look right with the complex titles and art I often deal with.
I too liked the gradiented look a lot, but I'll say having spent a ton of time with the site I love a current look so much more for a music site. Though rather than jumping on some train there was actually quite a few steps that brought me there. I tried a lot of things, but honestly it's just easier and cleaner to strip details when they are unnecessary, and I added some twists on it (I like having the flat bevel on the buttons).
Just thought I'd note that the influences often came from multiple not-obvious places and decisions, or even, maybe, just occasionally, perhaps, from me!