Spotify: please expose the essential parts of a song's waveform


Spotify: please expose the essential parts of a song's waveform


Spotify: please expose more of a track's waveform than the first 30 sec. 

There is a research community that wants to develop more powerful

song-waveofmr analytics than exists today.  For example, see Bill Protzmann's

post 3 weeks ago on Emotion Data.  This could be powerful information,

developed for you FOR FREE!!!  There are a number of ways to do this:

1) Develop a Python package that exposes the essential parts of a song: intro,

verse, bridge, chorus, break, ending.  You don't need to expose the entire

song, just its essential parts, excluding repetition.  Even just 5 seconds of

missing data between parts is enough to defeat pirating.

2) Watermark the songs by user (i.e. generate a watermark that is derived from

the user's SpotifyID).  As analytics are robust to watermarks, aggressive

watermarking is warranted.

3) Restrict access to qualified practitioners (academics, research-labs,

maker-spaces, etc).  Erik White's wrote a related post, "License for academic

research use".

1 Reply

Re: Spotify: please expose the essential parts of a song's waveform

Casual Listener

Hey so I'm just now starting to explore the possibilities of using Spotify's audio analysis to consistently determine the start times of (at least easy) choruses in songs. After reading through the API and seeing this visual demonstration right here:


I wondered if it might be possible to analyze those sections for similarity in order to determine which ones might likely be choruses?


Based on your thorough proposal above - I'm going to assume the task that I wish to complete is not possible just using information about sections that the glitch app demos - and that it would be more prudent to use more advanced tools to analyze the entire waveform (which we do not have access to) to solve this issue?