Analysis terminology

I was involved in a discussion about the difference between some terms that are frequently used: analysis, data processing, feature extraction, etc. To summarize my thoughts on how these terms are related, I made the little sketch below:


Rather than just storing it in my digital archive, I thought it might be useful for others, and could also hopefully lead to some interesting comments.

Music content processing

Dear child has many names. In this call for a special journal issue on music and robotics, I see the use of the word music content processing (MCP). I have been around in the larger music technology community for a while now, but I haven’t really thought of this as a concept in itself before.

Using Google as a reference, I see that “music content processing” actually returns 34 100 hits, so obviously it is being used quite extensively. As a comparison, “music information retrieval” (MIR) returns the double (62 100 hits). And here we need to remember that the MIR community has a society and an annual conference to help drive the number of pages.

This is quite clearly seen when we look at the results from Google scholar, which returns only 234 hits for ”music content processing”, but 7150 hits for “music information retrieval”. So from a scholarly point of view, it is quite obvious that MIR is much more established than MCP.

Here are some figures for other key terms/conference names (Google / Google Scholar):

  • “sound and music computing”: 174 000 /  1660
  • “new interfaces for musical expression”: 147 000 / 2300
  • “embodied music cognition”: 39 400170