1

I am an economist by training trying out my hands lately in Machine Learning applications. I recently got interested in music information retrieval theory. With no prior experience in learning any instrument I am planning to proceed as follows

  1. Study basic music theory
  2. Understand the physics of music
  3. Get my hands dirty by statistical analysis of various musical pieces

In terms of softwares am proficient in R and Python as well as various stastical learning techniques for analysing data.

I was wondering if there are any good free resources available for

  1. Understanding music theory
  2. Understanding how data related to music can be analysed
  3. Codes in Python or r which can be a good starting point to look at
  4. Sample musical data for analysis
Sycorax
  • 76,417
  • 20
  • 189
  • 313
Rajarshi Bhadra
  • 165
  • 2
  • 9
  • 2
    For low-level approaches to analyzing music data (e.g. feature extraction), as well as visualization tools, the resources [here](http://www.vamp-plugins.org/) may be useful. In terms of understanding the higher level "semantic content", you might begin by looking at information related to the (proprietary) [Music Genome Project](https://scholar.google.com/scholar?q="music+genome+project") that under-pins Pandora. – GeoMatt22 Sep 12 '16 at 15:57
  • 1
    Even after more than 120 years, [Helmholtz](https://archive.org/details/onsensationsofto00helmrich) is a great read on the physics of music. – whuber Sep 12 '16 at 16:23
  • @whuber How can it be not about statistics, machine learning? This can be a little bit broad, but it is an interesting question and is definitely about statistics / machine learning IMO. – Metariat Sep 12 '16 at 16:39
  • @Materiat Because it's *also* and *primarily* about music theory, physics, and coding. This is far too broad to fit the SE format for questions. – whuber Sep 12 '16 at 16:42
  • I think it can be approached like any other data related problem: using whatever learning technique is most suitable for the purpose I am using it for – Rajarshi Bhadra Sep 12 '16 at 16:45
  • 2
    I think perhaps the question could be re-phrased into something more CV appropriate. I interpreted the "essence" of the question as trying to get at what conceptual/theoretical frameworks are used in practical music-related machine learning. If the OP does a bit more research on the sub-topics/jargon (e.g. [here](http://c4dm.eecs.qmul.ac.uk/musicinformatics.html)), as well as the CV guidelines linked above, then I think it would not be too difficult for them to a re-focus the question to something CV-appropriate. – GeoMatt22 Sep 12 '16 at 19:52

0 Answers0