Recognition of the genre of musical composition

Question

Sound processing

There are a lot of tracks, for each track you need to determine its genre, you can’t rely on id tags.
We came to the conclusion that the best solution would be to write our own application for genre recognition based on machine learning.
The application will be able to compose a description of the track - sound speed, tone, duration, etc., and classify it based on this information. For correct recognition, we will first train the classifier.
Another question is how to write such a classifier?
Need to create your own neural network? Or are there already some libraries for such purposes?
This is the first time I've encountered such a task and I would like to hear the opinion of those who have experience in this subject.

Answer 1

Read the article " How Yandex recognizes music from a microphone " to get started. There, among other things, there are many links to the primary sources of music recognition.

Answer 2

About seven years ago, Brodsky's paths were found downloaded, huge, by the way, the idea to create a defendant appeared, it was tried and tested on Parker's resource. With a stupid search.
The number of semantic combinations is limited, and this seems to be the key to the idea.

Answer 3

Can you tell me in what genre and direction the tracks are performed:
indila - dernière danse (scott rill remix)
summertime sadness - lana del rey (mnex remix)
In the end by nippandab
Apologize (jvstin remake)
And the like?)

Answer 4

You found the answer to your question and is it possible to contact you in some social network, discuss, talk about this topic, I'm very interested?