将MIDI文件自动分割为数据集的短语 [英] Automatic segmenting of MIDI file into phrases for dataset

问题描述

我正在做一些研究，包括首先建立一个短的音乐短语的数据集。理想情况下，这将基于一些机器学习算法来进行，以猜测音乐的结构并将其分成短语。不幸的是，我还没有发现这样的算法的实现，或者甚至没有这样的算法存在的明确证据。

没有这样的算法，那么最好的事情是将每个MIDI文件分成2-3个栏的组。我还需要检测和删除空段。有没有一个简单的方法来自动化这个？如果我不需要这么多数据，我可能会考虑用手做。

我没有任何特定的语言。我看过music21，它似乎支持阅读和编辑MIDI文件，但我还没有找到我正在寻找的东西。
谢谢！

解决方案

Robert Rowe的书机器音乐家（MIT Press）有一章涉及不同的方法将音乐数据分割成短语和检测模式。 （显然，有更多的可以适应一个SO的回应..）

I am doing some research that involves first building a dataset of short musical phrases. Ideally, this would be done based on some machine learning algorithm to guess the structure of the music and split it into phrases accordingly. Unfortunately, I have not found an implementation of such an algorithm, or even any clear evidence that such an algorithm exists.

Without such an algorithm, that next best thing would be to split up each midi file into groups of say 2-3 bars. I would also need to detect and remove empty segments. Is there a simple way to automate this? If I didn't need so much data, I would probably consider doing it by hand.

I am not tied to any particular language. I have looked at music21 which seems to support reading and editing MIDI files, but I haven't yet found what I am looking for. Thanks!

解决方案

Robert Rowe's book 'Machine Musicianship' (MIT Press) has a chapter dealing with different approaches to segmenting musical data into phrases and detecting patterns. (Obviously, there's more to it than can fit in a SO response..)

这篇关于将MIDI文件自动分割为数据集的短语的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！