Audio processing with using Python language science libraries

摘要

The topic of the article is recognition of instruments and playing techniques of music for detection and correction of errors in a given music sample. It shows how to achieve characteristics of recorded sound and also how to compare amplitudes and frequencies of the same music piece, but played by different persons and also with using various instruments. For this aim the signal processing algorithms are used, which are available in standard Python libraries such as “numpy” or “scipy”. The key idea of the processing is detection of errors, but save playing technique and individual style of the player.