In this chapter, we aim to apply our DFT-approach to the well-known problem of automatic classification of musical style. In fact, one of the main goals of applying persistent homology and Topological Data Analysis to music is to provide some algorithms that would be able to “recognize” the style of a given music piece by analyzing the associated family of barcodes. There is already some work on this topic, and we can cite the famous article by Bergomi et al. (Computational Topology in Image Context. Lecture Notes in Computer Science, vol. 9667, pp. 88–100. Springer, Cham, 2016) which is a precursor of the subject. In this chapter, we propose a new way of approaching automatic style analysis by combining the DFT together with persistent homology. We will start by detailing the strategy, which consists of transforming a barcode into a family of points in \(\mathbb {N}^2\) , and computing statistical features on the length of the bars (Mean, Standard deviation and Entropy), as suggested in Mijangos et al. (Musical stylistic analysis: a study of intervallic transition graphs via persistent homology, 2022. https://doi.org/10.48550/arXiv.2204.11139 ). Therefore, we will select several MIDI files from of different musical styles, starting from Heavy Metal to Baroque, and compare them by clustering in \(\mathbb {R}^3\) . All the musical data we are going to use are listed in a database which is available on the following dedicated web page: https://math-musique.pages.math.unistra.fr/midi.html .