Music
Hardometer

Musical Hardness Estimation based on Acoustic Features (Czedik-Eysenberg et al., 2024)

Instructions: Select the audio file you wish to analyze below. For optimal performance, it is recommended to use short audio excerpts (approx. 10 seconds). Alternatively, you can upload a ZIP archive containing multiple audio tracks to perform a batch analysis with CSV export.
Batch Processing Progress:
Processed: 0 / 0 Files
Calculated Hardness Model Value
--.--
8
7
6
5
4
3
2
1
0
EXTREMELY HIGH HARDNESS VERY LOW HARDNESS
Select an audio file and click analyze to begin.
This tool is designed for academic research. Low-level audio feature extraction is based on the Essentia, Librosa, and pyloudnorm libraries. Estimations are generated via an empirically validated regression model derived from psychoacoustic listening experiments (predictive performance on test set: Pearson's r = 0.908). For use of the Hardness model please cite: Czedik-Eysenberg, I., Wieczorek, O., Flexer, A., & Reuter, C. (2024). Charting the Universe of Metal Music Lyrics and Analyzing Their Relation to Perceived Audio Hardness. Transactions of the International Society for Music Information Retrieval, 7(1). Czedik-Eysenberg, I., Knauf, D., & Reuter, C. (2017). " Hardness" as a semantic audio descriptor for music using automatic feature extraction. In INFORMATIK 2017 (pp. 101-110). Gesellschaft für Informatik, Bonn.