All musical audio feature extraction techniques require some form of signal processing as a first step. However, the choice of low level parameters such as window sizes is often disregarded, and arbitrary values are chosen. We present an investigation into the effects of low level parameter choice on different tonality estimation algorithms, and show that the low level parameters can make a significant difference to the results. We also show that the choice of parameters is algorithm specific, so optimisation is required for each different method.
Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member and would like to subscribe to the E-Library then Join the AES!
This paper costs $33 for non-members and is free for AES members and E-Library subscribers.