This paper presents a time-domain technique based solely on simple operations, like filtering and scaling, to continuously track a sound descriptor correlated to brightness and pitch. In opposition to existing algorithms this approach does not rely on input framing or frequencydomain transforms, ensuring a better temporal resolution and the possibility to model analog-like implementations. In the first part of the document we present the details of our approach to brightness estimation; then we will compare CoBE to the brightness estimation implemented in MIR Toolbox; we introduce and define the concept of “Equivalent Brightness Frequency” (EBF) and finally we show how to exploit CoBE and EBF to obtain a rough monophonic pitch estimator or an efficient pre-processor for zero crossing rate (ZCR) pitch trackers.
Continuous Brightness Estimation (CoBE): Implementation and its Possible ApplicationsWeisberg Division of Computer Science Faculty Research
Citation InformationPresti G., Mauro D.A. “Continuous Brightness Estimation (CoBE): Implementation and its possible applications." 10th International Symposium on Computer Music Multidisciplinary Research (CMMR), Marseille (France), 15-18 October, 2013.