Audio Feature Extraction is a technique to extract descriptive data from audio for use in a wide variety of applications. Meyda is an audio feature extraction library for the web audio API, providing a wide variety of audio features including spectral statistics like centroid (analogous to "brightness"), flatness (noisiness), kurtosis (pure pitchiness), as well as metrics like loudness (of various types) and Mel Frequency Cepstral Coefficients, which are used in speech recognition.
This talk covers the story of Meyda's development, and will attempt to provide inspiration for what you can do with this awesome tech!
Audio Feature Extraction is a technique to extract descriptive data from audio for use in a wide variety of applications. Meyda is an audio feature extraction library for the web audio API, providing a wide variety of audio features including spectral statistics like centroid (analogous to "brightness"), flatness (noisiness), kurtosis (pure pitchiness), as well as metrics like loudness (of various types) and Mel Frequency Cepstral Coefficients, which are used in speech recognition.
This talk covers the story of Meyda's development, and will attempt to provide inspiration for what you can do with this awesome tech!