Mel Frequency Cepstral Coefficients (MFCCs) are a representation of the short-term power spectrum of sound, commonly used in procesamiento de audio and reconocimiento de voz. They are derived from the Transformada de Fourier of a signal, capturing the frequency content in a way that mimics human perception del sonido.
El proceso de obtención de MFCCs implica varios pasos. Primero, la señal de audio se divide en cuadros superpuestos, y cada cuadro se ventana para reducir la fuga espectral. Luego, se aplica la transformada de Fourier a cada cuadro para generar un espectro de potencia. Este espectro se mapea luego a la escala Mel, que es una escala perceptual de tonos. La separación en la escala Mel está diseñada para reflejar la forma en que los humanos perciben el sonido, enfatizando las frecuencias bajas mientras comprime las altas.
After mapping to the Mel scale, the logarithm of the power spectrum is taken, followed by the application of a transformada discreta de coseno (DCT). The resulting coefficients represent the short-term power spectrum in a compact form, with the first few coefficients typically containing the most relevant information for tasks such as speaker recognition or phoneme classification.
MFCCs have become a standard feature set in various audio and speech processing applications due to their effectiveness in capturing the characteristics of the human voice and other sounds. They are widely utilizados en modelos de aprendizaje automático for tasks related to speech recognition, speaker identification, and even music genre classification.