2.Feature Extraction
Download as PPTX, PDF0 likes156 views
This document discusses various sound features that can be extracted for speech processing, including energy, tempo, pitches, zero crossing rate, spectrogram, and spectral centroid. Energy measures the loudness of a signal. Tempo estimates the speed of a musical piece in beats per minute. Pitches depend on the vibration frequency, with higher pitches for faster vibrations. Zero crossing rate indicates the rate of sign changes in a signal. A spectrogram plots the amplitude of different frequencies over time. The spectral centroid characterizes where the "center of mass" of the spectrum is located and relates to the brightness of a sound.
2.Feature Extraction
- 2. Sound Feature Sound features can be used to detect speakers, detect the gender, the age, diseases and much more through the voice.
- 3. Energy The energy of a signal is the total magnitude of the signal, i.e. how loud the signal is(Can use it as a classifier if someone anger or not.. . It is defined as:
- 4. Tempo (BPM). An estimate of the tempo in Beats Per Minute: the rate of speed of a musical piece or passage . Like a heartbeat, it can also be thought of as the 'pulse' of the music. Tempo is measured in BPM, or beats pe -can indicate that someone are arrested.
- 5. Bank of filters Digital filter-banks are an integral part of many speech and audio processing algorithms used in today’s communication systems. They are commonly employed for adaptive subband filtering. Another frequent task is speech enhancement by noise reduction
- 6. Pitches The Sound an object makes depending on how fast it is vibrating. When an object vibrates Quickly , high –pitched sounds are heard.
- 7. Pitches Low-pitched sounds come from things that vibrate more slowly. Humans can hear sounds of different pitches, But there are sounds that they cannot hear, human ears cannot detect very low-pitched noise known as infrasound, or very high-pitched noises, called ultrasound.
- 10. Zero Crossing Rate The zero-crossing rate (ZCR) is the rate at which a signal changes from positive to zero to negative or from negative to zero to positive. Its value has been widely used in both speech recognition and music information retrieval, being a key feature to classify percussive sounds. is an indicator function.
- 12. Spectrogram Ever heard of a spectrogram? It’s a 2D plot between time and frequency where each point in the plot represents the amplitude of a particular frequency at a particular time in terms of intensity of color. In simple terms, the spectrogram is a spectrum (broad range of colors) of frequencies as it varies with time.
- 14. Spectral Centroid The spectral centroid is a measure used in digital signal processing to characterise a spectrum. It indicates where the center of mass of the spectrum is located. Perceptually, it has a robust connection with the impression of brightness of a sound.
- 15. Spectral Centroid The spectral centroid is a measure that indicates where the “center of mass” of the spectrum is. Perceptually, it has a robust connection with the impression of “brightness” of a sound, and therefore is used to characterise musical timbre.