Published Paper Details:

: This paper presents a system that embodies the essence of sonic detective work. Like a musician, it first extracts certain features from the sound, identifies individual instruments, checks the notes or sounds they play, and then synchronizes the notes with the music. The end result is like a beautiful piece of music that takes the messy sounds and organizes them into something pleasant. It could be a new music mix or music score. People can listen to it and have a great time enjoying the music. This project is designed for listening to polyphonic music played by a variety of instruments together, and it does a lot of things well. For the purpose to advance the musical field we are developing the system that is capable of identifying instruments from polyphonic instrument sound from the feature extracted using Kernel PCA (Probabilistic Component Analysis) ,the classification is performed using SVM (Support Vector Machine).This allows system to efficiently identify the instruments. Then the pitch is detected from the instruments identified using Cepstrum method which is capable of detecting pitches from detected instruments. The detected pitches are then synchronized to provide a synthesized audio file as output using HMM (Hidden Markov Model) and DTW (Dynamic Time Wrapping).