The Importance of Dual Piano Performance Forms in Piano Performance in the Context of Deep Learning

Abstract

Abstract In the 21st century, computer technology covers human daily life, including the field of music. This paper uses a quaternion algorithm to describe the gesture posture during piano playing. Combined with the iteratively updated and extended IU-EKF algorithm, it realizes the fusion of piano playing gestures to fix the posture. The recognized piano playing gestures are output to the chord fingering automatic annotation board, the piano audio signal data are preprocessed and input to the 3D space, the annotation area is allocated, and the chord fingering features are extracted using the Boltzmann machine. Through spectral analysis and empirical investigation, we analyze the sound quality effect of the two pianos and the audience’s listening experience. The results show that in the spectral analysis of the two pianos, the time-domain waveforms of the pianos played using the gestures proposed in this paper have durations ranging from 0.75s to 1.25s, and the waveform graphs present triangular shapes, which are better in terms of sound quality. The listeners’ melodic memory of the songs played by two pianos is the best for the two pianos of Chinese art and folk songs, Chinese art and Chinese pop, with the average difference of 5.7899 and 5.6345 respectively. The two pianos form of playing can satisfy the listener’s need of listening to the piano songs to a certain extent.