Synthesis of Wind Instruments using BiLSTM and Gaussian Mixture Model

In this work, a source filter model incorporating the BiLSTM and the Gaussian Mixture Model (GMM) for the synthesis of woodwind instruments is presented. Magnitude-modulated and pitch-synchronous impulse signals are the sources. The filter is converted by the correspondent DCT coefficients, divided into the low-frequency and high-frequency parts. The high-frequency part is modeled by the GMM. The BiLSTM recurrent networks are used to predict the low-frequency part DCT coefficients of the filters. The proposed method can synthesize realistic and expressive tones and breath noise as well when compared to the conventional Digital Waveguide Filter-based method.