Music melody generation method based on pitch contour curve

Abstract

The invention relates to the technical field of music generation and specifically relates to a music melody generation method based on pitch contour curve. The method comprises the following steps that 1, long-term structure information of the pitch profile curve is extracted in a frequency domain, the long-term structure information comprises a low-frequency part in a frequency domain sequence ofthe pitch contour curve, and a long-term trend rule of melody is reflected; 2, through fitting the long-term structure information by using a neural network with label control, long-term structure information corresponding to the label is generated; and 3, another neural network is trained by using the long-term structure information and the melody length information of the music data, so that the music data has the capability of speculating the melody length information according to the long-term structure information. According to the invention, the music melody with a controllable long-term structure is generated by using the frequency domain characteristic of the pitch contour curve, so that the music generated by the method is closer to the real music distribution than the music generated by a long-short-term network.

Description

Technical field [0001] The invention relates to the technical field of music generation, in particular to a music melody generation method based on a pitch contour curve. Background technique [0002] Music generation has always been a direction that people are constantly exploring in the field of computer art. In the early days of computer development, people began to use traditional algorithms to generate music. In recent years, there have been more and more attempts to use deep neural networks for music generation, such as long and short-term memory networks, adversarial generation networks, convolutional neural networks, and improved variational autoencoders. The performance of short-term music generated by these networks is quite excellent, but the research on long-term music generation is slightly insufficient. How to make the melody of the generated long-term music have a reasonable phrase arrangement, and there is a satisfactory sequence and stable transition between di...

AI Technical Summary

Problems solved by technology

The performance of short-term music generated by these networks is quite good, but the research on long-term music generation is slightly insufficient

How to make the melody of the generated long-term music have a reasonable arrangement of phrases, and there is a satisfactory order and stable transition between different chapters, there is still no good solution

Images