Method for extracting melody of counterpoint based on GPU

Abstract

The invention provides a parallelization method for extracting the melody of counterpoint based on GPU (Graphic Processing Unit), and the method comprises the following three steps: firstly, conducting spectrum transform and pitch value saliency computing to the music data on the GPU; secondly, constructing pitch contour by using pitch saliency on the GPU, getting the relevant characteristics of the pitch contour, and conducting voice detection by using the characteristics; thirdly, conducting frequency doubling removing and outliers removing to the remained pitch contour and finally obtaining the melody track of the counterpoint. The method provided by the invention is based on counterpoint, and can be applied to music without background music and speech sounds. According to the method provided by the invention, the melody can be extracted based on GPU, the extracting time is decreased from the second level to the millisecond level, so as to achieve the standard of real time application. Furthermore, the extracting of required hardware resource is greatly reduced, and the developing speed of the algorithmic is greatly improved, so as to provide the method with wide use value and application prospect in the field of commercial application and scientific research.

Description

technical field [0001] The invention provides an acceleration method for extracting music melody, and in particular relates to a parallel method for extracting the melody of multi-part music by using GPU. It belongs to the field of parallel application of computer science. Background technique [0002] Given a polyphonic piece of music, the task of melody extraction is to automatically extract the melody track of the music from it. The most commonly used melody track today is the fundamental frequency (F0) sequence of the music. Most of the current algorithms are based on the saliency value. This type of algorithm constructs the melody trajectory by estimating the saliency value of the pitch. The general structure of the algorithm is: first obtain the spectral representation of the signal, and then use the spectral representation to calculate the time of the saliency value of the pitch. Frequency said. Peaks of significant values ​​are considered as possible F0 candidates...

AI Technical Summary

Problems solved by technology

Due to the influence of background sound, the complexity of the melody extraction algorithm is high, the operation is slow, and it cannot be applied in real time, which limits its scope of application.

The slow running time and huge test library also make it difficult to test and improve the algorithm, and the accuracy rate is much lower compared to the melody extraction of MIDI music

[0004] The current solution mainly relies on distributed computing to solve the problem of slow operation, but the speed improvement is limited and the construction and maintenance of clusters require high costs.

In addition, some people use FPGA to accelerate the algorithm, but the programming in this way is extremely complicated and the development cycle is too long, which limits its development.

Images