Research On Color-Music System Base On Audio Spectrum Recognition

Audio signal processing has been a popular research direction in the computer field.This project applied the scientific method to explore the basis of music and color mapping through the study of the relationship between music and color,and establish the model of color music.The connection existing between music and color is summarized and analyzed,and the embedded experimental system is developed on its own to conduct experiments on audio playback,audio sampling,audio signal processing,and audio signal spectrum recognition.Theoretical research on the relationship between music and color is conducted to establish the mapping relationship between note frequency and color;the spectrum analysis algorithm is optimized and a program algorithm suitable for real-time audio processing on the embedded system is implemented;the spectrum recognition method is studied to determine the spectral characteristics of audio through the spectrum data and to determine the color pattern that matches it;the pulse width modulation drives the light mixing module,to achieve the output display of the color,in order to achieve the purpose of matching music and color together accurately in real time.The main contents are as follows.(1)A study of the relationship between color and music.Starting from the basic properties of color and music,we explore the connection between color and music by analyzing the characteristics of music and color.Starting from the theory of sound,we find a "twelve mean rhythm" of light according to the twelve mean rhythms of sound.By analyzing the relationship between music and color,a model of color music was established.Based on the color music model,we realized the mapping of music and color matching,so that we can match the color of the music played in real time.(2)A study of spectral analysis algorithms.Using spectral analysis techniques,we start with the fast Fourier transform,further extend it to the short-time Fourier transform,and then continue to extend it to the constant Q-transform,and finally implement a tone level recognition algorithm based on the improved constant Qtransform.The algorithm is specially optimized for the twelve-mean-rhythm to better identify the chromatic signals in the twelve-mean-rhythm.Since the algorithm is to be implemented on an embedded platform,it is specially optimized for the practical performance of the embedded system to reduce the computational effort and enable the recognition of fundamental frequencies in terms of time frames..(3)Development of the experimental platform.In order that the experiment can be carried out smoothly,the embedded experimental platform is designed and developed,in which the audio signal can be processed in real time.The embedded system has the features of high portability and easy portability.Therefore,this topic chooses to use the embedded system as the research platform.The experimental platform designed in this topic adopts dual-machine communication: the host machine performs audio sampling and the slave machine performs spectrum analysis.The requirement of timing is satisfied by the collaboration of dual machines.