The Research On Digital Audio Effects Based On ARM Chip

Audio sound processing refers to the process of modifying or enhancing the sound.There are many scenes in life that have the need for audio sound processing.For example,the concert needs reverberation sound to make the singer's voice more pleasant.The speech scene needs to adjust the equalizer to make the speaker's voice clearer,or the studio needs to be right.A melody of the tune is trimming the pitch.Audio sound processing technology has been developed for a long time,and the classification is also quite complicated.It can be roughly divided into modified audio sound effects and compensatory audio sound effects.As the name implies,the former adds or enhances an effect on the audio of the source,while the latter compensates or masks the defects of the audio.With the hardware system equipped with the relevant audio sound processing algorithm,the above tasks can be completed in real time without delay,so that the sound of talking,singing or instrumental performance can achieve the specific effect and quality desired by the user.In the above classification,under the consideration of factors such as research significance and social value,each direction has been selected for research,including tube power amplifier analog sound effect and sound feedback phenomenon suppression technology:(1)For the simulation sound effect of the tube power amplifier,this paper takes the opposite research direction and the reverse modeling of the current mainstream forward modeling technology,and the core difficulty of the method lies in the decoupling identification of the Wiener Hammerstein model.In this paper,a neural network identification method based on feedforward loop is proposed,which is then tested using industry open data sets and compared with other algorithms in the industry.Finally,based on this,the embedded audio sound embedding algorithm of the tube power amplifier is realized;(2)This paper first studies the theoretical basis of the sound generation feedback phenomenon,and compares and analyzes the three mainstream suppression processing techniques: the frequency shift method,the adaptive filter method and the notch method.Through the comprehensive consideration of the three methods based on suppression ability,sound quality and impleme ntation cost,the notch suppressio n method was selected as the object of further research and improvement.Then,for the defects of the current mainstream notch method,the discrete-spectrum center-of-gravity correction algorithm,the zero-crossing detection algorithm,and the eigenvalue algorithm based on recursive filter are designed to improve the time domain and frequency domain respectively.Finally,the embedded algorithm of the improved notch method is completed,and the experimental simulation is ca rried out.The simulation results are analyzed.(3)In the end,this paper compares and analyzes the specific hardware scheme from the aspects of hardware cost and computing power,and chooses the three-chip scheme of RT1052 and RT1021 as the algorithm pro cessing core and STM32F103RCT6 as the data transmission processing core,and gives The key part o f the hardware scheme and the final finished diagram are presented.Then based on the hardware scheme,the corresponding main program processing flow is design ed,and finally the hardware platform that can meet the requirements of the algorithm designed in this paper is completed.Finally,the main work of the above paper is summarized,and the areas that can continue to improve and the prospects of related tech nologies are analyzed.