Embedded Linux Environment, The Audio Quality To Improve The Methods Of Research And Realization

With the fast popularization of the embedded audio systems, people have a growing demand on a higher quality of audio. Currently, how to improve the performance of audio in embedded systems whose resources are limited by applying software techniques is a hot research topic.This thesis focuses on how to improve the performance of audio system by applying software techniques under embedded Linux. A new software solution is presented in the thesis which redesigns and implements the driver of the embedded audio system in kernel context. This solution puts emphasis on the three parts: audio resampling, mixing and DMA engine with memory management. Firstly, in order to improve audio quality, Kaiser Window lowpass filter algorithm is used to resampling audio signals with different frequencies which can reduce the noise furthest. At the same time, the algorithm is optimized effectively to double the sampling efficiency. Secondly, mixing algorithm based on mixing theory is applied for supporting multi-channel outputs of audio flows synchronously. Finally, DMA engine and memory management are redesigned and implemented by using dual buffer design to improve the system performance which meets the requirement of real-time and continuity of audio. During implementation, MMX assemble commands are used to optimize the key parts of the algorithm which can also improve the system performance.From the experiment result we can know that the efficiencies of resampling and mixing and the performance of the system are improved greatly which means the proposed solutions are feasible.