Research And Implementation Of OPUS Audio Codec Based On STM32

The mobile Internet in the present age, embedded system has got its rapid development, speech technology has been widely applied to the embedded system, and embedded system storage size limit, which makes the speech coding technology, is essential.OPUS audio codec support narrow bandwidth to the full bandwidth of speech and audio, but also can support variable bit rate, integrated real time protocol(RTP), loss concealment and low delay algorithm has good package. At present, the codec delay than other codec is low, quality than the other encoder, OPUS codec has been defined as the audio codec standard, its application is very extensive, including the Internet phone, high quality audio streaming media information, the OPUS will be used as part of the Web RTC application in the future network video chat. The overseas is the standard basis, hope to develop speech coding algorithm with higher quality, lower latency, but OPUS in some hardware(such as stm32) support is not enough, the transplant procedure is more trouble, relatively large workload, this paper is to realize the porting based on stm32.In this paper, a detailed analysis of the coding and decoding principle of OPUS, first introduced the MDCT transform and psychoacoustic model, CELP algorithm, then puts forward CELT algorithm based on the improved CELP algorithm, introduces the key technologies, the CELT algorithm is one of the most important algorithm in OPUS. And then put forward the main now embedded real-time operating system, and do a comparison, found that RT-Thread has more advantages, then a simple introduction about the system, and describes the system transplanted to the process on the stm32, completed the system transplantation and verify. Finally, through the two aspects of OPUS compression and decompression procedures process expounds the process of OPUS implementation, and encoding and decoding of waveform and spectrum before and after verification of the compression quality and effect through software Adobe Audition observation, the codec code of high quality low delay and support the band width, and can be widely used in the embedded system. Then, the system under the OPUS codec is proposed based on the system optimization method, and optimized memory space becomes small.