Research And Realization Of DTS Decoder Based On DM642

Digital audio technology as the representative multimedia technology has developed rapidly recently. CD, DVD and other multimedia devices have already been applied in thousands of families. High-definition digital broadcasting is developing gradually worldwide. The traditional two-channel stereo could not meet people's requirements of high quality sound any more, so digital theater system (DTS) 5.1 surround sound encoding format came into being, which could provide people with a high quality acoustic enjoyment. Now DTS has been widely used in films, DVD, HDTV and other consumer electronics.The key technology of DTS standard was studied and analized at first. Then the floating-point decoder for DTS was implemented with C and made fixed point. It was transplanted to the TMS320DM642 audio test platform and optimized at last. The audio decoding was realized by DM642, and the transmission of pulse code modulation values to the multi-channel audio serial port was accomplished by enhanced direct memory access, and the analog signals were generated through digital-analog converter chip AIC23B.The masking effect of human's auditory system and the psycho-acoustic model based on critical band concept was studied. The core algorithm modules of DTS audio codec were studied in detail, including sub-band coding, vector quantization coding, block decoding, cyclic redundancy check, adaptive differential pulse code modulation decoding, the joint sub-band decoding and low-frequency effect channel interpolation. The floating-point decoder for DTS 5.1 channels was implemented, followed by a fixed point process, in which determination steps of dynamic range of the variable and quantization values were elaborated and the quality of the fixed point programe was measured by the signal to noise ratio. During the porting process, the play mechanism was designed and the decoding efficiency was improved through optimization options, structure and memory regulation.Real-time decoding has been achieved on DM642 platform at the frequency of 85MHz till now, with the average signal to noise ratio of up to 82dB, which is capable for the high-quality decoding.