The Research And Implementation Of DRA Codec Algorithms Based On FPGA

With the development of the information age, digital audio technology has been blending in our work and life deeply, as an indispensable part of our life. The digital audio standards being widely applied currently include MPEG-1, MPEG-2, Dolby AC-3 and DTS. All the standards'intellectual properties are belonging to foreign enterprises. It takes high patents fees when our country uses these technologies. National Standards Commission has published the "Specification for multichannel digital audio coding technology", the "DRA standard" in December 22nd,2008, which was authorized as the Audio National Standards in 2009. The DRA standard filled the blank of the core technology of digital audio industry in our country for its characteristics as high compression rate, good sound quality and low decoding complexity. It also promoted the development of our audio industry greatly and enhanced our international competitiveness.This paper focuses on researching the DRA codec standards. The Psychoacoustic Model, dividing of transient segment, a variable resolution filter bank, Huffman code book selection in the encoding process and the restore of window sequence, Huffman code book application, decoding quantization factor during the recoding process were all analyzed in detail in this paper. In the basis of in-depth research and analysis the DRA standard, the stream decomposition, header information decoder, Huffman decoding module and quantization factor decoding module were designed with the hardware programming language VHDL and verified on the FPGA chip, EP2C35F484C7 of DEI development board. The data fusion technology was adopted during the decoding module design process, which combined several simple algorithm decoding module functions together to make decoding operation. It saved many data reading time. In addition, an improved 16-tree look-up table computational method was proposed towards the Huffman decoding module, considering the relationship between look-up table memories and decoding rate. It also saved lots of storage space in the basis of not reducing the Huffman decoding rate.Except the researching of decoding algorithm, the overall design of voice alarm system was given in this paper. The alarm system, which gives a sound alarm for many simultaneous and multi-level measuring signals, can give operation tips for each alarm condition.