MPEG audio coding under the name MP3 has become one of the most popular standards for digital audio broadcasting and videos. High compression ratios offered by MP3 codecs in various stand alone players and hand held devices over the last few years has increased its popularity immensely. Internet users, music lovers who would like to download highly compressed digital audio files at near CD quality are the most benefited. Psychoacoustic model, Modified Discrete Cosine Transform (MDCT) and Huffman coding play a vital role in achieving such magnificent compression ratios. In this thesis, a thorough knowledge of MP3 decoder is obtained by going through the ISO standard and then some of the decoder blocks have been implemented for deeper understanding.
Reference URL 1: Visit Now
Author: Praveen Sripada
Source: Blekinge Institute of Technology
Contents
1 Introduction
1.1 How does MP3 work?
1.2 MPEG Audio Compression
2 Overview of audio compression formats
3 Inside an MP3 file
4 Overview of MP3 Encoder
4.1 Filter bank and Psychoacoustic model
4.2 Quantisation
4.3 Huffman coding
4.4 Bitstream formatting
5 MP3 Decoder
5.1 Audio Frame Header
5.1.1 Frame Header in detail
5.1.2 Frame Length Calculation
5.2 Decoding Side information
5.3 Main data
5.4 Decoding Scalefactors
5.5 Decoding Huffman data
5.6 Requantizing spectrum
5.7 Reordering spectrum
5.8 Stereo processing
5.8.1 Mid/Side stereo
5.8.2 Intensity stereo
5.9 Alias reduction
5.10 Inverse Modified Discrete Cosine Transform and Overlapping
5.11 Frequency inversion
5.12 Synthesis via polyphase filter bank
6 Implementation
6.1 Header information for frame 1
6.2 Side information details for frame 2
6.3 Problems encountered during implementation
7 Conclusions
8 References