Audio Codec And MIDI Of Mobile Multimedia

Audio is the most significant and fundamental application in Mobile field. And audio standard MP3 and MIDI, which well solved the limit of both the computation, memory capacity of mobile terminal and the band-with of wirelesscommunication. And these technique won it's reputation on the mobile multimedia development field.In this paper, I gave a introduction of Mobile Multimedia nowadays and it's application at the beginning. And expressed the importance of Audio technique in considering the computation, storage capacity of the mobile terminals,and the band-with of wireless communication. And then is the contend of human audio character, and psychoacoustic model of MP3. The coming chapters is the core technique of MP3, it include signal sub-band decomposition.MDCT, psychoacoustic model based bit allocation scheme,non-uniform quantization, entropy coding and etc.The core technique of signal sub-band decomposition is the filter group. In this paper we introduce two-channels filter group firstly, and then QMF, CQF, CMF, at last we gave the PQMF of MP3. It adopt FIR low pass filter as it's prototype filter, It is 512 order, linear Phase, and approximately ideal. We can get 32 sub-band filter though cosine modulation. So the input signal can be decomposed into 32 sub-band signal. And we can get 32 filters through this prototype filter. Here are the analvsis and synthesize filters;Since the MP3 psychoacoustic model is based on the analysis of audio signal in frequency-domain, and it needs higher frequency-domain resolution than 32 order to get accurate bit allocation scheme, so the 32 filters needs MDCT in order to get more refined audio signal spectrum. MDCT use data overlap technique, avoid marginal effect caused by DCT, and don't sacrifice compress efficiency. Psychoacoustic model computer the part of the frequency that could be apperceive, and the part that could be masked. And the outcome is SMR:Psychoacoustic model can recognize the signal that is masked, it saved much bit for the encoder and provide flexible time-frequency resolution ratio for MDCT and sub-band, and make the encoder to have optimal bit allocation scheme for stationary non-stationary signal. MP3 audio standard support multi-mono audio format, because there is correlation among different mono, and MP3 provide various audio coding mode, therefore correlation is avoided to improve compress efficiency of the signal.According to the features of mobile multimedia terminal, we introduced two improved algorithms for PQMF. The first algorithm needs only 80 additionand multiplication, and it improved the MP3 algorithm in order of magnitude. Sample of the k-th sub-band is:And synthesis filter have similar fast algorithm:The second one is based on DCT and IDCT, letthen the upon analysis filter algorithm is: It is just the 32 points IDCT. Because DCT is widely used in industrial field, and almost all multimedia modular is implemented by DCT. So this algorithm can be effectively implement the MP3 algorithm.In the following we present a improved MP3 algorithm(Chih-HsuYen,Yu-ShiangLinå'ŒBing-Fei Wu). It abandons psychoacoustic model and distortion control model, and adopt a new bit control algorithm, and achieve nice efficiency. In encoding end, it applies a low pass filter to control frequency-band and a new bit control algorithm, it include precise initialization,fast search of optimal quantize parameter and energy distribution based dynamic bit allocation. In decoding end, there are newly inverse quantization, IMDCT and sub-band synthesize.MIDI music synthesis has act as an important role in mobile multimedia field, resort on it's simple computation complexity and tiny storage requirement.From Mobile phone ring tones to animation background music, all are instances of the application of MIDI.The core technique of MIDI is sound synthesis, and in this paper, I introducedthe four most popular sound synthesis technique: frequency modulation,subtractive synthesis, wavetable synthesis and waveguide synthesis. These four algorithm have diversity in both computation complexity and storage requirement. In terms of computation complexity, they sort from low to high is: wavetable synthesis, waveguide synthesis, subtractive synthesis,frequency modulation. In terms of storage requirement, they sort from low to high is: frequency modulation, subtractive synthesis,waveguide synthesis,wavetable synthesis.Actually, the period sampling sample of sound signal saved by wavetable synthesis is the period sound signal sample in the model of frequency modulationand subtractive synthesis. Since frequency modulation and subtractive synthesis needs computation to get period sample of sound signal, the computationcomplexity of these two methods is higher than the wavetable synthesis that had the samples. On the other hand, wavetable synthesis needs memory to storage period sound signal sample, so it require more memory than sub- tractive synthesis than frequency modulation. In these four sound synthesis, waveguide synthesis moderate in both computation complexity and memory requirement, so it can be choose as a moderate scheme to system resource requirement.On specific sound synthesis effect, these technique have it's character respectively. Wavetable synthesis have the best simulation to the having sound. It could simulate some nature sound, and also mixed sound. The character of frequency modulation and subtractive synthesis is that synthesis takes more and natural sound takes less. Frequency modulation use modulationand carrier wave technique to synthesize sound frequency spectrum, and subtractive synthesis use filter technique.