| With the operation of 3G networks in the domestic formally and the emergency of various new multimedia mobile access technologies, a broad space and new opportunities have been brought for mobile multimedia applications. However, as the complexity of the mobile communications environment, the limitation of the wireless transmission frequency band resources, the variability of the transmission channel, the conventional audio codec technology encountered new embarrassment, therefore, technology based on 3G networks become a new key research in the field of mobile multimedia technologies. To meet the demanding application requirements of the mobile environment, several key technologies of mobile audio codec are researched deeply in this dissertation, including:the characteristic of human auditory perception and psychoacoustic model in mobile environment, bandwidth extension, error concealment and so on. Based on the above research, the most important results in this dissertation have formed as follows:1. Psychoacoustic model in mobile audio mixed codingIn the mobile multimedia applications, the audio signal contents are fairly complex. In order to solve the key problem of the complex audio signal encoding, in this dissertation, the weighting method of peripheral ear model of the psychoacoustic model in PEAQ coding standard is transplanted into EAAC+ coding standard so that the calculation of the masking threshold is more consistent with the physiological characteristics of the human ear, and its accuracy can be effectively improved. So the treatment can not only maintain the advantages of audio signal coding, but also improve the ability of speech signal coding.Modified algorithm does not exist the problem of coding mode switching, and the computational complexity is relatively lower. However, The time domain waveform of audio signal encoded with modified algorithm is close to the waveform of original signal and their quality are yet roughly the same.2. Bandwidth extension in mobile audio codingBased on EAAC+ coding algorithm, a detection and removal of sinusoidal component of the bandwidth extension algorithm is presented for the exist sinusoidal component of the original audio input signal. Firstly, the original audio input signal is divided into low frequency part and high frequency part. For the low frequency part, the implementation is still in accordance with the SBR algorithm. For the high frequency part, in accordance with the BBSM algorithm, the whole high frequency band is divided into low frequency sub-band and high frequency sub-band in order to reduce the following transmission bit rate. By detection and removal of sinusoidal component can achieve lower complexity of encoding algorithm.The subjective and objective evaluation results show that the high frequency spectrum of reconstructed audio signal maintains the fundamental frequency characteristics of original signal spectrum, according to the algorithm of bandwidth extension in the dissertation, and the coherence between high frequency spectrum and low frequency spectrum is good.3. Error concealment in mobile audio codingError concealment technologies are the important part of audio codec algorithm research, and when the transmission of data packet loss occurs, it is a effective solving means. In order to ensure human auditory system being not aware of the degradation of audio quality with the data packet loss, a error concealment scheme based on EAAC+ coding standard is presented.By detecting the transient signal of music signal in encoder, the relevant information as additional data is embedded into the compressed audio stream, and in decoder, the embedded additional information is decoded for error concealment in order to reconstruct the quality of audio signal.The theory and method for mobile audio codec key technology based on 3G communication is researched in this dissertation, its research methods and results laid the foundation of mobile audio technology application based on 3G communications. |
PDF Full Text Request