A way to reduce effects of packet loss in video streaming using Multiple Description Coding

Multiple description (MD) coding has appeared to be an attractive technique to decrease impact of network failures and increase the robustness of multimedia communications. Very common model of this technique is multiple-description lattice vector quantization, which is the best choice for robust data transmission over the unreliable network channels. However, MD coinciding lattice vector quantizer (MDCLVQ) is not considered discrete network conditions, so in this scheme, all videos are received or are not received. In this paper, this scheme is implemented in real network environment. So, raw video will be send in various packet, packets send independently and packets lose independently. The possibility of lossing all packets together is close to zero. Our object for increasing of resistance transfer in error- prone communication channels are used. This technique has been tested for standard videos "Akiyo", "Carphone", "Miss-America" and "Foreman". This results show that the quality of the reconstructed videos from the average PSNR values of the central decoder and the side decoders has been reached to grate degree, so increases error resilience over error-prone communication channels.

Keywords:

Multiple Description Coding , Real-Time Video , Lattice , Sub-lattice hexagonal , Quantization