Error Detection for DMB Video Streams

Abstract: The purpose of this thesis is to detect errors in Digital Multimedia Broadcasting (DMB) transport stream. DMB uses the MPEG-4 standard for encapsulating Packetized Elementary Stream (PES), and uses the MPEG-2 standard for assembling them in the form of transport stream packets. Recently many research works have been carried out about video stream error detection. They mostly do this by focusing on some decoding parameters related to frame. Processing complexity can be a disadvantage for the proposed methods. In this thesis, we investigated syntax error occurrences due to corruption in the header of the video transport stream. The main focus of the study is the video streams that cannot be decoded. The proposed model is implemented by filtering video and audio packets in order to find the errors. The filters investigate some sources that can affect the video stream playback. The output from this method determines the type, location and duration of the errors. The simplicity of the structure is one of advantages of this model. It can be implemented by three simple filters for detecting errors and a “calculation unit” for calculating the duration of an error. Fast processing is another benefit of the proposed model.