Packet-loss error detection system for DTV and set-top box functional testing

Video communication in present multimedia networks is generally prone to transport stream packet-loss error that is manifested as annoying visual artifact in the decoded video signal. The packet-loss error is generally due to noise and signal attenuation in the transport medium, which introduces errors in the modulated digital packetized signal. A case of special interest related to this issue, at which this paper is targeting, is digital broadcasting network and Digital TV receiver as the end application. At the receiver side, the DTV set performs demodulation and evaluates the errors in the incoming transport stream by a parametric model, based on which it decides whether to display the damaged video or simply acknowledge "no signal" state with black screen or frozen video. In order to test DTV set functionality in terms of detecting perceived packet-loss errors and correct switching between a "correct" signal mode (when no video degradation is present) to a "no signal" mode (when video degradation is detected), we propose an embedded system and a novel method for packet-loss detection based on decoded image information. The proposed system is highly efficiently implemented on a INTEL Quad Core platform with a simple connection to the TV set (under investigation) via CVBS. This facilitates easy usage of the proposed test system on different geographical locations, within the broadcast network, with the same type of the DTV sets in order to reliably verify their functionality under different network conditions. The proposed system comprises of a novel algorithm module for packet-loss error measurement, based on which the DTV functionality is assessed. The results indicate high sensitivity of the algorithm to packet-loss error with satisfying robustness against various video sequences with different image content.