We describe the runs for our participation in the Search sub-task of the Search and Hyperlinking Task at MediaEval 2012. Our runs are designed to form a retrieval baseline by using time-based segmentation of audio transcripts incorporating pause information and a sliding window to define the retrieval segments boundaries with a standard language modelling information retrieval strategy. Using this baseline system runs based on transcripts provided by LIUM were better for all evaluation metrics, than those using transcripts provided by LIMSI.