Hidden in Plain Sight: Evaluation of the Deception Detection Capabilities of LLMs in Multimodal Settings