The field of multimodal media generation is moving towards a more integrated approach, where audio and video are combined to create more realistic and engaging content. This is driven by the need for more sophisticated models that can capture the complex relationships between different modalities. Recent research has focused on developing new datasets and models that can effectively fuse audio and video information, leading to improved performance in tasks such as video generation and audio-visual denoising. Notable papers in this area include MVAD, which introduces a comprehensive multimodal video-audio dataset for AI-generated content detection, and Audio-Visual Affordance Grounding, which proposes a new task and model for learning visual affordance from audio. Other noteworthy papers include Does Hearing Help Seeing, which investigates the benefits of audio-video joint denoising for video generation, and Hear What Matters, which introduces a text-conditioned selective video-to-audio generation model.