With this workflow, you can generate synchronized audio for any video. Simply select a video, describe what sounds your scene would make in a natural language and it will generate the sounds and combines the audio with the video automatically.
This workflow is a slight modification of kijai's example workflow you can find here: https://github.com/kijai/ComfyUI-MMAudio
MMAudio's GitHub repository: https://github.com/hkchengrex/MMAudio
MMAudio
The MMAudio is a robust tool for generating and refining audio outputs based on video content or user-specified prompts. It excels at creating natural voiceovers that match the tone and pacing of visual inputs. With customizable parameters for detail and style, it adapts to various audio production scenarios.