The player has a volume icon and the file extension says MP4, so you assume the video must contain audio. Then you open it and hear nothing.
MP4 is only a container. It can hold video, audio, both or even other media tracks. A perfectly valid MP4 may contain no audio stream at all.
Find out whether an audio stream exists
Try the file in a different player and inspect its media information. If the file genuinely has no audio track, changing the volume will not help. If it contains AAC or another audio stream, investigate mute settings, device routing or encoding compatibility.
The distinction saves time. Playback settings cannot restore a stream that was never generated.
Why AI tools return silent video
Some models specialize in picture generation. Others offer an optional audio-generation setting. Still others preserve sound only in certain editing modes. A disabled audio option may be deliberate, especially if your workflow plans to reuse the original clip's dialogue.
For a reference-video edit, the application can sometimes restore sound after generation by combining the new picture with the source soundtrack. That is a media-processing operation, not a second round of voice generation.
Check duration before combining tracks
If the model changes the output length, audio can end too early or continue after the picture. Video and soundtrack need a deliberate synchronization rule. Trim or pad where appropriate and inspect dialogue at the start, midpoint and end.
Do not assume the original sound will line up perfectly if the AI reinterpreted mouth movements or camera cuts.
What to report to support
Include the generation ID, whether the source video had sound, whether the export contains an audio track, and the output duration. “The speaker is muted” and “the MP4 contains no audio stream” are very different bug reports.
See our guide to keeping original dialogue for a more complete overview.
