FacePikaFacePika
واپس بلاگ پر
Published Updated 2 min read

How to Keep Audio in a Face Swap Video

Learn why a face-swap result may lose audio and which Android-compatible audio formats give FacePika the best chance of preserving it.

FacePika Product Team
How to Keep Audio in a Face Swap Video

A video face swap changes the picture track, but viewers usually expect the original music, dialogue, or ambient sound to remain. Preserving that audio depends on whether the source track can be copied into the output MP4 container.

How FacePika handles source audio

FacePika processes and re-encodes the video track, then attempts to copy a compatible source audio track into the output. The copied audio is trimmed to the generated video duration so the two tracks end together.

The current Android muxing path accepts AAC, 3GPP/AMR-NB, and AMR-WB audio. If the source uses MP3, WMA, Vorbis, Opus, or another unsupported track, FacePika skips that audio rather than failing the entire video export.

Use AAC audio when possible

For the most predictable result, prepare an MP4 containing H.264 video and AAC audio. This combination is widely supported by Android and is the best match for FacePika's output pipeline.

How to check the source before processing

Many gallery apps show only the filename and duration. Use a media-information utility or a trusted editor when you need to confirm the codecs. Look for H.264 or AVC under video and AAC under audio. The extension alone is not enough: an MP4 container can still contain an audio codec that the output muxer cannot copy.

If the generated video has no sound

  1. Confirm that the original file actually contains an audio track.
  2. Convert or export the source as H.264 MP4 with AAC audio.
  3. Run the face swap again using the converted file.
  4. Review the whole result, not only the first preview frame.
  5. If necessary, add the original audio back in a video editor after the swap.

Why FacePika does not transcode every audio format

Copying a compatible track is faster and avoids another lossy audio encode. Automatically transcoding every possible input would add processing time and introduce more device-specific codec failures. The current behavior prioritizes completing the visual face swap and includes the audio only when the MP4 muxer can safely accept it.

See the recommended video format for Android face swap for preparation settings, or visit the video face swap app page.