Silent After Export: Tracking Down a Video's Missing Audio

First decide: is the file silent, or the player?

The same symptom, a picture with no sound, has two very different explanations. Note exactly what happens:

• No sound in any player on any device: the file probably has no audible audio.
• Sound in VLC but not in the default player, a browser or a phone: the player probably cannot decode the audio codec.
• Sound in one earbud only, or very faint: the audio may sit on one channel.
• Sound on headphones but almost none on a phone or laptop speaker: the stereo channels may be cancelling each other when mixed to mono.
• Only background sound between spoken lines in a dubbed video, or songs without their vocals: that is expected, because the dub removes the original speech and singing, as explained further down.

Common causes, from export to playback

Audio disabled at export

An export preset with audio switched off, or a video-only format option, produced a file with no audio stream

Muted or disabled track

A muted track, a disabled clip or a track left out of the export routing produced a silent but present stream

Codec the player cannot decode

PCM, FLAC or Opus audio inside MP4, or surround formats some browsers lack, plays in some apps and not others

Wrong track

The file has several audio tracks, and the player or the next tool picks one that holds silence or system sound

One-sided channel

A mono microphone recorded on the left channel only plays in one ear and is easy to miss

Phase cancellation

Two channels with opposite polarity cancel each other on mono speakers

Silent source

The recorder never captured audio, for example a screen recorder with the microphone or system sound turned off

Platform muting

A video site muted the upload or a section because of a copyright claim on the music

Diagnose in this order

1. Rule out the obvious: device volume, a muted browser tab, and audio routed to Bluetooth headphones that are not being worn.
2. Play the file in VLC or mpv on a computer. These players decode a wide range of codecs, so silence there points to the file.
3. List the file's streams with MediaInfo, or in a terminal with ffprobe, using the command below.
4. Read the result. No audio stream means the export left it out. An audio stream with an unusual codec explains player-specific silence. Several audio streams mean the wrong one may be playing.
5. Open the file in an editor and look at the audio waveform. A flat line means the stream exists but contains silence, so the problem happened at recording or in the timeline.
6. Check the export or conversion settings you used, and the original recording, to find the step where the sound disappeared.

ffprobe -hide_banner export.mp4

Repairs matched to each cause

Disabled or muted audio. Unmute the track, check the export routing and re-export with audio enabled, using AAC, the most widely supported audio codec in MP4.

Unsupported codec. Convert just the audio and copy the picture unchanged, which is fast and loses no video quality:

ffmpeg -i export.mp4 -c:v copy -c:a aac -b:a 192k export-aac.mp4

Wrong track. Keep only the track that holds the speech. This copies the video and the second audio track, counting from zero, into a new file:

ffmpeg -i recording.mkv -map 0:v -map 0:a:1 -c copy recording-voice.mkv

The article on video files with multiple audio tracks explains how to tell the tracks apart.

One-sided channel. Copy the good channel to both sides. This example duplicates the left channel:

ffmpeg -i export.mp4 -c:v copy -af "pan=stereo|c0=FL|c1=FL" -c:a aac export-fixed.mp4

Phase cancellation. Invert the polarity of one channel in an editor, or export mono from a single channel; mono vs stereo for speech covers why this happens.

Silent source. There is nothing to recover. Re-record, or recover audio from a second device that was recording at the same time.

A worked example: a screen recording that lost its voice

Mistakes and limits

• Re-encoding the whole video to fix audio wastes time and picture quality; convert only the audio stream.
• Adjusting a player's settings fixes your playback, not the file your viewers receive.
• Audio that was never recorded cannot be reconstructed.
• Muting by a video platform over copyrighted music is a licensing matter, not a file fault; replacing the music or licensing it are the only real fixes.
• Some players, browsers and phones do not play MKV files at all, which can look like a sound problem when it is a container problem.

What a mydubly translated video contains

A translated video from mydubly holds two things: the original video stream, copied without re-encoding, and the translated audio track, the same audio offered separately as an M4A download. AI vocal separation removes the original speech, and the remaining music, ambience and sound effects are mixed under the translated voice, lowered while it speaks. Passages without speech therefore carry only the original background, and stay quiet if the original had none. Sung vocals count as voice, so songs keep their instruments but lose their vocals, and in dense mixes the background can sound slightly thinner than the original. If you have your edit project, dubbing a dialogue-only export and mixing the translated voice with your clean music stem in an editor gives the cleanest result, as described in keeping background music in translated videos.

A few other properties matter when checking a result. If the video codec cannot go into MP4, the output is MKV, which some browsers and phones will not play inline; open it in a desktop player instead. On devices with too little memory for a very large file, you receive the translated audio to download instead of a merged video. And if your source has several audio tracks, only one is used, so export the source with the speech as the only or main track; a track without speech gives an empty or sparse transcript and a sparse dub. How the merge works in the browser is covered in browser video muxing.

When another workflow fits better

If viewers should be able to choose between the original audio and the translation, build a file with two audio tracks in an editor or with ffmpeg rather than replacing the sound. If the music must stay exactly as mixed, including any sung vocals, dub a dialogue-only export of your project and plan for an editing pass that mixes the translated voice with your clean music stem. And if the source itself has audio problems, fix them before translating, because the transcript and the dub both start from that audio.

Next step: export once, check once, then translate

Export your video with a single AAC speech track, confirm it plays with sound in two players, then run it through the video translator. The guide to dubbing a video with AI covers voice choice and review, and AI dubbing costs 50 credits per minute with a 2-minute minimum, so a 12-minute tutorial costs 600 credits (60¢).