Match Audio Speed to Fast Motion Video with FFmpeg

When creating fast-motion video effects in FFmpeg, speeding up the video stream without adjusting the audio results in desynchronization. To keep the audio track perfectly aligned with the accelerated video, you must use the asetpts (audio set presentation timestamp) filter in combination with audio tempo or sample rate adjustments. This article provides a straightforward guide on how to use asetpts to match your audio to fast-motion video, covering both pitch-preserved and pitch-altered methods.

Understanding the Math

To speed up a video, you reduce its presentation timestamps (PTS) using the setpts filter. For example, to make a video play at double speed (2x), you multiply the PTS by 0.5:

setpts=0.5*PTS

To match this fast-motion effect, the audio stream must be modified to fit the exact same timeline. This requires adjusting the audio speed and using asetpts=0.5*PTS to scale the audio presentation timestamps.

Here are the two ways to apply this filter depending on how you want the audio to sound.

Method 1: Speeding Up Audio with Pitch Preservation (Normal Voice)

If you want the audio to play faster to match the video but want to keep the voice pitch sounding natural (avoiding the “chipmunk” effect), combine the atempo filter with asetpts.

Use the following command to speed up both video and audio by 2x:

ffmpeg -i input.mp4 -filter_complex "[0:v]setpts=0.5*PTS[v];[0:a]atempo=2.0,asetpts=0.5*PTS[a]" -map "[v]" -map "[a]" output.mp4

How it works: * setpts=0.5*PTS speeds up the video by 2x. * atempo=2.0 speeds up the audio data by 2x without changing the pitch. * asetpts=0.5*PTS rescales the audio timestamps to ensure they match the accelerated video timeline perfectly.

Method 2: Speeding Up Audio with Pitch Change (Chipmunk Effect)

If you want the classic “high-pitched fast-forward” sound to accompany your fast-motion video, you should use the asetrate filter alongside asetpts.

For a 2x speed increase, you must double the input sample rate (e.g., from 44,100 Hz to 88,200 Hz) and then reset the timestamps:

ffmpeg -i input.mp4 -filter_complex "[0:v]setpts=0.5*PTS[v];[0:a]asetrate=44100*2,asetpts=0.5*PTS[a]" -map "[v]" -map "[a]" output.mp4

How it works: * asetrate=44100*2 changes the sample rate configuration, which increases the playback speed and raises the pitch. Note: Replace 44100 with your input file’s actual sample rate if different (e.g., 48000). * asetpts=0.5*PTS rebuilds the audio timestamps so that the player renders the high-sample-rate audio at the accelerated speed, syncing it with the video.

Formula Cheat Sheet for Different Speeds

To apply different fast-motion speeds, use this guide to calculate your filter values:

Note: The atempo filter in FFmpeg only supports values between 0.5 and 100.0.