Configure FFmpeg apad Filter to Stop with Video

When padding audio in FFmpeg using the apad filter, the output stream can run indefinitely because the filter pads the audio with infinite silence by default. This article provides a direct, step-by-step guide on how to configure FFmpeg so that the apad filter stops processing automatically the moment your video stream ends, preventing infinitely growing file sizes.

The Problem: Why apad Runs Forever

The apad (audio pad) filter is designed to add silence to the end of an audio stream. If you apply apad without any constraints, FFmpeg will continue generating silent audio frames forever, even after the video stream has finished. To stop this, you must instruct FFmpeg to terminate the output generation when the shortest input stream (the video) ends.

The Solution: Using the -shortest Flag

The most efficient way to stop the apad filter when the video ends is to use the global -shortest option. This tells FFmpeg to stop writing to the output file as soon as the shortest stream (typically the video) finishes.

Here is the standard command structure:

ffmpeg -i input_video.mp4 -i input_audio.wav -filter_complex "[1:a]apad[padded_audio]" -map 0:v -map "[padded_audio]" -shortest output.mp4

How This Command Works:

• -i input_video.mp4: Defines the video source (Stream 0).
• -i input_audio.wav: Defines the audio source (Stream 1).
• -filter_complex "[1:a]apad[padded_audio]": Takes the audio from the second input (1:a), applies the apad filter to add silence to the end, and labels the output stream as [padded_audio].
• -map 0:v: Maps the video from the first input to the final output.
• -map "[padded_audio]": Maps the padded audio stream to the final output.
• -shortest: This is the key parameter. It forces FFmpeg to stop encoding the moment the video stream (0:v) ends, cutting off the infinite silence generated by apad.

Alternative: Limit Padding by Sample Count

If you do not want to use the -shortest flag, you can configure the apad filter to add a specific amount of silence instead of an infinite amount. This is useful if you know the exact duration differences beforehand.

• whole_len: Specifies the target minimum length of the audio stream in samples.
• pad_len: Specifies the exact number of silent samples to append.

Example of adding exactly 44,100 samples of silence (1 second of silence at 44.1 kHz):

ffmpeg -i input_video.mp4 -i input_audio.wav -filter_complex "[1:a]apad=pad_len=44100[padded_audio]" -map 0:v -map "[padded_audio]" output.mp4