How to Use the FFmpeg Vocoder Filter

This article provides a straightforward guide on how to use the vocoder filter in FFmpeg to apply a classic robotic voice effect to your audio. You will learn the fundamental mechanics of how the filter processes audio, the key parameters you can adjust, and practical command-line examples for combining voice and carrier signals on the fly.

Understanding the FFmpeg Vocoder Filter

The vocoder filter in FFmpeg requires a stereo input to function. It uses one channel as the modulator (typically a human voice) and the other channel as the carrier (typically a synthesizer, instrument, or noise).

By default: * The Left channel is treated as the modulator (speech). * The Right channel is treated as the carrier (synthesizer).

The filter analyzes the frequency characteristics of the modulator and applies them to the carrier, resulting in the classic “talking synthesizer” effect.

Basic Filter Parameters

You can customize the effect using the following parameters:

• bands: Sets the number of frequency bands (default is 30). A higher number of bands (up to 240) makes the speech more intelligible, while a lower number makes it sound more abstract and robotic.
• scale: Sets the output gain scale (default is 1.0).
• cutoff: Sets the carrier bandpass cutoff frequency in Hz (default is 20000).
• pitch: Sets the pitch scale (default is 1.0).

Practical Command Examples

Example 1: Applying Vocoder to a Pre-Prepared Stereo File

If you already have a stereo audio file where the left channel contains the voice and the right channel contains the synthesizer, use this simple command:

ffmpeg -i input_stereo.wav -af "vocoder=bands=40:scale=1.5" output.wav

Example 2: Combining Two Separate Files (Voice and Synth)

Most of the time, your voice and synthesizer tracks will be in separate files. You can use the amerge filter to combine them into a single stereo stream before passing them to the vocoder filter:

ffmpeg -i voice.wav -i synth.wav -filter_complex "[0:a][1:a]amerge=inputs=2[stereo];[stereo]vocoder=bands=40:scale=1.2[out]" -map "[out]" output.wav

In this command: 1. -i voice.wav is input 0 (the modulator). 2. -i synth.wav is input 1 (the carrier). 3. amerge=inputs=2 merges the two mono inputs into a single stereo stream where voice.wav becomes the left channel and synth.wav becomes the right channel. 4. vocoder processes this combined stream to generate the final effect.