How to Apply Vocoder Effect to Vocals in FFmpeg

This article provides a straightforward, step-by-step guide on how to create a classic robotic vocoder effect using FFmpeg. By combining a modulator track (your vocal recording) and a carrier track (a synthesizer chord or drone), you will learn the exact FFmpeg command-line arguments and filter parameters needed to merge these signals and generate a high-quality vocoded audio output.

Understanding the Vocoder Setup

To create a vocoder effect, you need two distinct audio inputs: 1. The Carrier (Synth): A rich harmonic sound source, such as a synthesizer pad, saw-wave chord, or bright drone. This provides the musical pitch and tone. 2. The Modulator (Vocal): The speech or vocal track. This provides the rhythmic and phonetic characteristics (the words being spoken).

FFmpeg’s built-in vocoder filter requires both inputs to be merged into a single stereo stream first, where one channel acts as the carrier and the other acts as the modulator.

The FFmpeg Command

Below is the standard command to apply the vocoder effect. This command takes a mono synth track and a mono vocal track, merges them into a stereo pair, and applies the vocoder filter.

ffmpeg -i synth.wav -i vocal.wav -filter_complex \
"[0:a][1:a]join=inputs=2:channel_layout=stereo[stereo]; \
 [stereo]vocoder=carrier_channel=0:modulator_channel=1:bands=30:gain=2.0[out]" \
-map "[out]" output.wav

Command Breakdown

• -i synth.wav: The carrier audio file (Input 0).
• -i vocal.wav: The modulator vocal file (Input 1).
• -filter_complex: Initiates FFmpeg’s complex filtergraph, which is required when processing multiple inputs.
• [0:a][1:a]join=inputs=2:channel_layout=stereo[stereo]: This joins the audio stream from the first input ([0:a]) and the second input ([1:a]) into a single stereo stream named [stereo]. By default, the first input goes to the Left channel (index 0) and the second goes to the Right channel (index 1).
• vocoder=...: Applies the vocoder filter to the merged stream.
  • carrier_channel=0: Specifies that the Left channel (index 0, the synth) is the carrier.
  • modulator_channel=1: Specifies that the Right channel (index 1, the vocal) is the modulator.
  • bands=30: Sets the number of frequency bands.
  • gain=2.0: Applies a volume boost to the resulting output to compensate for volume loss during filtering.
• -map "[out]": Maps the final processed filter output to the destination file.
• output.wav: The rendered vocoded audio file.

Fine-Tuning the Parameters

You can adjust the parameters within the vocoder filter to customize the texture and clarity of the effect:

• Adjusting Bands (Clarity vs. Robot Effect): The bands parameter determines the resolution of the vocoder.
  • Lower values (e.g., 10 to 16) result in a vintage, highly synthetic, and chunky robotic sound.
  • Higher values (e.g., 30 to 50) provide much better speech intelligibility and a smoother, more modern sound.
• Handling Volume Levels: Vocoding can significantly reduce the output volume. Use the gain parameter (e.g., gain=3.0) to boost the output level directly inside the filter, or normalize the input tracks before processing.