Extracting Audio from Video Files with Xabe.FFmpeg

Introduction And Context Setting

Welcome back! In the previous lesson, you learned how to preprocess audio files using FFmpeg and the Xabe.FFmpeg library in C# . You explored how to normalize audio, convert it to a standard format, and prepare it for tasks like transcription. This foundational knowledge is essential for working with audio data in real-world applications. Today, we will build on that foundation by introducing a new and very practical skill: extracting audio from video files. Many times, the content you need to process or transcribe is not in a standalone audio file but is embedded within a video — such as a recorded meeting, a lecture, or a podcast episode published as a video. Being able to extract just the audio track from these files is a key step in many media workflows.

Lesson Objectives And Expected Outcome

By the end of this lesson, you’ll know how to extract audio from a video file using C# and Xabe.FFmpeg, making it ready for further processing or transcription. You’ll also understand why this step is often necessary and advantageous in modern workflows. Extracting audio is useful for several reasons. Audio files are usually much smaller than their video counterparts, making them easier to upload, share, and process. Many APIs and cloud services for transcription or speech recognition enforce file size limits (e.g., 25MB per upload), so working with just the audio ensures you remain within those boundaries. Focusing on audio also reduces bandwidth costs and speeds up processing, especially when the video content isn’t needed.

Understanding The Audio Extraction Process

When extracting audio from video, it’s important to produce output that is compatible with your downstream tasks, such as transcription. For best results, you want a mono, 16kHz WAV file—this is the standard for most speech recognition APIs. Here’s what the key FFmpeg parameters do in this workflow: -i "input": Specifies the source video file. -vn: Ignores the video stream, processing audio only. -ar 16000: Sets audio sample rate to 16,000 Hz, which is standard for speech. -ac 1: Forces mono (single channel) output. -acodec pcm_s16le: Saves as 16-bit signed PCM in little-endian format, which is widely accepted for WAV files. Additionally, you can extract only a specific segment of the audio by using the -ss (start time) and -t (duration) parameters: -ss {startTimeSeconds}: Start extracting from this timestamp (in seconds). -t {durationSeconds}: Extract audio for this duration (in seconds). By using these parameters together, you create a clean, compact, and compatible audio file ready for transcription, and you can focus on just the segment you need.

Join the 1M+ learners on CodeSignal

Be a part of our community of 1M+ users who develop and demonstrate their skills on CodeSignal

public async Task ExtractAudioFromVideoAsync(string videoPath, string audioPath, double startTimeSeconds, double durationSeconds)
{
    var parameters = $"-ss {startTimeSeconds} -i \"{videoPath}\" -t {durationSeconds} -vn -ar 16000 -ac 1 -acodec pcm_s16le \"{audioPath}\"";

    await FFmpeg.Conversions.New()
        .AddParameter(parameters, ParameterPosition.PreInput)
        .Start();
}

public async Task ExtractAudioFromVideoAsync(string videoPath, string audioPath, double startTimeSeconds, double durationSeconds)
{
    var parameters = $"-ss {startTimeSeconds} -i \"{videoPath}\" -t {durationSeconds} -vn -ar 16000 -ac 1 -acodec pcm_s16le \"{audioPath}\"";

    await FFmpeg.Conversions.New()
        .AddParameter(parameters, ParameterPosition.PreInput)
        .Start();
}

public async Task ExtractFullAudioFromVideoAsync(string videoPath, string audioPath)
{
    var parameters = $"-i \"{videoPath}\" -vn -ar 16000 -ac 1 -acodec pcm_s16le \"{audioPath}\"";

    await FFmpeg.Conversions.New()
        .AddParameter(parameters, ParameterPosition.PreInput)
        .Start();
}

public async Task ExtractFullAudioFromVideoAsync(string videoPath, string audioPath)
{
    var parameters = $"-i \"{videoPath}\" -vn -ar 16000 -ac 1 -acodec pcm_s16le \"{audioPath}\"";

    await FFmpeg.Conversions.New()
        .AddParameter(parameters, ParameterPosition.PreInput)
        .Start();
}

// Initialize the Audio Processor
AudioProcessor.Initialize();

var transcriptionService = new TranscriptionService();

// Extract mono 16kHz WAV audio from a specific segment of the video
string videoPath = Path.Combine("Assets", "sample_video.mp4");
string audioPath = Path.Combine("Assets", "sample_audio_from_video.wav");

// For example, extract audio from 30s to 90s (60 seconds duration)
double startTimeSeconds = 30;
double durationSeconds = 60;

var audioProcessor = new AudioProcessor();
await audioProcessor.ExtractAudioFromVideoAsync(videoPath, audioPath, startTimeSeconds, durationSeconds);

Console.WriteLine($"Audio extracted to: {audioPath}");

// Perform transcription on the extracted audio
AudioTranscription transcription = transcriptionService.TranscribeAudio(audioPath);

// Print the transcribed text
Console.WriteLine("Video Transcription:");
Console.WriteLine(transcription.Text);

// Initialize the Audio Processor
AudioProcessor.Initialize();

var transcriptionService = new TranscriptionService();

// Extract mono 16kHz WAV audio from a specific segment of the video
string videoPath = Path.Combine("Assets", "sample_video.mp4");
string audioPath = Path.Combine("Assets", "sample_audio_from_video.wav");

// For example, extract audio from 30s to 90s (60 seconds duration)
double startTimeSeconds = 30;
double durationSeconds = 60;

var audioProcessor = new AudioProcessor();
await audioProcessor.ExtractAudioFromVideoAsync(videoPath, audioPath, startTimeSeconds, durationSeconds);

Console.WriteLine($"Audio extracted to: {audioPath}");

// Perform transcription on the extracted audio
AudioTranscription transcription = transcriptionService.TranscribeAudio(audioPath);

// Print the transcribed text
Console.WriteLine("Video Transcription:");
Console.WriteLine(transcription.Text);

await audioProcessor.ExtractFullAudioFromVideoAsync(videoPath, audioPath);

await audioProcessor.ExtractFullAudioFromVideoAsync(videoPath, audioPath);

Audio extracted to: Assets/sample_audio_from_video.wav
Video Transcription:
[Transcribed text from the video’s audio track]

Audio extracted to: Assets/sample_audio_from_video.wav
Video Transcription:
[Transcribed text from the video’s audio track]

public async Task ExtractAudioFromVideoAsync(string videoPath, string audioPath, double startTimeSeconds, double durationSeconds)
{
    var parameters = $"-ss {startTimeSeconds} -i \"{videoPath}\" -t {durationSeconds} -vn -ar 16000 -ac 1 -acodec pcm_s16le \"{audioPath}\"";

    await FFmpeg.Conversions.New()
        .AddParameter(parameters, ParameterPosition.PreInput)
        .Start();
}

public async Task ExtractFullAudioFromVideoAsync(string videoPath, string audioPath)
{
    var parameters = $"-i \"{videoPath}\" -vn -ar 16000 -ac 1 -acodec pcm_s16le \"{audioPath}\"";

    await FFmpeg.Conversions.New()
        .AddParameter(parameters, ParameterPosition.PreInput)
        .Start();
}

// Initialize the Audio Processor
AudioProcessor.Initialize();

var transcriptionService = new TranscriptionService();

// Extract mono 16kHz WAV audio from a specific segment of the video
string videoPath = Path.Combine("Assets", "sample_video.mp4");
string audioPath = Path.Combine("Assets", "sample_audio_from_video.wav");

// For example, extract audio from 30s to 90s (60 seconds duration)
double startTimeSeconds = 30;
double durationSeconds = 60;

var audioProcessor = new AudioProcessor();
await audioProcessor.ExtractAudioFromVideoAsync(videoPath, audioPath, startTimeSeconds, durationSeconds);

Console.WriteLine($"Audio extracted to: {audioPath}");

// Perform transcription on the extracted audio
AudioTranscription transcription = transcriptionService.TranscribeAudio(audioPath);

// Print the transcribed text
Console.WriteLine("Video Transcription:");
Console.WriteLine(transcription.Text);

await audioProcessor.ExtractFullAudioFromVideoAsync(videoPath, audioPath);

Audio extracted to: Assets/sample_audio_from_video.wav
Video Transcription:
[Transcribed text from the video’s audio track]

Introduction And Context Setting

Lesson Objectives And Expected Outcome

Understanding The Audio Extraction Process

Join the 1M+ learners on CodeSignal

Implementing Audio Extraction in C#

Using Audio Extraction in Your Application

Real-World Considerations And Best Practices

Summary And Next Steps