Free Video to Audio Extractor

What audio extraction actually does

A video file is a container (MP4, WebM, MOV, MKV, AVI, etc.) holding multiple elementary streams: typically one video stream, one or more audio streams, and sometimes subtitles or chapter markers. Audio extraction is a demuxing operation: the container is opened, the audio stream is read out, and either copied verbatim into a new audio file (stream copy, lossless and fast) or re-encoded to a different audio codec (transcode, slower with a small quality cost). FFmpeg handles both modes through its -c:a copy versus -c:a mp3/aac/opus flags. This tool picks the right path automatically based on your target format.

The audio codec inside a video file is usually AAC (most MP4s), Opus (newer WebM and MKV files), or sometimes MP3 (older files), AC-3 (DVD-style), or PCM (uncompressed in MOV/AVI). Each is a different way of representing the same waveform data; converting between them involves decoding to raw PCM samples and re-encoding to the target codec. Lossy codecs (AAC, MP3, Opus, Vorbis) discard inaudible frequencies and patterns to save space; lossless codecs (FLAC, ALAC, raw PCM in WAV) preserve every sample. If your goal is the cleanest possible audio, extract to FLAC or WAV. If your goal is a portable file for casual listening, MP3 at 192 kbps is the universal default.

Bitrate is the audio version of CRF: how many bits per second of audio are allotted. 320 kbps MP3 is "transparent" to most listeners (indistinguishable from CD quality in blind testing). 192 kbps MP3 is the default for most music libraries. 128 kbps is acceptable for voice/podcasts but audibly degrades music. 64 kbps is voice-only territory. AAC and Opus achieve the same perceived quality at lower bitrates (Opus is the most efficient: 96 kbps Opus rivals 192 kbps MP3 for music).

How this tool works under the hood

Same ffmpeg.wasm engine as the other video tools: FFmpeg compiled to WebAssembly via Emscripten, ~30 MB browser-side binary, runs entirely in the tab via SharedArrayBuffer multi-threading. When you drop a video, the file is read into the WebAssembly virtual filesystem via a streaming reader.

The extraction command looks like -i input.mp4 -vn -c:a libmp3lame -b:a 192k output.mp3 for MP3 at 192 kbps, where -vn drops the video stream entirely. For WAV (lossless): -i input.mp4 -vn -c:a pcm_s16le output.wav. For AAC: -c:a aac -b:a 192k output.m4a. For OGG/Opus: -c:a libopus -b:a 128k output.opus. The audio decoder built into FFmpeg handles whatever codec is inside the source (AAC, Opus, MP3, AC-3, FLAC, etc.) and outputs the chosen target codec.

When the source audio codec and the target codec match (rare in this tool because targets are usually MP3/WAV while sources are usually AAC), the tool can do a stream copy (-c:a copy) instead of re-encoding. Stream copy preserves the original bits exactly and finishes in seconds even for hour-long videos. For all other cases the tool re-encodes, which is bound by your CPU. Progress messages stream from FFmpeg's stderr and update the on-screen progress bar in real time.

Brief history of audio codecs and extraction

• MP3 (MPEG-1 Layer III), 1993. Fraunhofer Institute publishes the MP3 specification. The format dominates the late 1990s and early 2000s consumer audio scene: Napster (1999), iPod (2001), iTunes Music Store (2003). Patent restrictions limited free implementations until 2017, when the last MP3 patents expired.
• AAC (Advanced Audio Coding), 1997. MPEG-2 Part 7 and later MPEG-4 Part 3. Achieves similar quality to MP3 at lower bitrates. Apple adopts AAC for iTunes Music Store (2003) and iPhone (2007). AAC becomes the default audio codec in MP4 containers.
• FLAC (Free Lossless Audio Codec), 2001. Lossless compression cutting raw audio size by roughly half. Becomes the standard for archival audio, audiophile distribution, and CD ripping. Open source and royalty-free from inception.
• Vorbis and OGG container, early 2000s. Xiph.Org Foundation publishes Vorbis as a royalty-free alternative to MP3, packaged in the OGG container. Adopted by some games (Quake III, Halo) and open-source projects, but never reaches mainstream consumer adoption.
• Opus, 2012. IETF standardizes Opus, combining elements of SILK (voice, from Skype) and CELT (music). Becomes the dominant codec for low-latency real-time communication (Discord, WebRTC, Zoom) and increasingly for streaming (YouTube WebM, Spotify Web). At 96 kbps it rivals MP3 at 192 kbps.
• ffmpeg.wasm and browser audio extraction, 2019 to 2026. Jerome Wu publishes the first ffmpeg.wasm in 2019. By 2024 the 4.x line stabilizes at ~30 MB with SharedArrayBuffer multi-threading. Browser-side audio extraction from video becomes practical, replacing many cloud-based extraction services for personal use.

How It Works

1. Upload your video: Select any video file (MP4, WebM, MOV, AVI) containing the audio you want to extract.
2. Choose audio format: Select MP3, WAV, OGG, or AAC as the output format and set the audio quality/bitrate.
3. Extract and download: Click Extract Audio and download the audio file when processing is complete.

Why Use Video to Audio Extractor?

Extracting audio from video is a common need, pulling music from a video file, getting voiceover audio from a recording, saving podcast audio from a video interview, or extracting commentary from a gameplay video. Dedicated audio software or command-line tools are overkill for this simple task. This browser-based extractor strips the audio track from any video and saves it as a standalone audio file without any software installation.

Features

• Multiple audio formats: Export as MP3 (universal), WAV (lossless), OGG (open), or AAC (Apple/mobile).
• Bitrate control: Choose audio quality from 64 kbps (voice/podcast) to 320 kbps (high-quality music).
• Trim audio: Optionally specify start and end times to extract just a segment of the audio.
• Channel and sample rate: Advanced options for mono/stereo and sample rate for specialized audio use.
• Privacy-first: Video files are processed locally in your browser, never uploaded to any server.

Real-world extraction workflows

• Pulling audio from a video podcast recording. Many podcasts record video and audio simultaneously (Riverside, Zencastr, even Zoom). Extracting the audio track gives you the podcast-ready MP3 to upload to the podcast hosting platform. For voice-only content, 128 kbps mono MP3 is plenty; for music-heavy podcasts, 192 to 256 kbps stereo.
• Saving lecture audio for offline listening. Recorded lectures, conference talks, or webinars often have far more video than you need (slides barely change). Extracting just the audio cuts file size dramatically (a 1 GB hour-long video becomes a 30 MB MP3) and lets you listen during a commute or workout without the video drain.
• Ripping music from a music video. Sometimes the only version of a song you have is in a music video. Extracting to high-bitrate MP3 (256 to 320 kbps) or FLAC for lossless gives you a playable audio file for your music library. Make sure you have the rights or it's for personal use only.
• Saving voiceover or narration tracks. If you recorded a video tutorial with a high-quality voiceover and want to repurpose just the audio (for a podcast, audiobook, or different video), extraction gives you the standalone audio file to mix or remaster separately.
• Audio track for dubbing or subtitling. Translators and dubbing studios often work from extracted audio rather than full video, especially when bandwidth or storage matters. Extracting at lossless WAV gives the cleanest possible source for the translation/dubbing workflow.
• Pre-processing for transcription services. Most automated transcription services (Whisper, Otter, Rev) accept audio uploads faster than video uploads. Extracting to MP3 before uploading saves time and bandwidth. Audio-only inputs also tend to transcribe faster than video on the same service because no video decode is needed server-side.

Common pitfalls and what they mean

• Lossy-to-lossy re-encoding adds quality loss. Extracting AAC from an MP4 to MP3 means transcoding from lossy AAC to lossy MP3, which adds another generation of quantization noise. The result is often indistinguishable but slightly degraded. For pure preservation, extract to FLAC or WAV (lossless) and convert to MP3 later if needed.
• Bitrate higher than source doesn't help. If the video's audio is encoded at 128 kbps AAC, extracting to 320 kbps MP3 doesn't recover detail the source didn't have. It just bloats the file. Match the source bitrate or step slightly higher if you plan to do more processing; going dramatically higher is wasted space.
• Silent videos have no audio to extract. Screen recordings without a microphone, time-lapse videos, drone footage shot in noise-isolating conditions: these may contain no audio stream at all, or a stream of pure silence. The tool detects no-audio cases and notifies you; for pure-silence streams, the extracted file will be technically valid but inaudible.
• Multi-channel surround tracks downmix to stereo. Some videos (Blu-ray rips, 5.1 conference recordings) have multi-channel surround audio. Extracting to MP3 or AAC typically downmixes to stereo, losing the spatial separation. To preserve channels, extract to FLAC or multi-channel AAC; standard MP3 has no native multi-channel support.
• ID3 tags don't come for free. Video files don't carry MP3-style ID3 metadata (title, artist, album, cover art). Extracted MP3s come out with empty tags. If you need tags filled, use a desktop tool like Mp3tag or MusicBrainz Picard after extraction to add titles, artists, and cover art.
• DRM-protected video isn't supported. Netflix downloads, Apple TV+ files, and other DRM-protected video cannot be extracted by this tool (or any consumer tool). The audio is encrypted in the container. Extraction only works on unencrypted files you own.

Privacy: your video never leaves your device

Cloud video-to-audio services (Online Audio Converter, Audio Extractor, Convertio, dozens more) all upload your full video, run FFmpeg on their hardware, and send back the extracted audio. For a 200 MB phone video that's 200 MB up plus 10 to 50 MB down through their infrastructure. Audio content commonly includes voices of family or coworkers, conversations from meetings, private music, narration of sensitive topics. Most operators publish privacy policies committing to delete uploads within 1 to 24 hours and encrypt in transit, and major ones hold ISO/IEC 27001 certifications. They have strong business reasons to honor those policies. But "deleted within an hour" is not "never seen." During that window the file sits on operator infrastructure, accessible to any process or person with the right permissions, visible in logs and backups per the operator's retention policy.

This tool never uploads anything. The full pipeline (file selection, decoding via browser-native readers, extraction via ffmpeg.wasm WebAssembly, download via the browser's blob API) runs inside your browser tab. No uploads, no network requests carrying video data, no log entries. You can verify by opening browser dev tools on the Network tab before extracting: no requests leave with video content. Only the initial page load and the one-time roughly 30 MB ffmpeg.wasm download (cached for subsequent visits) touch the network. Put the browser in airplane mode after page load and the extractor still works on local files.

When another tool is the right pick

• Files over 2 GB. Browser memory limits become a wall above roughly 2 GB. Use desktop FFmpeg CLI: ffmpeg -i big_video.mkv -vn -c:a copy big_audio.aac works in seconds on any file size because it streams from disk and uses stream copy when codecs match.
• Multi-track audio extraction. Some videos have multiple audio tracks (director's commentary, alternate dub languages, narration variants). FFmpeg CLI with -map 0:a:N selects each track explicitly. Browser tools typically extract only the default audio track.
• Batch extraction across many files. A shell script with FFmpeg CLI processes a folder in one command: for f in *.mp4; do ffmpeg -i "$f" -vn -c:a libmp3lame -b:a 192k "${f%.mp4}.mp3"; done. Much faster than running a browser tool 100 times manually.
• Heavy ID3/metadata workflow. If you need to attach proper ID3 tags, cover art, lyrics, and chapter markers to extracted MP3s, use Mp3tag (Windows) or MusicBrainz Picard (cross-platform) after extraction. These have rich metadata editors that browser tools lack.

Frequently Asked Questions

Will the extracted audio quality match the video?

Yes, if you extract to a lossless format (WAV) or at the same bitrate as the video's audio track. The video's audio is already encoded, re-encoding at a lower bitrate will reduce quality. For best results, choose WAV or the highest bitrate option.

What if the video has no audio track?

If the video has no audio track, the tool will notify you. Silent videos (like screen recordings without microphone) contain no audio data to extract.

Can I extract audio from YouTube videos?

This tool works with video files from your device. Downloading or extracting audio from YouTube videos may violate YouTube's Terms of Service. Use files you own or have permission to process.

Other frequently asked questions

What's the difference between MP3, AAC, OGG, and WAV?

MP3 (1993) is the universal lossy format, plays everywhere. AAC (1997) is more efficient than MP3 at the same quality, default on Apple devices. OGG/Vorbis (early 2000s) is the royalty-free open alternative; Opus (2012) is its modern successor and the most efficient lossy codec for voice and music. WAV is uncompressed PCM (huge files but lossless). FLAC is lossless compression (about half the size of WAV with no quality loss). For sharing: MP3 or AAC. For archiving: FLAC. For editing: WAV.

What bitrate should I pick?

For music: 192 to 320 kbps MP3 or AAC (320 is transparent for most listeners; 192 is the de facto library default). For voice/podcasts: 64 to 128 kbps mono is plenty (smaller files, no quality loss for speech). For archival: don't pick a bitrate, use lossless (FLAC or WAV). Picking a bitrate higher than the source's audio bitrate doesn't help; match or step slightly higher only.

Will extraction be lossless or lossy?

Depends on the target format. Extracting to WAV or FLAC is lossless: every sample of the decoded source is preserved. Extracting to MP3, AAC, OGG, or Opus is lossy: the encoder discards some inaudible detail to save space. If the source audio was already lossy (most MP4s have AAC inside), even a lossless extract is a "lossless copy of lossy source," not a perfect reproduction of the original recording.

Why is extraction sometimes near-instant and sometimes slow?

If the source audio codec matches the target (rare in this tool because targets are usually MP3/WAV while sources are usually AAC), the tool can do a stream copy that just rewrites the container without re-encoding. Stream copy finishes in seconds for any file size. Re-encoding is bound by your CPU and takes time proportional to the audio duration: 1 to 5x real-time for MP3/AAC on a typical laptop.

Is there a desktop or command-line equivalent?

Yes. FFmpeg CLI: ffmpeg -i input.mp4 -vn -c:a libmp3lame -b:a 192k output.mp3. For stream copy: ffmpeg -i input.mp4 -vn -c:a copy output.m4a. VLC has a Convert/Save dialog that extracts audio through its built-in FFmpeg. Audacity can import video and export audio in any format. All produce essentially identical output to this browser tool because they share the same underlying codecs.

Can I extract just a portion of the audio?

In this tool, not directly: the extraction processes the full audio track. To extract a segment, use a video trimmer first to cut your video to the desired range, then extract audio. Or in FFmpeg CLI: ffmpeg -i input.mp4 -vn -ss 00:01:30 -to 00:03:45 -c:a libmp3lame -b:a 192k clip.mp3 extracts from 1:30 to 3:45.