How to extract lossless wav audio from any video — free
- Step 1Drop your video onto the tool — Drag an MP4, MOV, MKV, WebM, or AVI onto the dropzone above (it also accepts audio files — MP3, WAV, FLAC, M4A, OGG, Opus). The file is read into the page; nothing uploads. FFmpeg 8.1 loads its WASM core on first use, then stays warm for subsequent files.
- Step 2Confirm the detected sample rate and channels — The file card shows the source sample rate (e.g.
48,000 Hz) and channel layout (stereo / mono). Use this to decide whether to keep the native rate or resample — most video is48 kHz, most music delivery is44.1 kHz. - Step 3Pick the output sample rate — Choose
44.1 kHz · CD,48 kHz · video,96 kHz · studio, or16 kHz · speech. Leaving it on the source rate avoids an extra resample. There is no "original" passthrough toggle — pick the value that matches your source if you want a clean copy. - Step 4Pick stereo or mono —
Stereo(default) preserves both channels.Monodownmixes to a single channel — correct for a one-mic interview where both channels are identical, and it halves the WAV size. - Step 5Run the extraction — Press the action button. FFmpeg runs
-c:a pcm_s16lewith your-ar(sample rate) and-ac(channels) flags, carrying source metadata across with-map_metadata 0. Progress is local CPU work — a 10-minute 1080p clip is typically a few seconds. - Step 6Download the WAV — Save the resulting
.wav. It is a standard 16-bit PCM file ready to drop straight into Reaper, Audition, Logic, Pro Tools, Ableton, or a transcription pipeline — no further conversion needed.
The two controls (and what they actually do)
These are the only options the video-to-wav tool exposes. The output codec is fixed at pcm_s16le (16-bit) — there is no bit-depth selector.
| Control | Choices | Default | FFmpeg flag | When to change it |
|---|---|---|---|---|
| Sample rate | 44.1 kHz · CD / 48 kHz · video / 96 kHz · studio / 16 kHz · speech | 44.1 kHz | -ar 44100 / -ar 48000 / -ar 96000 / -ar 16000 | Match 48 kHz for video timelines; drop to 44.1 kHz for music delivery; 16 kHz for speech-to-text |
| Channels | Stereo / Mono | Stereo | -ac 2 / -ac 1 | Mono for single-mic interviews — halves file size with no quality loss |
| Bit depth | (not exposed — fixed) | 16-bit | -c:a pcm_s16le | Not adjustable in this tool. Need 24-bit? FFmpeg here only emits 16-bit PCM for this slug |
| Output format | (fixed) | WAV | container .wav | This tool always writes WAV. For MP3 use the sibling video-to-mp3 tool |
Sample-rate presets mapped to real workflows
Pick the rate that matches the destination, not the largest number. Resampling 48 kHz video down to 44.1 kHz is a one-time, transparent operation.
| Preset | Rate | Typical source | Best for |
|---|---|---|---|
| CD | 44,100 Hz | Music videos, mastered tracks | Music delivery, CD-quality archives, Spotify/Apple-bound stems |
| Video | 48,000 Hz | Most cameras, screen recordings, NLE exports | Keeping the native rate of camera/NLE audio — avoids an extra resample |
| Studio | 96,000 Hz | High-res field recorders, archival masters | Long-term archival headroom and pitch/time work that benefits from oversampling |
| Speech | 16,000 Hz | Interviews, voice memos, lecture capture | Feeding Whisper / speech APIs that expect 16 kHz mono — smallest correct file |
Per-tier limits for video-to-WAV jobs
WAV is ~10 MB per stereo minute at 44.1 kHz, so the size cap and the duration cap can both bite. Values read from the live tier configuration.
| Tier | Max file size | Max duration / file | Files per job |
|---|---|---|---|
| Free | 50 MB | 30 minutes | 1 |
| Pro | 200 MB | 120 minutes | 10 |
| Pro-media | 100 GB | Unlimited | 100 |
| Developer | 100 GB | Unlimited | — |
Cookbook
Real extraction recipes. Each shows the source, the controls you set, and the FFmpeg command the tool runs under the hood.
Editable WAV master from a 48 kHz camera MP4
Camera and NLE audio is almost always 48 kHz. Keep that rate to avoid a needless resample, and keep stereo if the mics were panned. This is the master you hand to your DAW.
Source: interview-final.mp4 (H.264 + AAC 48 kHz stereo)
Controls: Sample rate = 48 kHz · video, Channels = Stereo
ffmpeg -i interview-final.mp4 -c:a pcm_s16le -ar 48000 -ac 2 \
-map_metadata 0 interview-final.wav
Output: interview-final.wav (16-bit PCM, 48 kHz, stereo)
~5.5 MB per minuteMono WAV from a single-mic talking-head clip
A lavalier or shotgun mic recorded to both channels gives you two identical tracks. Extracting mono halves the file with zero audible difference and simplifies the DAW session.
Source: vlog-take3.mov (single mic, duplicated to L+R)
Controls: Sample rate = 48 kHz · video, Channels = Mono
ffmpeg -i vlog-take3.mov -c:a pcm_s16le -ar 48000 -ac 1 \
-map_metadata 0 vlog-take3.wav
Output: vlog-take3.wav (16-bit PCM, 48 kHz, mono)
~2.75 MB per minute (half the stereo size)44.1 kHz WAV for a music-video stem headed to a DAW
A music video shot at 48 kHz but destined for a 44.1 kHz session: resample once here so every later bounce stays sample-rate-consistent.
Source: band-liveset.mkv (48 kHz stereo)
Controls: Sample rate = 44.1 kHz · CD, Channels = Stereo
ffmpeg -i band-liveset.mkv -c:a pcm_s16le -ar 44100 -ac 2 \
-map_metadata 0 band-liveset.wav
Output: band-liveset.wav (16-bit PCM, 44.1 kHz, stereo)Extract, then convert the finished edit back to MP3
WAV is the working format, not the delivery format. After you have edited the WAV in a DAW, send the bounce to a lossy format for distribution using the sibling converter.
Step 1 (this tool): screencast.webm -> screencast.wav (16-bit, 48 kHz)
Step 2 (DAW): edit, EQ, normalise -> screencast-final.wav
Step 3 (sibling): /audio-tools/wav-to-mp3 -> screencast-final.mp3
Why: each lossy encode is one-and-done at the very end,
so no artefacts compound during editing.Tighten silence on the WAV before you import it
Long dead-air at the head or between takes wastes DAW timeline space. Pull the WAV here, then strip the pauses with the silence tool — both run locally, no re-upload.
Step 1 (this tool): zoom-recording.mp4 -> zoom-recording.wav
Step 2 (sibling): /audio-tools/silence-stripper
threshold -40 dB, min silence 0.5 s
Result: a tight WAV with the gaps removed, still 16-bit PCM,
ready to drop on the timeline.Edge cases and what actually happens
Source has no audio stream
Fails (no audio)A screen recording captured with audio disabled, or a B-roll clip with a muted track, has no audio stream for FFmpeg to demux. The job errors rather than producing a silent WAV. Check the source plays with sound before extracting.
Free tier — file over 50 MB or longer than 30 minutes
Rejected (tier limit)The Free tier caps each file at 50 MB and 30 minutes. A long video can trip the duration cap even when the byte size looks fine. Trim the source first, or upgrade — Pro is 200 MB / 120 min, Pro-media is 100 GB with no duration limit.
Expecting 24-bit or 32-bit-float WAV
By design (16-bit only)This tool always writes 16-bit PCM (pcm_s16le). There is no bit-depth selector. 16-bit is correct for almost all editing and transcription; if you genuinely need 24-bit for high-headroom mastering, this is not the right tool.
Multi-language / multiple audio tracks in the container
First track onlyAn MKV or MP4 with several audio tracks (e.g. an English and a commentary track) yields the first audio stream FFmpeg maps. There is no track picker. To choose a specific language track, demux it in a desktop FFmpeg first, then bring the single-track file here.
5.1 surround source
DownmixedIf the video carries 5.1 audio and you select Stereo, FFmpeg downmixes 6 channels to 2 using its standard matrix. Selecting Mono collapses further to one channel. There is no surround passthrough — WAV output here is mono or stereo only.
WAV is bigger than the source video
ExpectedUncompressed PCM has no codec savings, so a heavily-compressed clip can produce a WAV larger than the original file. That is normal and the point of WAV — it is the editing master, not the delivery file. Convert to MP3/AAC for distribution.
Variable-frame-rate (VFR) screen recording
SupportedVFR affects video timing, not the audio stream. Audio extraction is unaffected and the WAV duration matches the audio track length. (If you later need to re-sync to video, the audio is the reliable reference.)
Corrupt or partially-downloaded video
Fails (demux error)If the container is truncated or the moov atom is missing (common with interrupted downloads), FFmpeg cannot demux it and the job errors. Re-download the source or repair the container in a desktop tool before extracting.
DRM-protected video
Rejected (encrypted)Files with DRM/encryption (e.g. some purchased downloads) cannot be decoded by FFmpeg and will fail. This is by design — the tool only handles media you can actually decode.
Frequently asked questions
Why extract to WAV instead of MP3?
WAV is uncompressed 16-bit PCM, so no lossy artefacts get added. If you plan to edit — EQ, compress, normalise, denoise — start from WAV so quality does not degrade across passes. For final delivery where size matters, convert the finished WAV to MP3 with the sibling wav-to-mp3 tool.
What bit depth is the output?
Always 16-bit (pcm_s16le). This tool does not expose a 24-bit or 32-bit-float option. 16-bit gives ~96 dB of dynamic range, which is plenty for editing, transcription, and almost all production work.
What sample rates can I choose?
Four: 44.1 kHz (CD), 48 kHz (video), 96 kHz (studio), and 16 kHz (speech). The default is 44.1 kHz. If you want a clean copy with no resampling, pick the rate that matches your source — most video is 48 kHz.
Which video formats are accepted?
Anything FFmpeg can demux — MP4, MOV, MKV, WebM, AVI, and others. The tool reads the audio stream inside (AAC, AC-3, MP3, PCM, Opus) and re-wraps it as PCM WAV. The dropzone also accepts plain audio files (MP3, WAV, FLAC, M4A, OGG, Opus).
Can I use the WAV in Reaper, Audition, Logic, or Pro Tools?
Yes. 16-bit PCM WAV is the most universally compatible audio format. Every major DAW — Logic, Audition, Reaper, Pro Tools, Ableton — imports it natively with no conversion.
Is my video uploaded anywhere?
No. Extraction runs entirely in your browser via FFmpeg 8.1 compiled to WebAssembly. The video and the resulting WAV never leave your device, which is why this is safe for unreleased footage and client work.
How big will the WAV be?
Roughly 10 MB per minute for 44.1 kHz / 16-bit stereo, ~11 MB/min at 48 kHz stereo. Mono halves that. WAV has no codec compression, so plan for a file that can be larger than the source video.
Can I extract a specific audio track from a multi-track MKV?
No — the tool maps the first audio stream FFmpeg finds; there is no track picker. To isolate a particular language or commentary track, demux it with a desktop FFmpeg first, then bring the single-track file here.
What is the largest file I can process?
Free: 50 MB and 30 minutes per file. Pro: 200 MB / 120 min, 10 files per job. Pro-media: 100 GB with no duration limit and 100 files. These are real per-file limits — the duration cap is separate from the size cap.
Will it keep the original sample rate automatically?
There is no automatic passthrough — you choose from the four presets. To avoid resampling, read the source rate on the file card and pick the matching value (e.g. 48 kHz for most camera footage).
Can I extract mono to save space?
Yes. Set Channels to Mono. For a single-mic source where both channels are identical, this halves the WAV size with no audible difference. For genuinely stereo content, keep Stereo.
What if I need to clean up the audio after extracting?
Extract the WAV here, then chain a sibling tool — silence-stripper to cut dead air, ai-noise-reducer for background hiss, or loudness-normalizer for consistent levels. All run locally, so no re-upload.
Privacy first
Every JAD Audio tool runs entirely in your browser via FFmpeg (WebAssembly) and RNNoise. Your audio files never leave your device — verified by zero outbound network requests during processing.