How to extract podcast audio from a video recording
- Step 1Export the recording from your platform — Save the session as a video file: Zoom and Teams produce MP4; OBS commonly produces MKV or MP4; Riverside can give you a video file; a camera gives MOV/MP4. Any of these works as the source.
- Step 2Open the extractor — Load audio-track-extractor. FFmpeg.wasm initialises in-browser; nothing is uploaded — important for guest confidentiality.
- Step 3Drop the episode recording — Add one file (MP4/MOV/MKV/WebM/AVI). Free tier caps a single source at 1 GB; a long episode at modest video resolution usually fits. Pro raises this to 10 GB.
- Step 4Pick the format for your edit — In the Format dropdown: WAV for an editing master, FLAC for a lossless archive, or MP3 for a quick reference. WAV/FLAC are the right call before any cleanup work.
- Step 5Run the extraction — The engine runs
-vn -c:a pcm_s16le(WAV),-c:a flac(FLAC), or-c:a libmp3lame -b:a 192k(MP3). It extracts the default audio track in one pass — no mixing, no processing. - Step 6Take the audio into your editor — Import the WAV/FLAC into your DAW for noise reduction and loudness normalisation (this tool doesn't do that step). Need to anonymise the source video's metadata first? Use metadata-scrubber.
Best output format for each podcast step
All five live in the one Format dropdown. For podcasting, pick a lossless master for editing and a portable copy for sharing.
| Step | Format | Why |
|---|---|---|
| Editing in a DAW | WAV | 16-bit PCM, opens natively everywhere |
| Lossless archive of the raw audio | FLAC | Lossless, ~half the size of WAV |
| Quick reference / send to co-host | MP3 | 192 kbps, plays on any device |
| Source is AAC, want untouched copy | AAC | Stream-copy, bit-identical (no re-encode) |
| Voice-message snippet | Opus | 128 kbps, efficient for speech |
Recording platform to source file
What each common podcast-recording setup produces, and how the extractor handles it.
| Platform | Typical export | Audio codec | Extract note |
|---|---|---|---|
| Zoom | MP4 | AAC | AAC copy or WAV master both work |
| Riverside | MP4 / WebM | AAC / Opus | WAV/FLAC re-encode handles either |
| OBS Studio | MKV / MP4 | AAC (or PCM if configured) | WAV master for editing |
| Microsoft Teams | MP4 | AAC | Default track only — single mix |
| Camera (DSLR/phone) | MOV / MP4 | AAC / LPCM | WAV preserves a PCM source losslessly |
Cookbook
Real podcast-extract jobs. The extractor's role is to get clean audio out of the video; mixing and mastering are separate downstream steps.
Zoom interview MP4 to a WAV editing master
A recorded Zoom interview comes out as MP4 with AAC audio. WAV gives a lossless PCM master to edit before cleanup.
Source: ep42-zoom.mp4 (62 min, AAC) Format: WAV Command: ffmpeg -i ep42-zoom.mp4 -vn -c:a pcm_s16le ep42-zoom.wav Result: ep42-zoom.wav (16-bit PCM, ready for the DAW) Then: noise reduction + loudness in your audio editor.
OBS MKV to FLAC archive
An OBS capture in MKV. You want a lossless archive of the raw episode audio without WAV-sized files.
Source: show-raw.mkv (95 min) Format: FLAC Result: show-raw.flac (lossless, ~45% smaller than WAV) Keep this as the archival master; edit a copy.
Quick MP3 reference to send a co-host
Before editing, you want to send the co-host a listenable cut of the recording. MP3 at 192 kbps is small and universal.
Source: raw-session.mp4 Format: MP3 Result: raw-session.mp3 (192 kbps) Good enough to review; not the editing master.
Camera MOV with PCM audio kept lossless
Some cameras record LPCM. Extracting to WAV keeps that signal lossless (it re-encodes into the WAV container but loses nothing).
Source: cam-A.mov (LPCM audio) Format: WAV Result: cam-A.wav (16-bit PCM, lossless) Use this for sync against other camera angles in the DAW.
When you expected stems but got one track
If your platform recorded a single mixed track (most do by default), the extractor gives you that one mix — not per-speaker stems. The tool can't split a mixed recording into separate voices.
Source: mixed-session.mp4 (single mixed stereo track) Format: WAV Result: one stereo WAV (the mix) — NOT separate stems For per-speaker stems, record local tracks per guest at source (Riverside/Zoom local recording) before extracting.
Edge cases and what actually happens
Recording has separate-speaker tracks but only one is extracted
By designThe extractor pulls the default (first) audio stream — there's no track-picker. If your file is a true multitrack recording (one stream per speaker), you get the primary track only, not all stems. Capture local per-speaker recordings at source if you need separate stems.
This tool doesn't clean noise or normalise loudness
By designExtraction is a clean demux/re-encode, not a mastering step. Hum removal, de-noise, EQ, and loudness normalisation all happen downstream in your DAW. Extract a lossless WAV/FLAC first so the cleanup has the best source to work from.
Long episode hits the 1 GB Free cap
RejectedA two-hour high-resolution recording can exceed 1 GB. Either upgrade (Pro = 10 GB) or first trim dead air at the start/end with lossless-trimmer to shrink the source before extracting.
Source has no audio (camera mic was muted)
FailsIf the recording captured no audio, -vn leaves nothing to encode and FFmpeg errors out. Confirm the file actually has sound — a common surprise with cameras whose external-mic input wasn't enabled.
WAV master is huge for a long episode
ExpectedUncompressed PCM is ~10 MB/minute stereo, so a 90-minute episode is ~900 MB. That's near the Free cap on the way out and uses disk. Choose FLAC for the archive (lossless, about half the size) and edit from there.
Audio plays slightly out of sync after import
PreservedThe extracted audio keeps the source's original timing, including any track offset some recorders write. When syncing against a separate video edit, you may need to nudge the audio in your DAW to line it up — the extractor preserves rather than corrects timing.
WebM/Opus source and you chose AAC
FailsRiverside and browser recordings often produce WebM with Opus audio. AAC stream-copy can't wrap Opus packets, so it fails. Choose WAV/FLAC (lossless re-encode) or MP3 for these sources instead.
Want to remove guest's identifying metadata from the video
Use metadata-scrubberAudio extraction doesn't strip the source video's metadata (it just makes a new audio file). If the original recording carries identifying data you need gone, run the video through metadata-scrubber before or alongside extraction.
Frequently asked questions
What format should I extract for podcast editing?
WAV (16-bit PCM) for the editing master — it opens natively in every DAW and is lossless. FLAC if you want a lossless archive at about half the size. MP3 only for quick reference cuts, since it re-encodes to lossy 192 kbps.
Does this clean up the audio or normalise loudness?
No. The extractor produces a clean audio file from the video — it doesn't de-noise, EQ, or normalise. Do that in your DAW after extracting. Pull a lossless WAV/FLAC so the cleanup works from the best possible source.
Can I get separate tracks for each speaker?
Only if your recording stored them separately and the one you need is the default stream — and even then the tool extracts a single (default) track, not all stems. For true per-speaker stems, use local per-guest recording at the source platform and extract each file individually.
Is my interview uploaded anywhere?
No. Everything runs in your browser via FFmpeg.wasm, so the recording is processed locally and never sent to a server. That's what makes it safe for NDA guests and embargoed conversations.
My Zoom recording is an MP4 — will this work?
Yes. Zoom MP4s carry AAC audio. You can stream-copy it untouched (Format = AAC) or re-encode to WAV/FLAC/MP3. For editing, a WAV master is the usual choice.
My Riverside file is WebM — why does AAC fail?
Browser-based recorders often produce WebM with Opus audio, and AAC stream-copy can't wrap Opus packets. Choose WAV or FLAC (lossless re-encode) or MP3 instead — those work regardless of the source codec.
How long an episode can I extract?
There's no duration cap — only file size. Free tier allows sources up to 1 GB, which covers most episodes at modest video resolution. Pro raises it to 10 GB and Pro + Media to 100 GB for very long, high-bitrate recordings.
Will the audio be in sync when I edit?
The extract preserves the source's original timing. If you're cutting audio against a separate video edit, you may need to nudge it in your DAW to align — extraction keeps timing as-is rather than re-syncing.
Can I extract from several episode files at once?
No — this tool processes one file per run. Extract each recording individually. The extractor doesn't accept multiple inputs.
Do I need software like Audacity or FFmpeg installed?
No. The FFmpeg engine runs as WebAssembly in the browser. You only need a DAW (Audacity, Reaper, etc.) afterward for the actual editing — the extraction itself needs nothing installed.
Should I scrub the video before extracting for privacy?
If the original recording carries identifying metadata you want gone, run it through metadata-scrubber. The audio extractor makes a fresh audio file but doesn't itself strip the source video's metadata.
What's the fastest extract if my source is already AAC?
Choose AAC — it stream-copies (-c:a copy) with no encode pass, so it's near-instant and bit-identical. See the AAC stream-copy guide. Use WAV/FLAC instead when you want a lossless master for editing.
Privacy first
Every JAD Video tool runs entirely in your browser via WebCodecs and FFmpeg (WebAssembly). Your video files never leave your device — verified by zero outbound network requests during processing.