Extracting audio for video editing, podcasts and transcription

Recommended settings by task

Konvert offers five formats: MP3 at 128, 192 or 320 kbps, plus M4A, OPUS, WAV and FLAC. Here's where each one fits. The reasoning for every row is explained further down.

Not sure what the formats themselves are? Our audio formats guide compares all five, and the MP3 bitrate guide explains the difference between 128, 192 and 320 kbps.

Before you start

Three things are worth checking before you paste a link.

• You have the right to use it. Your own uploads, public-domain material, Creative Commons content that allows your use, and anything you have permission for are all fair game. Everything else deserves a second look. Our copyright basics guide walks through it.
• The source sets the ceiling. Konvert saves the best audio stream the video offers. A choppy phone recording stays choppy in FLAC. Higher settings avoid losing quality; they can't add any.
• It fits the length limit. Videos longer than 1 hour aren't supported, so split long projects across shorter videos where you can.

Video editing

Pulling audio out of a video is handy when you want to reuse a voice-over, replace a noisy camera track, or build a music bed from a performance you recorded and uploaded yourself.

Why WAV or FLAC beats MP3 here

Video platforms store audio in a compressed, lossy format. That's the first generation of loss, and nothing can undo it. If you save it as MP3, you add a second lossy encode. When you export your finished video, your editor compresses the audio again, which makes a third.

WAV and FLAC stop the chain at one. Konvert decodes the original stream once and stores the result exactly, so the only lossy step left is your final export. The file sounds the same as the source, not better, but it won't get any worse on the way through your project.

Sample rate: expect 48 kHz

Most video projects run at 48 kHz, the long-standing standard for video. Konvert doesn't resample: WAV and FLAC files keep the sample rate of the stream they came from. On YouTube the best audio stream is usually Opus, which is 48 kHz, so your file will usually match a video timeline exactly. If a file ever arrives at 44.1 kHz, modern editors convert it on the fly when you drop it on the timeline.

Importing into your editor

Every mainstream video editor imports WAV, whether it's a professional app on your desktop or a simple editor on your phone. FLAC support is less universal, so if an editor refuses a FLAC file, convert again as WAV.

1. Download the file and move it into your project folder before importing, so the project doesn't lose track of it later.
2. Give it a descriptive name. Konvert names files after the video title, which may not mean much in a crowded media bin.
3. Import or drag it into your media bin, then onto an audio track.
4. If you're lining it up with other footage, zoom in on the waveforms and nudge the clip until distinct peaks, like a clap or a hard consonant, match.

Trim before you import

Konvert always saves the full audio track. If you only need a few seconds, trimming first keeps your project light. A free, open-source audio editor such as Audacity handles this well: open the file, select the part you need, and export the selection as WAV. Keep in mind that WAV takes about 10–12 MB per minute, so a long recording adds up quickly.

Podcasts

A podcast might use extracted audio for a guest's clip, a moment from a talk you're discussing, or your own video appearance repurposed for an audio feed.

Clear the rights for every clip

Once you publish an episode, you're distributing whatever's in it. Before a clip goes in, make sure it's yours, it's in the public domain, its license allows reuse (and you follow the license's terms, such as crediting the creator), or you have permission. Short excerpts used for commentary or criticism may qualify as fair use or fair dealing in some countries, but that's a narrow, case-by-case judgment, not a general rule. Crediting the source is good practice. On its own, it isn't permission.

A simple editing workflow

1. Extract the clip as WAV.
2. Trim it to the part you need, adding a short fade in and out so it doesn't start or stop abruptly.
3. Place it in your episode and match its level to your voice.
4. Export the finished episode once, in your publishing format.

Loudness basics

Listeners notice volume jumps more than almost anything else. A clip pulled from a video is often mastered much louder or quieter than your recorded voice, so level it by ear first, then check with a meter. Loudness is usually measured in LUFS. Many podcast guidelines aim for an integrated loudness of around −16 LUFS for stereo, with true peaks no higher than about −1 dBFS, but check what your podcast host recommends. Many audio editors, including Audacity, include a loudness normalization effect that does most of the work for you.

Choosing a publishing format

MP3 is still the most widely compatible format for podcast feeds. 192 kbps leaves comfortable headroom for music and sound effects, and spoken-word shows often use less. M4A (AAC) is a good alternative that sounds slightly better at the same size. Either way, you'll export this from your editor, not from Konvert. Konvert's MP3 192 option is useful when you just want to listen back to a source before deciding whether to use it.

Transcription and captions

If you're turning a talk into notes, an article or a caption file, the format matters much less than you might think. Transcription services and speech-recognition apps listen for clear speech, and every format Konvert offers carries speech with no audible difference.

• Choose OPUS or M4A for the smallest files, which upload fastest. OPUS takes about 1 MB per minute, and on YouTube it's often the original stream saved without re-encoding.
• If a tool rejects OPUS, use M4A or MP3. Both are accepted almost everywhere.
• Timestamps line up. Konvert saves the complete audio track from the beginning, so times in your transcript match times in the video.
• Review before you publish. Automatic transcripts get names, jargon and overlapping speakers wrong. Budget time for a careful read-through, especially for captions other people will rely on.

If the video's creator already publishes captions or a transcript, start there. It may save you the whole job.

Study and language learning

Audio is easy to take with you. A lecture you'd never sit and watch again becomes something you can listen to on a walk, and a short dialog in a new language becomes something you can loop until it sticks.

• Stick to material you're allowed to access, like lectures your course publishes for students, talks released under an open license, or your own recordings. If a course's terms say its videos are for streaming only, respect that.
• MP3 at 128 kbps is plenty for speech and keeps a semester's worth of lectures small. OPUS is even smaller if your player supports it.
• Use playback speed. Most podcast and audio apps can slow speech down for shadowing practice or speed it up for review.
• Cut phrases into short clips. For language practice, trimming a sentence or two into its own file (Audacity works well for this) makes repetition much easier than scrubbing through a long track.

Sampling and music production

Sampling is where rights matter most, because your track may end up distributed far and wide. Work only with material you've made, material in the public domain, or material licensed for it: sample packs, royalty-free libraries, or Creative Commons works whose license allows what you plan to do.

• Read the license. A Creative Commons license with NC rules out commercial releases, and one with ND rules out remixing entirely. Royalty-free usually means no ongoing fees, not no conditions.
• Use WAV for your DAW. Every music production app loads it natively, and there's no extra compression step before you start chopping and time-stretching.
• Keep a record. Save the source link, the creator's name and the license alongside your project. If a question ever comes up, you'll have the answer, and you'll need the details for attribution anyway.

See copyright basics for how Creative Commons licenses work and how to credit them properly.

Limits to plan around

Maximum length

1 hour per video

Per conversion

One video

Per hour

Up to 10 conversions

File lifetime

About 30 minutes

Longer recordings aren't supported, and playlists and channels aren't accepted, so paste one video link at a time. Files are deleted about 30 minutes after you start a conversion, so download yours right away.

If a conversion doesn't go through, the troubleshooting guide explains every message you might see. For general questions, see the FAQ, or read how to use Konvert for a quick walkthrough.

Questions about editing workflows