Turning a screen recording into audio

Screen recordings are ordinary video files, so extracting the audio is easy. The hard part is what got recorded in the first place.

A screen recording is an ordinary video file, so pulling the audio out of it is no different from any other extraction. The hard part is almost never the conversion — it is what the recorder captured in the first place.

What kind of file you are holding

In every one of those rows, MP3 is available too — it is simply a re-encode rather than a copy. If the recording is a meeting or a lecture you are going to listen to once, that is a perfectly sensible trade for a file that plays anywhere.

The silent recording problem

This is the actual reason most people end up searching. A screen recording can contain no audio at all, and the recorders do not make it obvious.

• iPhone. The screen recorder captures app audio by default, but the microphone is a separate toggle — long-press the record button in Control Centre to turn it on. If a recording captured neither (some apps block their audio from being recorded), there is nothing in the file.
• macOS. ⇧⌘5 records the microphone if you choose one under Options, and does not record system audio at all without an extra audio driver. A recording of a video call made this way contains your voice and none of theirs.
• Windows Game Bar. Captures game and system audio plus, optionally, the microphone. It refuses to record some windows entirely.
• Browser tab recording. Only includes audio when "Share tab audio" was ticked in the sharing dialog. Nobody ticks it the first time.

The extractor tells you this directly: if the file has no audio stream it says so, by name, rather than producing an empty MP3 and letting you discover the problem later.

Doing the extraction

1. Open the extractor and drop the recording in.
2. Check the panel: it shows the soundtrack's codec, sample rate, channel layout and bitrate — which is also the quickest way to tell whether the microphone was actually recorded, since mic-only captures are usually mono.
3. For a meeting or lecture, fold it to mono and choose 64 or 96 kbps. A one-hour recording comes out around 30 MB instead of 90, and speech sounds the same.
4. For anything musical, leave it in stereo and take M4A if you want it untouched.

Long recordings

Screen recordings get big — an hour of 4K can be several gigabytes. That is not a problem here, because the video is never loaded into memory: the extractor mounts the file and reads the audio packets as it walks through, skipping the picture entirely. The ceiling is around 2 GB for the file itself, above which a desktop FFmpeg is the right tool.

It also means extraction is fast in a way that surprises people: a copy of the soundtrack out of a 3 GB recording is seconds, because nothing decodes the picture at any point.

If it is going to a transcription service

Most of them cap uploads well below the size of an uncompressed hour, so do not reach for WAV by default. Mono at 64 kbps is plenty for speech recognition, is accepted by every major service, and turns an hour of meeting into roughly 30 MB. Only choose WAV when the service explicitly asks for uncompressed input.

On the phone

iPhone screen recordings land in Photos, which means the app can read them directly: pick the recording, choose MP3 or M4A and a bitrate, convert. No Save-to-Files step, no browser, and it works with no signal — useful on the train home from the lecture you just recorded.