Recording a screen with the sound that matters

You recorded the screen to show someone a problem, and the video has no sound. Or it has your voice but not the app's audio. Or it has the app's audio but you can hear the room. Screen recording is built into every phone and computer, and the part people get wrong is never the recording — it is which of two completely separate audio sources ended up in the file.

There are two microphone-shaped things involved and they are unrelated: system audio, meaning the sound the device itself is producing, and microphone audio, meaning the room. Every platform handles these differently, and the default is rarely what you wanted.

The two audio sources

Getting this straight solves most of the problem.

On iPhone and iPad, screen recording captures system audio by default and the microphone only if you turn it on. Press and hold the record button in Control Centre to reveal the microphone toggle — a single tap starts recording without ever showing you that switch, which is why so many recordings are missing narration.

On Android, the recorder asks each time what to record: no audio, device audio, microphone, or both. Read the dialogue rather than dismissing it, because the choice is made there and cannot be changed afterwards.

On a Mac, the built-in recorder captures the microphone but not system audio, which surprises people. Capturing what the Mac is playing requires either a screen-capture tool that supports it or an audio routing utility.

On Windows, the Game Bar records system audio and optionally the microphone.

The most common screen recording failure is not a missing feature. It is a microphone toggle that was never shown, because the button was tapped rather than held.

Before you press record

Five preparations that determine whether the recording is usable, in the order they matter.

1. Silence notifications. A message banner mid-recording is both a distraction and, frequently, a privacy problem. Turn on Do Not Disturb or a focus mode first.
2. Decide the audio sources, per the table above, and verify with a ten-second test rather than assuming.
3. Check what is visible. The status bar carries your carrier, battery and time; open tabs, bookmarks and desktop files carry more. Close what should not be in frame.
4. Rehearse the path once. A recording where you hunt for a menu is three times longer than one where you know the route.
5. Record a few seconds of nothing at the start and end, giving yourself something to trim to.

Point three is the one that produces real consequences. Screen recordings are shared, and they routinely include a notification, a browser tab title, or an autofill dropdown that the person never intended to send.

How do you get narration that sounds decent?

If your voice is in the recording, it is now an audio production problem, and three things do most of the work.

Get closer to the microphone. A phone at arm's length in a normal room picks up more room than voice. Holding it closer, or using earbuds with a microphone, is a bigger improvement than any processing.

Record somewhere soft. Hard walls and bare floors produce reverb that cannot be removed afterwards. A room with a rug, curtains or a sofa sounds dramatically better, and this is free.

Turn off anything that hums. Air conditioning, a fan, a fridge. Continuous noise is more objectionable to a listener than an occasional sound, and it is exactly what a listener notices within seconds.

The alternative approach is to record the screen silently and add narration afterwards. This is slower but it means a stumble costs one sentence rather than a whole take, and it produces a script you can reuse — a scripted voice track from a tool like SpeakFile also gives you the subtitle text for free, since the words already exist as text.

What do you do with the file afterwards?

A screen recording is long, large and in the wrong shape, and three steps fix it.

Trim the ends. The Photos app on both platforms does this without re-encoding, so it is instant and lossless. This alone removes the awkward start and finish.

Check the size. A few minutes of screen recording at full resolution is hundreds of megabytes, mostly because the frame rate is high and screen content changes constantly. If it is going by email or message, it needs to be smaller.

Reduce it deliberately rather than letting a messaging app do it. Exporting at a lower resolution, or converting at a controlled quality, produces a far better result than sending it through an app that recompresses it to fit its own limit.

For content that is more slideshow than demonstration, assembling stills into a video is smaller and clearer than a screen recording — which is what tools like Stillora are for, and it avoids the size problem entirely.

Recording a call or a meeting

This is the case with the most constraints, and two of them are not technical.

The legal position varies. Recording a conversation you are part of is permitted in some places and requires everyone's consent in others, and this differs not only by country but sometimes within one. Check where you are before relying on a recording.

The social position is simpler. Ask. "Do you mind if I record this so I can write it up?" is normal, takes three seconds, and removes the entire problem — including the part where a recording turns out to be unusable because you cannot share it.

Technically, three points matter:

• The other side's audio may not be captured. On a phone, screen recording during a call often captures your microphone and not the call audio, because the system routes call audio separately. Test before relying on it.
• Meeting apps have their own recording, which captures both sides properly and usually produces a transcript. Where it exists, it is better than a screen recording in every respect.
• A separate audio recording is often the right answer. If you need the words rather than the picture, record audio only: smaller files, better quality, and easier to transcribe afterwards.

The third point is the one people miss while fighting with screen recording. Most meeting recordings exist to answer "what was said", and that is an audio problem with an audio solution — the screen was never the content.

When it will not record

Four causes, and the first two account for most of it.

• Protected content. Streaming services block screen recording of video, and the result is a black rectangle with sound, or nothing at all. This is deliberate and there is no setting to change.
• The recorder is not enabled. On iPhone, Screen Recording has to be added to Control Centre in Settings before it appears at all.
• Storage is full. Recording stops without a clear message, and long recordings consume space quickly.
• A call interrupted it. Incoming calls end a recording on most phones, and what was captured up to that point is usually saved.

The protected-content case is worth understanding rather than working around: it is enforced below the app layer, and tools promising to defeat it are the category most likely to be doing something else as well.

More free utilities in free utilities, the editing side in video tools, and general picks in online tools. Apple documents the Control Centre recorder including the microphone toggle that causes most of the confusion.

The short version

Screen recording captures two independent audio sources — the device's own sound and the room — and the default is rarely the combination you wanted. On iPhone, press and hold the record button to reach the microphone toggle; on a Mac, the built-in recorder does not capture system audio at all.

Silence notifications first, check what is visible in the frame before recording rather than after sharing, do a ten-second test to confirm the audio, and trim in Photos afterwards because that is lossless. If your voice is in it, getting closer to the microphone in a soft room beats anything you can do in software later.