How to record your screen with audio
Choose the sound source as carefully as the picture. A microphone and a shared tab record different things.
Free toolScreen RecorderRecord a screen, window or tab with a live sound check, optional camera bubble and trimming. Save the video without uploading it.Record your screenRecord one short test before the real take
Choose a playing tab, enable shared audio and add your microphone if you need narration. Record a short sentence, then play it back.
- Share the smallest useful source.
A single tab avoids unrelated windows. Tick Share tab audio in the browser picker when that option appears.
- Check the sound you actually received.
A moving microphone meter proves a signal reaches the recorder. A connected track alone does not prove anyone is audible.
- Listen to a saved test.
Play both the beginning and the end. Check voices, source audio and timing before recording your full explanation.
Microphone, tab sound and system audio are different
Microphone captures your voice. Tab sound captures audio from a shared browser tab. System audio covers computer sound when supported.
| What you need | What to choose | What to check |
|---|---|---|
| A silent visual demonstration | Screen or window, audio off | No private notifications appear. |
| Your explanation over a website | Tab or window plus microphone | Your voice meter moves and playback is clear. |
| A video’s own soundtrack | Playing tab plus Share tab audio | The shared-audio meter moves. A mic is unnecessary unless narrating. |
| A meeting plus your commentary | Tab audio and microphone | Use headphones. Check the meeting’s recording rules and tell the people involved. |
| Sound from a desktop app | System audio if offered, or an OS/desktop recorder | Do a test. A shared window can supply no audio track. |
Choose a recording method for your device
A browser suits a short walkthrough you want to keep as a local file. Built-in recorders and OBS suit different jobs.
A browser on a computer
Use the screen recorder to check sound, add a camera bubble and save a short clip. Source-audio support varies by browser and operating system.
Windows 11
Open Snipping Tool, choose Record and select your area. Use its available audio controls, then check a test. Microsoft’s Snipping Tool instructions cover current controls.
Mac
Press Shift, Command and 5. Choose full-screen or selected-area recording. Options can select a microphone. For other audio needs, see Apple’s screen recording instructions and test your chosen method.
iPhone and iPad
Use Screen Recording in Control Centre. Hold its control to choose microphone behaviour when available. Some apps block recording. Apple explains the built-in recorder.
Android
Swipe down to Quick Settings and choose Screen record. Select the sound options your device offers. Labels vary by manufacturer. Follow Google’s screen recording instructions.
Long sessions or several scenes
Use OBS Studio for sources, scenes and disk recording. Setup takes more work, but suits long tutorials and production control.
Fix a screen recording with no sound
Start with the missing input. Turning up your speakers does not add a shared-audio track that the browser never supplied.
| What you see or hear | What to try |
|---|---|
| No shared audio was provided | Choose a playing browser tab again and tick Share tab audio. If your picker has no sound option, use another browser or recording method. |
| Microphone permission refused | Allow the microphone in site settings and check your operating system’s privacy controls. Reconnect it if missing. |
| Input connected, meter still | Check the device mute switch, app volume and whether the source is playing. Record a test and listen. |
| Voice doubled or echoing | Wear headphones so speaker audio does not enter the microphone. Avoid monitoring the recorder’s mixed sound through speakers. |
| Sound clips or distorts | Reduce the microphone or shared-input level. Two loud inputs can overload a mix. |
| Sound stops halfway through | Check device disconnection or ended sharing. The floi recorder stops when an active audio source ends so you can review the partial take. |
Read a sound check without overclaiming it
A meter measures the signal now. It cannot prove the right speaker will be audible later. Treat it as one part of a check.
The floi recorder follows each input from capture to export. These states mean different things:
- Requested: you asked for shared sound, but the source has not been chosen.
- Connected: the browser returned a live audio track. Play sound to test its level.
- Muted: a track exists but contributes zero signal while muted.
- Captured: the finished recording contains the input mix. Listen to check quiet sections and muted moments.
Two sources are mixed at half gain each by default. If both reach a peak amplitude of 1 together, their sum stays at 1: 0.5 × 1 + 0.5 × 1 = 1. This reserves headroom; it does not match perceived loudness.
Resolution, file size and pauses
Choose enough detail for the words you show. A 4K source does not help if your audience cannot read tiny text.
| Choice | Practical effect |
|---|---|
| Source resolution, 4K or 8K | Source resolution keeps the available detail. A lower height limit can reduce encoder work. Increase the app’s text size for readability. |
| Frame rate | 30 suits slides; 60 or higher can help motion. Floi offers targets up to 240 fps and Source maximum. Review shows the actual average from encoded frames. A higher target cannot create motion your source never supplied. |
| Lower target bitrate | Aims for a smaller file. Busy footage may show more compression. Static screens can be much smaller than the target. |
| Pause | Removes your waiting time from the recorded timeline. Marked moments follow recorded time, not wall-clock time. |
| Camera and frame composition | Composes incoming frames separately from the page where supported. This keeps effects independent of the preview’s paint loop. A compatibility fallback asks you to keep the recorder visible. |
For a rough storage budget, 6 Mb/s for 60 seconds is 6,000,000 × 60 ÷ 8 = 45,000,000 bytes before audio and container overhead. This is arithmetic from the target bitrate, not a measured file size.
Keep an original before trimming or converting
Save the complete original first when the recording matters. Trimming and MP4 conversion can require another encode.
In floi, Save original bypasses the trim selection and format choice. MP4 export refuses to continue if the browser would discard a media track.
The recorder keeps encoded chunks in memory. It stops at 30 minutes or 512 MiB; edited exports support selections up to 10 minutes. For long sessions, write to disk with a desktop recorder.
A short recording, checked in October 2026
A controlled Chrome 156 test used an animated 960 × 540 source, a 440 Hz shared tone and a 660 Hz microphone tone. It included a camera bubble, a frame and a 0.9-second pause.
On an M1 Max MacBook Pro with 32 GiB RAM, the recorded timeline was 2.5819 seconds. Decoding confirmed both tones in the original and edited output.
| Result | Picture and sound | Observed duration | File bytes |
|---|---|---|---|
| Original WebM | 960 × 540 VP9, Opus audio | 2.5819 seconds | 132,808 |
| MP4, selected 0.4 to 1.6 seconds | 960 × 540 H.264, AAC audio | 1.2000 seconds of video; 1.2800 seconds in the container | 58,919 |
The MP4 container lasted longer because of its encoded audio boundary. A selected duration and a player’s displayed file duration can differ slightly.
This was a synthetic functional check on 28 October, before the frame-rate update. It is not a cross-device speed or quality benchmark. Actual file sizes vary with content and encoder settings.
What the browser can promise
It can ask you to choose a screen source. It cannot promise every operating system will provide every sound source.
The Screen Capture API documentation explains secure contexts, permission prompts and optional audio. The MediaRecorder documentation describes recording the stream your browser returns.
This guide’s signal example and storage calculation are implementation analysis dated 28 October 2026. They are not a cross-browser benchmark. Device permissions and codecs should be checked on your own browser with a short test.