How to record high quality audio in your browser

You can record audio in your browser right now. Open a tab, click a button, and you have a recording. The problem is that most browser recorders give you bad audio. Mono signal, low sample rate, compressed output. It sounds like a phone call from 2005.

The browser is capable of recording CD-quality audio. 44.1 kHz stereo, uncompressed WAV. The Web Audio API supports it. Most recorders just choose not to use it because high-quality audio means bigger files, and bigger files cost more to store on their servers. Orec records locally, so file size is your device's problem, not ours. That changes what we can offer.

Sample rate is the thing most recorders get wrong

Sample rate determines the highest frequency your recording can capture. The math is simple: a recording can reproduce frequencies up to half its sample rate. This is called the Nyquist frequency.

At 22 kHz (what most browser recorders use), the highest frequency you can capture is 11 kHz. Human hearing goes up to about 20 kHz. So a 22 kHz recording is throwing away almost half the audible spectrum.

That missing top half is where consonants get their crispness. The difference between "s" and "f" lives up there. Cymbal shimmer, acoustic guitar sparkle, the breathiness in a vocal take. All above 11 kHz. All gone at 22 kHz sample rate.

Orec records at 44.1 kHz or 48 kHz, depending on what your microphone and browser support. 44.1 kHz is CD quality. 48 kHz is the standard for video production. Both capture the full range of human hearing.

Stereo vs. mono

Most browser recorders default to mono. One channel. Every sound source collapsed into a single point. For a quick voice memo, mono is fine. But for anything you want to sound good (a podcast intro, a music demo, an interview with two people), stereo makes a real difference.

Stereo captures spatial information. The slight difference in timing and volume between left and right channels tells your brain where sounds are in the room. A guitar panned slightly left, a voice in the center, room reverb spreading wide. Stereo recordings sound like you are in the room. Mono recordings sound like you are on the phone.

Orec records in stereo by default when your microphone supports it. If you have a USB condenser mic or a stereo field recorder, you get both channels.

The echo cancellation problem

Your browser has built-in audio processing designed for video calls. Echo cancellation, noise suppression, and auto gain control. These features are great for a Zoom meeting. They are terrible for recording.

Echo cancellation removes parts of the signal it thinks are reflections. On a call, this prevents feedback loops. On a recording, it can cut out parts of your audio that sound similar to other parts. Noise suppression applies an aggressive filter that makes quiet sounds disappear and adds a watery, digital quality to everything else. Auto gain control pumps the volume up during silence and yanks it down when you get loud, killing your natural dynamics.

Orec disables all three by default. Your recording captures exactly what the microphone hears, with no algorithmic processing. If you want to clean up noise or adjust levels, you can do that afterward with a tool designed for it. But you cannot undo processing that already happened during recording.

WAV export gives you the raw audio

Most browser recorders export to MP3 or OGG. These are compressed formats. The encoder analyzes your audio, decides which parts are "less important," and throws them away to make the file smaller. This works well enough for sharing, but every round of compression loses detail.

If you record in a compressed format and then edit the file in a DAW, you are editing a lossy copy. If you export again after editing, you lose even more. This is called generation loss, and it adds up fast.

Orec exports to WAV. Uncompressed, lossless, exactly what the microphone captured. A one-minute stereo WAV at 48 kHz is about 11 MB. That is a big file compared to a 1 MB MP3, but storage is cheap and your recording quality is worth it. You can always convert to MP3 later for sharing. You cannot convert an MP3 back to WAV and recover what was lost.

Practical tips for the best recording

Your recording is only as good as the signal going into your microphone. The browser can capture everything faithfully, but it cannot fix a bad source.

Use a decent microphone. The built-in mic on your laptop picks up fan noise, keyboard clicks, and room echo. A USB condenser microphone (even a $50 one) is a massive upgrade. Test your mic before recording to make sure it is working and picking up your voice cleanly.

Get close to the microphone. Six inches is a good starting distance for speech. The closer you are, the louder your voice is relative to room noise. This is the single biggest improvement most people can make.

Reduce room noise. Close the window. Turn off the fan. Record in a room with soft surfaces (carpet, curtains, a couch) rather than a room with hard walls and floors. Sound bounces off hard surfaces and creates reverb that muddies your recording.

Wear headphones. If you are monitoring your recording through speakers, the microphone picks up the speaker output and records it. Headphones eliminate this. They also let you hear exactly what the microphone is picking up, so you can catch problems before they ruin a take.

Check your levels. Record a test clip and listen back. If the waveform is tiny, you are too far from the mic or the gain is too low. If the waveform is clipping (hitting the top and bottom of the track), you are too loud. You want peaks around 70-80% of maximum. You can use our LUFS normalizer to check your levels after recording.

Start with good audio, fix less later

Every professional audio engineer says the same thing: get it right at the source. No amount of post-processing can fix a recording that was captured at 22 kHz mono with echo cancellation eating half the signal. But a clean 48 kHz stereo WAV gives you everything you need to edit, mix, and publish something that sounds professional.

Orec records at 44.1 or 48 kHz stereo, exports to lossless WAV, and disables all call-oriented processing by default. Open the page, check your mic, and hit record. The browser does the rest.