Start a project →

Free Room Echo Checker: how echoey does your voice sound?

Choose a recording of your voice and get an estimate of how much your room echoes: a decay time in seconds, a simple verdict from dry to very echoey, the lows, mids and highs apart, and every moment it used, so you can click and hear the tail yourself. Then fix the room with the tips below. Everything runs in your browser: your file never leaves your device.

Updated · Built by Kevin Tabares, long-form YouTube editor at Umbrella Creators.

Check the room echo in a voice recording

Your file stays on your device. Nothing is uploaded: your browser decodes the audio and measures it right here, in this tab.

How to use it

  1. Record a minute or more of you talking normally in the room you want to check, with no music. The raw file from the camera, the phone or the recorder is best (MP4, MOV, WEBM, MP3, WAV or M4A, up to 60 minutes).
  2. Press Check the echo. The browser decodes the audio and follows how the sound dies away every time you stop talking; a long file takes a few seconds.
  3. Read the decay time and the verdict, look at the lows, mids and highs, and click any moment to hear the tail after your words. Change one thing in the room, record again and compare.

The estimate needs pauses: moments where you stop and the room is the only thing left. Normal speech has plenty of them. Music under the voice, a noise gate, noise reduction or an AI voice fill or cut those pauses, so check the raw voice when you can.

Step-by-step guide

For the whole process around this tool, read How to reduce echo in voice recordings: how to stop it while recording and what really works to clean it in the edit.

FAQ

What does the echo number mean?
It is a decay time in seconds: how long the sound of your room would take to fall by 60 dB after you stop talking. That is how the International Electrotechnical Commission defines reverberation time (IEV 801-31-07, read on 20 October 2026), the number often written RT60; here it is estimated from your own speech instead of a test sound. It is read on the mids (500 to 2000 Hz), where most of the voice is; the lows and highs are shown apart. A drier room gives a smaller number. A single distinct echo, your words coming back a moment later from a hard wall further away, is a different thing: the tool looks for it separately (50 to 300 ms after the words, down to about 14 dB quieter than your voice) and tells you when it hears one.
How does it measure the echo from my voice?
It looks for the moments where you stop talking and the sound dies away: a short peak followed by a fall of at least 25 dB with no new words on top. On each of those tails it fits a straight line to the fall from 5 to 25 dB below the peak and extends that slope to 60 dB. Tails that are cut by new speech, sink into the background noise or are not straight are thrown away, and the result is the middle value of all the tails left. Speech itself takes a moment to stop, so the dry narrations I measured (7 voices) already read about 0.06 to 0.11 s; the tool subtracts 0.07 s for that.
How accurate is it?
I tested it on three dry voices (a documentary voiceover and two synthetic voices, mono and stereo, 44.1 and 48 kHz) mixed with decays I built with known lengths, some with steady background noise 20 to 35 dB under the voice: 55 tests in all. Rooms of 0.6 s and longer read within 13% of the real value and a 0.4 s room within 16%, but a 0.2 s room read 2 to 33% too low, because speech itself takes a moment to stop and that moment is different for every voice. The verdict was right in every test except right at the 0.8 s line, where readings fell on both sides of it. It is still an estimate from speech, not a lab measurement with a sweep or a clap. A mic close to your mouth makes the same room read drier (in my test, a 0.6 s room with the voice 6 dB louder than the room read 0.50 s), because your voice is much louder than the room, and that is also how it sounds. Music under the voice fills the pauses and makes the reading unreliable, so check a clip of just your voice.
Why does my cleaned or gated audio read so dry?
A noise gate, noise reduction or an echo removal tool cuts the end of each tail, so the sound dies away faster than the room really makes it. When many tails drop straight to silence, the tool says so: if the reading is dry or has some room, it tells you it can't judge the room from that file at all, and if it is echoey, that the room may be even more echoey. AI voices and edited gaps of pure silence do the same. To judge the room itself, check the raw recording from the camera or the recorder.
What is a good number for a YouTube voice?
My rule of thumb, not an official standard: under 0.3 s sounds dry and close, 0.3 to 0.5 s has some room, like a furnished room, 0.5 to 0.8 s is echoey and listeners notice it, and over 0.8 s sounds like an empty room. A dry recording is the easiest to edit: you can add room sound later, but removing echo afterwards is harder and can leave a watery sound.
Does it upload my file?
No. Your browser decodes the audio and analyses it on your own device; the file never leaves this tab, and nothing is downloaded either.

Who made this

I'm Kevin Tabares. I've edited long-form YouTube videos since 2015 (1,277 videos and 152.2M+ views tracked on YT Jobs) for channels like Rex & Alexa, Puff and Mud. Room echo is one of the first things I listen for in raw talking-head files: on one channel I edit, the raw camera audio reads about 0.52 s (echoey) here and the cleaned version about 0.19 s (dry). I built this so anyone can check a room in seconds, without uploading anything. If you want an editor for your channel, get in touch.

More free tools

A background noise checker for hiss and hum, a mic test you can run before you record and a LUFS meter for loudness and true peak, all free and without signup.

See all tools →