video tool
Audio waveform viewer
Open an MP3, WAV, OGG or M4A file and this audio waveform viewer decodes every sample in the page and draws the waveform, with the peak and the average level in dBFS, a count of samples at full scale and every silent stretch shaded and listed. Click the waveform or a silence to jump there and listen with the player underneath.
Loading the tool…
- Runs on your device, nothing is uploaded
- Free, no sign in
- No watermark, no file limit
How it works
- 1
Open the audio
Choose a file or drop it on the panel. The length is checked first, then every sample is decoded on your device.
- 2
Read the waveform
Taller means louder. Amber shading marks silent stretches and red columns mark samples at full scale. Click anywhere to move the playhead.
- 3
Check the numbers
Cards show the duration, channels, peak, average level and how many samples hit full scale. Notes below say what the figures mean for this file.
- 4
Tune the silence finder
Change the threshold and minimum length to suit the recording. Each silent stretch is a button that jumps the player to it.
Reading the levels
Digital audio is measured against full scale, the largest value a sample can hold, written 0 dBFS. Everything else is negative: a peak of -1 dBFS is close to the ceiling, -6 dBFS is half the full amplitude. The peak card shows the single highest sample across every channel. The average card is the RMS level, the square root of the mean of all squared samples, which tracks how loud the file feels better than the peak does. Spoken word recordings commonly average between -24 and -16 dBFS; music mastered for release sits much higher. Samples at exactly full scale are counted, because long runs of them mean the waveform was flattened against the ceiling, which is clipping.
How silences are found
The file is cut into 50 millisecond windows and the RMS level of each is measured. A run of windows below the threshold, lasting at least the minimum length, is a silent stretch. At -60 dBFS only near digital silence counts, such as the gaps between tracks. At -50 the hiss and room tone of a typical voice recording counts as silence, which is the setting for finding pauses in speech. At -40 quiet passages count too. The list shows start, end and length, and each is a button that moves the player there so you can hear whether it is a real gap or a quiet line. The total number of silent stretches and their combined length appear in the last card, which is a quick way to judge how much a recording could be tightened.
Music and voice for a video
Before a track goes under a video, the waveform shows whether it starts on silence that would delay the music, ends abruptly, or clips. In the OneCraft video builder you can upload your own music track, up to 25 MB, or pick one of 13 in the library. Music plays at 25% volume by default and dips to 22% under narration, so a track that is already quiet, with an average well below -30 dBFS, may sink out of hearing under a voice. The note under the cards says whether the file fits the upload cap.
Limits worth knowing
Decoding uses the browser's own audio engine, so the formats are whatever it supports: MP3 and WAV everywhere, OGG Vorbis and Opus in most browsers, M4A with AAC in most, and Apple Lossless or protected M4A files in few. When decoding fails you get an error, never a partial waveform. Every sample is held in memory while you look, so recordings over 60 minutes are refused. The sample peak can sit slightly under the true peak an encoder later produces, so leave a little headroom. Nothing is uploaded and nothing is stored, ever.
Limits
- Files the browser cannot decode, which includes some M4A files, show an error rather than a partial result.
- Recordings longer than 60 minutes are refused, because every sample is held in memory, about 21 MB a minute in stereo.
- The peak is a sample peak and the average is RMS. Neither is a loudness reading in LUFS, and the true peak between samples can be slightly higher.
What it does not do
- It does not edit, trim, normalise or convert the audio.
- It does not measure loudness in LUFS for streaming targets.
- It does not show a spectrum or frequencies.
Questions people ask
What is a good peak level for a voice recording?
Aim for peaks between about -6 and -3 dBFS when recording, which leaves headroom for loud moments and for the small overshoot an encoder adds. A finished file can sit closer to the ceiling, around -1 dBFS. If the peak card shows 0.0 and red columns appear, the recording clipped and the loudest syllables are distorted.
Why does my file show samples at full scale but sound fine?
A few isolated samples touching full scale are inaudible; mastering often leaves music right at the ceiling. Clipping you can hear comes from runs of them, where the top of each wave is flattened, and shows as solid red stretches on the waveform. If the red is scattered single columns, the file is loud but probably clean.
Is dBFS the same as the loudness platforms ask for?
No. Streaming and broadcast targets use LUFS, a loudness measure that weights frequencies the way hearing does and gates out silence. A file's RMS average in dBFS moves in the same direction as its LUFS reading but is not the same number. Use a loudness meter for a LUFS target; use this page to see the shape and the peaks.
Can I see each channel of a stereo file on its own?
The waveform combines the channels, drawing the highest and lowest sample of either at each point, which is what you need to spot clipping and silence. The peak and average cards also cover both channels together. If one side is silent or much quieter, the channels card still reads stereo, so listen on headphones to check the balance.
Can I save the waveform as an image?
Yes. Save the waveform as PNG writes a 1600 by 320 pixel image of the full file with the silent stretches shaded and any clipping in red, without the playhead. It suits a podcast episode page, a social post about a track, or a note to an editor showing where the problem sits.
Put your music under a narrated video
Take what you made here into videos.
Create a video with OneCraftMore tools like this
- Video metadata viewerA video metadata viewer for MP4, MOV and WebM: duration, resolution, aspect ratio, frame rate, bitrate, codecs and audio, read in your browser. No upload.
- Screen recorder onlineA free screen recorder online: record a screen, window or tab with or without your microphone, review it, and save a WebM file. Nothing to install, no upload.
- Speaking time calculatorPaste a script and see it laid out on a clock, section by section, against the slot you have. Six measured speaking rates, and a rehearsal timer.
- SRT to textTurn SRT or VTT subtitles into plain text: timestamps and numbers removed, lines joined into paragraphs, speaker names kept if you want. Download as .txt.
Background reading: the question page on this.
Written and checked by the OneCraft team. Last checked .