Audio Waveform Video Maker

How to make a waveform video

1. Choose the audio. One MP3, WAV, M4A (AAC) or OGG file, up to 20 minutes long. Your browser decodes it; a file it cannot decode is named in a plain line, and nothing else changes.
2. Pick the clip. The overview shows the whole file; the first 30 seconds (or all of a shorter file) are picked. Drag either handle, or type a start and an end. Play clip plays exactly that part with the moving preview.
3. Choose the look. Bars, Wave or Circle, a colour for the bars and the words and one for the background. Add cover art or a photo: behind the bars (darkened so they stay readable) or inside the circle.
4. Add words, if you like. A title and a second line (the artist, the episode, a date), up to two lines each, drawn into every frame in this site's Geist typeface. Tick the progress bar for a line that fills as the clip plays.
5. Pick a size and save. Square, Vertical or Wide, full size or Faster 720p. Save MP4 shows the progress with a Cancel button and downloads waveform-1080x1080.mp4. Save this frame as PNG gives a still at the same size, for a cover or a thumbnail.

What the bars show

Every frame is worked out from the sound itself, not recorded from the screen, so the same file and settings always give the same video. For the frame at time t, the page takes 4,096 samples (93 ms at 44.1 kHz) centred on t, shapes them with a Hann window and runs a Fourier transform. Bars splits 40 Hz to 16 kHz into 48 bands spaced evenly in pitch, so each band is about 2.2 semitones wide, and each bar's height is the band's level from −70 dB (no bar) to 0 dB, the level of a full-scale sine wave. A 440 Hz A, for example, lands in bar 20 of 48 (428.6 to 485.6 Hz); a 2 kHz whistle in bar 32 (1,917 to 2,172 Hz). Silence draws no bars at all. Circle uses the same measure with 72 bands, clockwise from the top. Wave draws 33 ms of the sound wave across the frame, so a low hum shows as a slow curve and a voice as a busy, jagged line.

Sizes and file size

The sizes are upper estimates: the video bitrate plus 128 kbit/s of sound, times the length. A plain background with bars usually comes out smaller, a busy photo behind the bars closer to the estimate.

The sound in the MP4

The sound is the clip itself, at 44.1 kHz, written as AAC at 128 kbit/s, the sound format almost every phone, TV and computer plays in an MP4. Where your browser's own AudioEncoder can write AAC, the page uses it. Where it cannot (Chrome on Linux, on our test machine, for one), the page uses its own small AAC encoder: it codes everything up to 15.9 kHz and drops the very top of the treble, and it spreads its rounding noise evenly instead of hiding it where the ear is least sensitive, so it is a little less clean than a studio encoder on bright, dense music, and fine for speech. A stereo file stays stereo and a mono file stays mono.

What it cannot do

• No captions or lyrics from speech. There is no speech recognition here; the only words in the video are the title and second line you type.
• One audio file per video. No playlist, no second track, no mixing two clips.
• It does not post anything. It saves a file to your device; you upload it to an app yourself.
• The MP4 needs WebCodecs. Making the MP4 needs a browser with the WebCodecs VideoEncoder. Where it is missing, the page records a WebM in real time instead, and says so before it starts.
• Files up to 20 minutes, clips up to 10 minutes. On our test machine (headless Chrome on a 4-core server) reading a 20-minute stereo MP3 took 3.9 s and about 640 MB of memory, because the sound is held unpacked: 44,100 samples × 2 channels × 4 bytes for every second, plus a mono copy for the bars. A phone with little memory may stop on a long file, so cut it shorter first. Making the video ran at 5.7× real time for Square 1080 (a 10-minute clip in about 1 min 45 s), 9.8× for Square 720 and 3.7× for Vertical 1080 × 1920 (about 2 min 40 s for 10 minutes); a phone can be slower.
• The picture is a still. JPG, PNG or WebP, drawn once behind the bars or inside the circle; it does not move or zoom.
• Nothing is saved for later. Close the tab and the file, the clip and the settings are gone.

Questions

How do I turn audio into a video with a waveform?

Open this page and tap Choose an audio file (MP3, WAV, M4A or OGG). Drag the two handles on the overview, or type the start and end, to pick the part you want; the first 30 seconds are picked for you. Choose Bars, Wave or Circle, your colours, a picture and a title if you like, and a size, then tap Save MP4. Your browser draws every frame from the sound and writes an MP4 with H.264 video and AAC sound; nothing is uploaded and no watermark is added.

Can I make a music visualizer video for Instagram or TikTok?

Yes, as a file you post yourself. Pick Vertical (1080 × 1920, 9:16) for an upright phone video or Square (1080 × 1080) for a feed post, then Save MP4: the file holds H.264 video and AAC sound at 30 frames a second. This page does not post anything and does not check any app's rules on length or size; you upload the saved file in the app as you would any video. Use only music you have the right to share.

How long can the video be?

Up to 10 minutes, cut from a file of up to 20 minutes. Measured on our test machine (headless Chrome on a 4-core server, a 30-second clip): Square 1080 × 1080 was made at 5.7 times real time, Square 720 × 720 (Faster) at 9.8 times, Vertical 1080 × 1920 at 3.7 times and Wide 1920 × 1080 at 3.9 times, so a 10-minute vertical video takes about 2 min 40 s there. A phone can be slower; the progress bar shows the speed on your own device. The WebM fallback always takes as long as the clip.