Audio to video converter, with your image and a moving waveform

MP3 to MP4 with an image, ready for YouTube

YouTube does not accept sound on its own. Its help page says audio files such as MP3, WAV or PCM cannot be uploaded to create a YouTube video, and it lists MP4 among the formats it does take (YouTube Help). So a song, a talk or a podcast episode needs a picture before it can go up, and that is the part this page adds.

For a regular YouTube video, choose Wide 16:9 at 1080p. Under Image placement, the Whole image, on a blurred copy setting keeps square cover art complete, with a soft blurred copy of it filling the sides, and the Fill the frame setting suits a wide photo, though it trims whatever does not fit. The result is H.264 video in an MP4 with AAC sound, the container and codecs named in YouTube's recommended upload settings (YouTube Help).

For a YouTube Short, choose Vertical 9:16 or Square 1:1 and keep a part of up to 3 minutes: YouTube takes Shorts up to 3 minutes long, in a square or vertical shape (YouTube Help).

Want only a black screen under the sound? Keep Color as the background, set both the Background and Waveform colors to black, and uncheck the progress bar. Every frame comes out black, with your title in its own color if you type one. In our test, 10 seconds at 720p came to 177 KB, almost all of it sound.

The right shape for Instagram Reels, Shorts, WhatsApp Status and feeds

Pick the shape for the place the video is going. Every shape is made at 30 frames a second, at 1080 or 720 pixels on its short side.

In vertical videos the title starts about an eighth of the way down and the waveform stops above the bottom fifth, the areas where apps tend to lay their own names, captions and buttons over a video.

Voice memos and voice notes, on an iPhone or Android phone

On an iPhone, open Voice Memos, tap the recording, tap the More button, then Share, then scroll down to Save to Files. Apple says recordings are exported as .m4a files by default (Apple). Open this page in Safari 26 or later, tap Choose an audio file and pick the recording from Files. When its sound is AAC, it goes into the video exactly as it was recorded.

A voice message in a chat leaves WhatsApp when you export the chat with its media. On Android, Export chat sits under More in the chat's menu, where Include media attaches the chat's most recent media; on an iPhone it sits under the contact or group name, where you pick With all media, then Save to Files (WhatsApp Help). The iPhone export is a ZIP file, and tapping it in Files makes a folder with the files inside (Apple). Opus, OGG and M4A recordings all open here. To keep a voice message as sound only, WhatsApp audio to MP3 saves it as an MP3 instead.

When the video is ready, Chrome on Android saves it to your download location, where the Files app lists it (Google), and Safari on an iPhone saves it to the Downloads folder in Files (Apple). From there, share it to the app you want like any other video.

Phones have less memory than computers, and on a phone the browser holds the finished video in memory until you save it. Clips of a minute or two at 720p are the easy case. If a video would pass about 200 MB, the tool says so before you start and suggests a shorter part or 720p.

Bars, line or circle: an audio visualizer that follows the sound

In podcasting, a video like this is called an audiogram (the same word names the chart from a hearing test, which this is not). Every frame here is measured from your own sound, so the waveform jumps on a drum hit, swells with a voice and settles in a pause. Nothing is looped or made up.

Each frame looks at about 46 milliseconds of sound around that moment and splits it into 64 bands of pitch, from deep bass at 50 Hz up to 16 kHz, where cymbals and the hiss of an s sit. Each bar, or each point of the line, is one band: the lowest pitches in the middle and the highest at the edges. The circle puts the lowest at the top.

Loudness is measured against the loudest moment of the last few seconds, so a quiet voice memo fills the frame as fully as loud music. Bars jump up quickly and settle gently, which keeps fast speech readable instead of jittery.

The preview uses the same drawing code as the video. While the sound plays, it measures what you hear; when you pause, it works out the frame the video will have at that moment, the same way the video does.

What happens to your sound

Inside an MP4 the safest sound format is AAC: MDN notes that AAC in an MP4 container is supported by all major browsers (MDN). So the sound in the video is always AAC, copied when it already is and converted when it is not:

Converted sound always gets an explicit bitrate, never an encoder's default, and the bitrate steps down only when the browser's AAC encoder will not take it. A part of an AAC sound keeps only the pieces (about 23 milliseconds each) that lie wholly inside your start and end, so nothing from before the start is hidden in the file.

Long podcasts, large files and how long it takes

This page sets no file size limit and no length limit; your device does. The video is encoded at a fixed bitrate for each size, so the file grows steadily with its length. These are the most a minute can take, next to times from our tests in Chrome on a Windows laptop (Edge was slightly faster with the podcast):

Our files came to between half and nearly all of these figures: the least for bars on a plain color, the most for the line style or a detailed photo. A one hour episode at 1920 x 1080 can reach about 2.3 GB, so for a full episode 1280 x 720 is the practical choice, at up to about 1.2 GB an hour.

A sound longer than 10 minutes opens with only its first minute chosen, so a long episode does not turn into a long video by accident: drag the handles to pick your clip, or close Use only part of the sound to use all of it. In Chrome and Edge on a computer, a video over 500 MB is written straight into a file you choose instead of being held in memory, which is how large files get through. A phone can take longer than a laptop, and the progress bar shows how much time is left.

Without uploading: what stays on your device

The page opens in your browser like any online tool, but your sound file and your image are read by this tab and never uploaded. The page and the video engine it loads when you open a file come from uncrop.si; after that, the work happens on your device. Under the download button, a line counts what this page has sent since you added your file, so you can check it yourself.

Sound files often carry more than sound. MDN notes that an MP3 is often accompanied by metadata such as ID3 tags (MDN), which can hold a title, an artist, album art, comments and dates. None of it is copied into the video: the MP4 holds the frames drawn on this page and the sound, and nothing else. Your image is used only as pixels, so the location and camera details saved in a phone photo stay behind too.

Limits, and which browsers can make the video

No captions, lyrics or subtitles of what is said: the video shows the waveform and the words you type, and nothing is transcribed. No AI is used anywhere in this tool. Each video takes one sound and one image. A title can be up to 80 characters, on up to three lines, and a subtitle up to 100 characters, on up to two.

The waveform is always drawn, so a still picture with no waveform over it is not an option; the black screen settings above come closest. The image itself stays still, except that a round image pulses gently with the sound.

Making the video needs WebCodecs, the browser feature for encoding video on the device (caniuse). Before anything starts, the tool checks that this browser can encode H.264 at the size you picked and, for sound that is not AAC, encode AAC. If it cannot, it says what is missing and what to try.