Audio Speed Changer

What this tool does and when it helps

A speed change is the simplest edit with the widest range of uses. Musicians slow a difficult passage to learn it note by note. Students and researchers replay interviews at a pace that suits transcription. Commuters get through an hour-long talk in forty minutes. Dancers and runners match a track to the tempo of a routine. All of these need the same two decisions: how much faster or slower, and whether the pitch should follow.

Here the speed runs from 0.25x, a quarter of normal, to 4x, four times normal. The start value is 0.8x, a gentle slowdown that makes fast phrases easier to hear while keeping them recognizable. The Keep the original pitch switch is on by default, so music stays in tune and voices sound like the same person speaking more slowly or quickly.

The tool edits files that you already have. It does not fetch audio from video or streaming sites, so use recordings you made, bought or have permission to change.

How to slow down audio and how to change speed in practice

Load the file, then choose a preset. 0.9x is a mild reduction for pieces you almost know. 0.75x is the usual practice speed, and 0.5x is half speed, where most notes become distinct even in fast runs. On the faster side, 1.25x, 1.5x and 2x suit spoken material, and 4x is mainly for scanning through a long recording to find the section you want.

Drag the Speed knob up or down for an in-between value, and double-click it to return to 0.8x. Because the preview responds within about a quarter of a second, you can turn it while the audio plays and hear the change without restarting. The Result and Original buttons switch at the same moment of the track, so you can check that a difficult phrase is audible at the new speed.

When the setting is right, save the file. The name keeps your original and adds the tool name, so a lecture called week3 becomes a separate file that you can load onto a phone or player. Use Copy settings link if you want to return to the same speed later.

Time-stretch versus resampling: the one idea worth understanding

There are two fundamentally different ways to make audio run at another speed. The first is resampling, sometimes called varispeed. The recording is simply played back at a different rate, exactly like a tape machine or a vinyl record at the wrong setting. Every sound wave becomes shorter or longer, so the frequency changes with the speed. Play at half speed and everything drops one octave; play at double speed and everything rises one octave. The voice of a speaker turns into a growl or a squeak.

The second is time-stretching, and it is what the pitch switch activates. The tool cuts the signal into short overlapping windows, then lays them back down closer together or further apart. Within each window the waveform is untouched, so the frequencies, and therefore the pitch, stay the same. The technique used is called waveform similarity overlap-add, or WSOLA. Before joining two neighboring windows, it looks around a tiny range for the position where their waveforms line up best, so the seam is not heard as a click. The code is our own, written for this site.

Time-stretching has limits. Because it repeats or skips tiny pieces, sustained tones can develop a slight flutter at extreme settings, and sharp drum hits may lose a little of their snap when slowed a lot. Resampling has no such side effects, but it moves the pitch. A good rule is to pick time-stretch for speech, instruments you want to learn and anything that must stay in its key. Pick resampling when the pitch change is the effect you are after, as with a tape-style slowdown.

Switch the option on and off while listening to the same bar. The difference is immediate, and it will tell you quickly which version suits your purpose.

Lengths at different speeds

Length is the length at normal speed divided by the speed factor. The table in the Numbers tab shows a three-minute song and a one-hour recording at each factor. At 1.5x, 60 minutes becomes 40, which answers a common question: 1.5x is 50 percent faster, so you finish in two thirds of the time. At 2x the hour becomes 30 minutes, and at 0.5x it grows to 2 hours.

The last column of that table treats a 100 beats-per-minute practice track. At 0.75x it plays at 75 beats per minute, at 1.25x at 125. That makes the tool a handy tempo changer for rehearsal: pick the BPM you want to start from, divide by the original, and use the result as the speed. If a piece is marked at 120 and you want to practice it at 90, set 0.75x.

For a longer-term routine, build a ladder. Start at 0.6x until the passage is clean, move to 0.7x, then 0.8x, 0.9x, and finally full speed. Save a file at each rung if you want to practice away from the computer.

Speed settings for different tasks

For learning a song by ear, 0.75x with the pitch kept is a comfortable first pass; shift to 0.5x for the hardest bars. For transcribing speech, 0.7x to 0.85x gives the ear time to catch each word without making the voice sound drunk. For podcasts and lectures, 1.25x to 1.5x is a good speed-listening range, and many listeners can follow spoken material up to about 2x with practice.

For checking a mix, a slower speed can reveal timing slips between instruments. For dance classes or workouts, adjust the track so the beat matches the pace you need. If you know the target BPM from the bpm finder or the tap tempo page, divide the target by the original to get the factor.

For effects, switch the pitch lock off. A tape-style slowdown at 0.5x produces a dark, thick version of the track, and double speed gives a high-pitched squeal. For a ready-made slow-and-wet treatment or a fast-and-bright one, other pages in this site have dedicated controls.

Getting the best quality

Start with the highest-quality copy of the recording, since slowing audio makes weak spots last longer. An MP3 at a low bitrate may reveal its compression artifacts at 0.5x, with a swirly or watery quality on cymbals. WAV, FLAC or a high-bitrate MP3 holds up much better.

Keep moderate settings when you use the pitch lock. Between 0.7x and 1.5x, results are usually clean on both music and speech. Outside that range, listen carefully and, if you hear fluttering, try a slightly less extreme speed or accept the tape-style option for that section.

Saving as WAV avoids another round of lossy compression, so choose it when you will edit the file further. MP3 at 320 kbps is a good choice for listening on a phone. The estimated size is displayed before you commit.

What to expect at the extremes

At 0.25x a four-second phrase stretches to sixteen seconds. With the pitch lock on, individual notes become very long, and plucked or struck sounds can start to feel like a series of separate pieces rather than one smooth decay. This is normal for any time-stretch at that depth. It is still useful for hearing the exact order of fast notes, but it is not a faithful picture of how the music sounds.

At 4x the opposite happens. Speech becomes a stream of fragments and drum parts blur into a buzz. Treat this range as a way to skim a long file for a section rather than to listen to it. If the goal is scanning, 2x with the pitch lock on is often the better compromise, because words stay recognizable.

In the middle of the range, between about 0.7x and 1.5x, results are the cleanest. Past that stretch on music, expect some flutter: compare against Original often, and consider switching off Keep the original pitch for the most extreme settings, where the tape-style sound has no flutter at all.

Formats, limits and privacy

Common formats such as FLAC, Opus, Ogg, WAV, M4A and MP3 are accepted, along with other audio the browser can decode, with a size ceiling of 250 MB. At slow factors the output grows: a ten-minute track at 0.25x runs forty minutes and the saved file expands to match, so watch the size estimate. Long files need more time on slower phones. A lecture recorded as video can be slowed down as well, because most MP4, MOV and WebM files open and their sound is processed; the saved file is audio only.

Every step happens on your own device and no copy of the recording is sent anywhere. Your browser keeps the last file you opened on this device, so other tools can offer to continue with it, and the links under the player pass the processed audio onward through local storage. Clearing the site data removes both.