How to stretch a sound into a drone in 3 steps
- 01
Load a sound
Click Open file or drag audio into the drop zone, or press Try a demo to start from a built-in loop. Short, tonal source material — a chord, a pad, a single note — tends to work best.
- 02
Set the stretch factor and window size
Choose a stretch factor from 2x to 100x to decide how far the sound is drawn out, and adjust window size from 0.1 to 1 second to control how smooth or grainy the result sounds.
- 03
Render and download
Rendering runs in a background worker with a progress bar, since a heavy stretch takes real processing time; cancel at any point if you want to change a setting. Output is capped at 10 minutes, so download it once it finishes.
What you can do with the Evaporator
- Spectral time-stretching from 2x to 100x, built on the same general approach as PaulStretch-style extreme stretching, so a few seconds of audio can become minutes of slow-moving texture.
- Window size from 0.1 to 1 second, which trades off smoothness against graininess — a larger window sounds smoother, a smaller window keeps more transient detail.
- Output length up to 10 minutes per render, enough to build a genuinely long ambient piece from a short source clip.
- Rendering runs in a background worker with a visible progress bar, so the browser tab stays responsive while a heavy stretch processes.
- Cancel a render at any point if you want to change the stretch factor or window size instead of waiting for it to finish.
- Works on any source sound, not just music — a field recording, a spoken phrase, or a single instrument note all stretch into distinct textures.
- Render preview, an Original/Processed A/B comparison, and Download audio once the stretched result sounds right.
What is extreme audio time stretch?
Extreme audio time stretch takes a sound and draws it out far beyond a normal tempo change, using a spectral technique in the general style of PaulStretch rather than the pitch-preserving stretch used for slowing a song down to practice it. Instead of just spacing the original waveform out, it rebuilds the sound from its frequency content over time, which is what allows a stretch of 50x or 100x without the sound falling apart into audible clicks or repeats.
The result is closer to a new piece of ambient audio than a slowed-down version of the original — pitch stays put, but rhythm, words, and individual notes dissolve into a continuous, evolving texture. It works on any sound you load, though tonal, sustained source material tends to produce a smoother, more recognizably musical drone than something percussive or full of silence.
Stretch factor and window size explained
Stretch factor sets how much longer the output is than the source — a 10-second clip at 20x becomes about 200 seconds, and at 100x becomes well over five minutes, up to the tool's 10-minute output cap. Lower factors, in the 2x to 10x range, still sound clearly stretched but keep more of the original's identity; higher factors move further toward abstract, ambient texture where the source becomes hard to recognize.
Window size controls the size of the audio chunks the spectral process works with, and it shapes the character of the result more than most people expect. A smaller window, closer to 0.1 seconds, keeps more of the original detail and can sound slightly grainier; a larger window, closer to 1 second, smooths everything into a more continuous, blurred wash. There is no single correct setting — it is worth trying both ends on the same source.
Building drones and ambient beds
A single sustained chord, a pad sound, a bowed instrument note, or even a short vocal phrase are all strong starting points for a drone, since the stretch has continuous tonal content to work with rather than sharp transients or silence. Stretching a short clip at a high factor is often a faster way to build an evolving ambient bed than programming one from scratch, since the source material's natural harmonic movement carries into the result.
Because rendering happens in a background worker, you can set up a stretch, let it process, and check back once the progress bar completes rather than waiting with the tab frozen. Once you have a drone you like, use Continue in another tool to add space with the reverb tool, or push the texture even further through the granular synthesizer.
What to expect: an experimental Lab tool
The Evaporator lives in the Lab because extreme audio time stretch is a deliberately extreme, exploratory process rather than a general-purpose audio effect — it exists to dissolve a sound into something new, not to preserve it. Very high stretch factors on complex, busy source material can produce muddy or indistinct results rather than a clean drone, and there is no way to selectively stretch only part of a mix; the whole file stretches together.
A heavy stretch, especially a long output length at a high factor, takes real processing time and taxes your device's memory, so an older or lower-powered device may be slower to render or struggle with a full 10-minute output. If a render feels stuck, cancel it and try a shorter output length or a lower stretch factor before assuming something is wrong.
Tips for better results
- Start with a short, tonal source clip — a chord or a sustained note stretches into a cleaner drone than a busy full mix.
- Try a lower stretch factor first (2x–10x) to hear what the tool does to your source before committing to an extreme 50x or 100x render.
- Compare a small and a large window size on the same source; the difference is often bigger than the stretch factor itself.
- Cancel a render early if the progress bar stalls or a quick preview sounds wrong, rather than waiting out a long process.
- Layer a stretched drone with reverb or granular processing in another tool for an even more spacious, evolving texture.
- Watch render time on long, high-factor stretches on an older device — shorter test renders help you dial in settings faster.
Frequently asked questions
Is the Evaporator free to use?
Yes. Stretching, rendering, and downloading are all free, with no sign-up and no watermark on the result. It runs entirely in your browser, using your own device to process the audio from start to finish.
What does "extreme audio time stretch" actually mean?
It means stretching a sound far beyond a normal tempo change — often tens or even a hundred times its original length — using a spectral technique that rebuilds the sound from its frequency content rather than simply spacing out the waveform, which is what keeps a 100x stretch from falling apart into clicks.
How long can the output be?
Up to 10 minutes per render. A short source clip at a high stretch factor can easily reach that cap, so a longer source file may need a lower factor to fit comfortably within the limit.
Does this upload my audio anywhere?
No. The source file you load and the stretched result are processed entirely on your device in a background worker. Nothing is sent to a server at any point in the process, and no account is needed.
Why does rendering take a while?
A heavy spectral stretch, especially at a high factor or a long output length, is genuinely processing-intensive, so it runs in a background worker with a progress bar rather than instantly. You can cancel at any point if you want to change a setting instead of waiting it out.
What is window size, and how should I set it?
It controls the size of the audio chunks the stretch works with. A smaller window keeps more detail and can sound grainier; a larger window smooths the result into a more continuous wash. There is no single correct value — try both ends on your source and pick what sounds right.
What kind of source audio works best?
Tonal, sustained material — a chord, a pad, a bowed note, a held vocal — tends to stretch into a cleaner, more musical drone. Percussive or silence-heavy material can stretch into something thinner or less continuous, though that is not necessarily a bad result for sound design.
Does the pitch change when I stretch audio this much?
No. The stretch changes duration and texture, not pitch — a note stretched 50x still centers on roughly the same pitch, just spread across a much longer, more diffuse span of time than the original recording.
Is this the same as the tempo changer or slowed + reverb tools?
No. Tempo changer and slowed + reverb apply moderate speed changes, generally well under 3x, meant for practice or a musical effect. The Evaporator applies spectral stretching from 2x up to 100x, built for dissolving a sound into ambient texture rather than a musical tempo shift.
Can I use this on my phone?
The tool works in a mobile browser, but extreme stretches are processing-heavy, so a long output at a high factor may render slowly or use noticeable memory on an older or lower-powered phone than on a desktop.
What can I do with a stretched drone once I have one?
Download it as audio, or use Continue in another tool to add reverb, run it through the granular synthesizer for further texture, or use the stretched result as an ambient bed underneath other material in a mix.