Time-Stretching Algorithms Explained: Modes and Artifacts

What is time-stretching in music production?

Time-stretching changes how long a piece of audio lasts without changing its pitch. You use it to fit a loop to your project tempo, tighten a performance, or slow a sample down while keeping it in key.

It is different from re-pitching, which speeds up or slows down playback and moves pitch with it, like a turntable. True stretching has to invent or discard audio, and that is where artifacts come from. Researchers Jonathan Driedger and Meinard Müller put it plainly in their 2016 review of time-scale modification: because music mixes harmonic, percussive, and transient components, no single method copes with every signal equally well^[1].

That is why every DAW offers several modes. Once you know the three or four underlying approaches, the menus in Ableton Live, Logic Pro, FL Studio, and Pro Tools stop being guesswork. If you work mostly in Logic or Live, our Flex Time vs Warp comparison covers the editing workflow; this article focuses on the algorithms and their sound.

What are the main time-stretching algorithms?

Most stretch modes fall into four families: time-domain overlap-add methods (including granular ones), the phase vocoder, transient slicing, and hybrid engines that combine them. Each keeps one quality of the sound intact and sacrifices another.

What are the time-stretch modes called in each DAW?

Each DAW names its modes after the material they suit rather than the math behind them. The table maps the same jobs across four DAWs, based on each manufacturer's documentation. For every Ableton control in detail, see our guide to Ableton warp modes.

Two details are easy to miss. FL Studio's time-stretching uses zplane's Elastique Pro version 3 for its e3 modes, and its offline Elastique options give the highest quality but do not follow tempo changes^[4]. Pro Tools lists X-Form as its highest-quality algorithm and notes that it requires rendering^[5].

Which time-stretch mode is best for drums?

Use a transient-based mode for drums: Beats in Ableton, Slicing in Logic, Slice stretch or Slice map in FL Studio, Rhythmic in Pro Tools. These modes keep each hit intact and only move it, so kicks and snares keep their punch.

In Ableton's Beats mode, set Preserve to Transients for percussive audio, then choose a Transient Loop Mode for the gaps: Loop Off lets each segment play and fall silent, Loop Forward loops the tail until the next transient, and Loop Back-and-Forth alternates direction. The Transient Envelope control fades each segment, where 100 means no fade^[2]. FL Studio splits the job by tempo: Slice stretch is meant for drum loops at or below the project tempo, and Slice map for loops at or above it^[4].

1. Check the transient markers
   Zoom in and confirm there is a marker on every kick, snare, and important hat. Add or delete markers before judging the sound.
2. Pick the gap behavior
   When slowing a loop down, try Loop Off with a short envelope for tight, gated drums, or Loop Forward for fuller sustain.
3. Shorten slices when speeding up
   If hits overlap after speeding up, shorten the slices. Logic's Slice Length parameter does exactly this and can also create a gated effect^[3].
4. Compare against re-pitch
   For stretches under a few percent, re-pitching the loop may sound more natural than any stretch mode. The pitch change is small enough to go unnoticed.

If a drum loop needs heavy tempo changes, chopping it and re-sequencing the hits is often cleaner than stretching. Our guide to chopping samples covers that route.

Which time-stretch mode is best for vocals?

Use a monophonic mode for a single voice or solo instrument: Tones in Ableton, Monophonic in Logic and Pro Tools, e3 Mono in FL Studio. These modes track the pitch of one sound, so they stretch it more smoothly than a general-purpose mode.

The conditions matter. Apple notes that Logic's Monophonic algorithm expects a relatively dry recording without audible reverb, and suggests Polyphonic otherwise; its Percussive option protects transients in plucked or percussive tonal parts^[3]. In Ableton's Tones mode, small grain sizes suit material with distinct pitch changes, while larger grains reduce noise but can introduce artifacts^[2]. FL Studio describes e3 Mono as specialized for vocals and solo instruments, with automatic formant preservation^[4].

A vocal sample with reverb, doubles, or a band behind it is not monophonic. Treat it as a full mix, or separate the vocal first and stretch the stem on its own.

Which time-stretch mode is best for full mixes and chord loops?

Use the polyphonic or complex mode for full songs, chord loops, and anything with several instruments: Complex or Complex Pro in Ableton, Polyphonic in Logic, e3 Generic in FL Studio, Polyphonic or elastique Pro in Pro Tools.

Ableton recommends Complex for audio that combines beats, melodies, and textures, and describes Complex Pro as a variation that may give higher quality results, especially on polyphonic textures and full songs, at a higher CPU cost; the manual suggests freezing or resampling tracks that use it^[2]. For pads, drones, and noise without a clear pitch, Ableton's Texture mode adds a Fluctuation control that introduces randomness, which can hide repetitive grain patterns^[2].

When a mix stretches badly in every mode, split it. Separate drums from the melodic parts with a stem tool, stretch drums with a transient mode and the rest with a polyphonic mode, then recombine. That mirrors the harmonic-percussive approach from the research^[1]. Our comparison of AI stem separation tools can help you pick a splitter.

How do you recognize and fix time-stretch artifacts?

Name the artifact first, then change the mode or the settings that cause it. Most problems come from a mode that does not match the material.

How far can you time-stretch a sample before it sounds bad?

Small changes are usually transparent, and artifacts grow with the ratio. As a working rule, changes within about 5 to 10 percent rarely draw attention on suitable material, while stretches beyond 20 to 30 percent start to sound processed in most modes.

Pitch shifting runs into the same limits, because one common way to shift pitch is to resample the signal and then time-stretch it back to its original length, so stretch artifacts carry over^[1]. Our guide on pitch-shifting samples without artifacts covers that side, including formants.

Three habits help. Stretch toward the sample's original tempo when you can, choosing a project tempo near it. Render or freeze the stretched audio once it sounds right, so the result does not change with CPU load or later edits. And when a stretch still sounds wrong, embrace the other option: re-pitch the loop, which is how many classic sample-based records got their character.