Audio5 min

How to Speed Up or Slow Down Audio Without Changing Pitch

Time-stretch audio 0.5×–2× while keeping pitch intact — perfect for lecture playback, transcription, and music practice.

Speed changes run locally via Zro7 Audio Speed Changer using ffmpeg.wasm. Lectures and recordings never leave your browser.

To speed up or slow down audio without the chipmunk effect, use Zro7 Audio Speed Changer. It uses FFmpeg's atempo filter — a WSOLA time-stretch algorithm — which changes duration while preserving pitch. Perfect for 1.5× lecture playback or slowing music to 0.8× for practice. All in your browser.

How pitch-preserving time-stretch works

Naive playback speed changes both duration and pitch (record at 2× → play at 1× → sounds slow AND lower). WSOLA (Waveform Similarity Overlap-Add) breaks audio into short windows, overlaps them at a new rate, and blends similar waveforms — output is faster but pitch and formants are unchanged. Modern implementations sound near-transparent up to ~1.7×.

Useful ratios

  • 1.25×–1.5× — sweet spot for lecture / podcast catch-up.
  • 2.0× — still intelligible for spoken word; a bit choppy for music.
  • 0.5×–0.8× — transcription, learning song parts, dictation.
  • 3.0×+ — chain multiple atempo passes; quality degrades noticeably.

Steps

  1. Open Audio Speed Changer.
  2. Drop the file (MP3, WAV, M4A, FLAC).
  3. Pick a speed ratio (0.5–3.0).
  4. Preview the result.
  5. Export.

Pitch vs speed — pick your tool

  • Want to change speed but keep pitch? This tool.
  • Want to change pitch but keep speed? Use Voice Changer.
  • Want both to change together (classic tape effect)? Use resample instead of atempo.

Frequently asked questions

Any quality loss?

Time-stretch has small artifacts (transient smearing on drums, slight ghosting on speech). Under 1.7× it's mostly imperceptible; above 2× it's noticeable.

Can I speed up a video?

Extract audio, speed it, and remux — or use a video speed tool that handles both streams together.

Does the output file get shorter?

Yes, proportional to the speed ratio. Bitrate stays the same, so file size shrinks.

Is it lossless?

The DSP is high-quality but not bit-lossless (time-stretch inherently reconstructs signal). Output re-encoded at the source's format.

Any upload?

None. ffmpeg.wasm runs the atempo filter in a Web Worker.

Related posts

← Back to blog