Speed / Pitch

This tool lets you speed up or slow down audio for practical listening and practice workflows. Make a long podcast faster to get through it on a commute, slow a language lesson down so you can catch every phoneme, take a recorded lecture at 1.5× and cut an hour off your review time, or drop a musical backing track by a few semitones to fit your vocal range. The workflow is the same in every case: upload a file, pick a speed (and optionally a pitch change), and download an MP3 at the new tempo.

Two related but independent settings do the work: speed (or tempo), and pitch. Speed is how fast the waveform plays back, measured as a ratio relative to the original (0.5× is half speed, 2.0× is double speed). Pitch is how high or low the tone sounds, measured in semitones where 12 semitones equal a full octave. Depending on the mode you choose, changing speed can preserve pitch (time‑stretch) or shift pitch along with it (the classic “tape speed” effect), and pitch can be shifted independently without changing speed.

Time‑stretch (preserve pitch) is the right mode for almost all listening use cases. Doubling a podcast’s speed without pitch preservation turns every voice into a chipmunk; with pitch preservation the speaker just sounds faster, which is usually what you want. For language learning, 0.75× or 0.5× with pitch preservation gives you clearer consonants and vowels without the weird low‑pitched drone that a naïve slowdown would produce.

Pitch shifting in semitones is where musical use cases live. A singer whose range doesn’t quite fit a recording can drop the backing track by 2 or 3 semitones to get a comfortable key. Guitar players practicing along with a reference can shift the track up or down to match their tuning. Songwriters sketching cover ideas can try a song in different keys in seconds. One semitone is the smallest step between adjacent notes in Western music; twelve semitones make a full octave.

Safe ranges and artifact behavior: small changes (0.85×–1.25× speed, ±2 semitones pitch) are almost transparent on most sources. Moderate changes (0.6×–1.6× speed, ±5 semitones) still sound natural for speech but start showing processing artifacts on music — slight “swishing,” transient smearing on drums, or a metallic tone on complex mixes. Extreme changes (below 0.5× or above 2.0×, or more than one octave of pitch shift) sound noticeably processed and are usually only useful for effect or analysis rather than enjoyable listening.

Practical listening presets people actually use: 1.25× is a gentle podcast speedup that still feels natural. 1.5× is the sweet spot for review and study content, shaving a third off the runtime. 1.75× and 2.0× work for information‑dense material but require more attention. 0.75× is a comfortable slowdown for dense technical explanations. 0.5× is useful for transcription and language drills but sounds obviously slowed. For music, ±1–2 semitones is standard for key matching; ±3 is the upper end before a song stops feeling like itself.

Why audio sounds “robotic” at extreme settings: time‑stretch algorithms work by analyzing short windows of the signal and repeating or thinning them to change duration without changing pitch. When changes are small, those repetitions are imperceptible; when they are large, the windows start to become audible as a texture, and transient events (drum hits, hard consonants) smear across adjacent windows. Complex polyphonic music stresses the algorithm most; plain speech stresses it least. Smaller increments always sound cleaner than one extreme jump.

Quality and workflow tips: test a short region first before committing to the full file, especially on music. Work from the highest‑quality source you have — a heavily compressed MP3 will show more artifacts after stretching than a WAV or lossless source. Do the speed/pitch change once, not in multiple passes, because each pass adds both algorithmic artifacts and (for this tool’s MP3 export) lossy encoding. If the final file will sit next to other audio, follow up with Normalize so the loudness matches.

A sensible pipeline: (1) if your source is a video, extract the audio first with MP4 → MP3 or MP4 → M4A, (2) trim the exact region you want to play with, (3) apply speed and/or pitch changes here, (4) normalize loudness at the very end. Doing the speed change before normalization is important because changing tempo and pitch can shift perceived loudness; normalizing afterwards produces a consistent final level. Save the original source file so you can rerun with different settings without compounding quality loss.

Things this tool is not for: elaborate multi‑band pitch correction, vocal tuning, removing vocals from a track, or changing speed for only a section of the file. It applies one speed and (optionally) one pitch shift uniformly to the entire file. For surgical musical tasks — auto‑tune‑style pitch correction, key‑change automation inside a song, isolating a specific part — a full DAW with pitch/time plugins is the right tool. This tool is built for simple, whole‑file transformations you want done in seconds.

Duration math: the new duration equals original duration divided by the speed ratio. A 60‑minute lecture at 1.5× becomes 40 minutes; at 2.0× becomes 30 minutes. A 3‑minute song at 0.75× becomes 4 minutes. This is exact — the tool doesn’t round — so you can plan review windows or playlist slots confidently. If you need a very specific target duration, pick the speed that divides the original evenly into it (e.g. a 48‑minute target from a 60‑minute source needs 1.25×).