Fast speech,
slowed to your pace.

What slowing does to a voice — the honest version

In this tool, tempo and pitch are linked: slow a recording and the voice gets deeper as it gets slower, the way a record does when the turntable loses speed. At 75%, speech sits about five semitones lower than the speaker actually sounds.

For language learning, that trade is usually a good one. You are not studying the speaker's voice — you are studying the words, and the words are untouched. Every syllable, every liaison between words, every swallowed ending is still exactly where the speaker put it, just stretched out enough for your ear to catch it. A deeper voice pronouncing the same sounds is a small price for finally hearing where one word ends and the next begins.

If you want slowed speech at the speaker's true pitch — natural-sounding, just slower — that is unlinked tempo and pitch, and it is an app feature: the Y2K Player app slows tempo while holding pitch, and saves the setting per recording. The web tool on this page is the quick, linked version, and for comprehension work it carries you a long way.

75% is the working speed — voices stay clear to about 70%

Slowed speech has a comfort zone. Down to roughly 70% of the original speed, a voice remains easy to parse: lower, slower, but plainly a person talking. Push much below that and the linked pitch drop starts to work against you — the voice turns unnaturally deep and the flow of speech stops sounding like speech.

That is why this page starts at 75%. It is inside the clear zone with room to spare, and it is slow enough to matter: a quarter more time on every syllable is the difference between hearing a blur and hearing three separate words.

Slow is for hearing, not for staying

The point of slowing speech is to catch what fast speech was hiding — then return to full speed and confirm you can still hear it. A pass at 75% followed by a pass at 100% teaches more than three passes at 75%, because native speakers will not slow down for you. Treat the slow speed as a magnifying pass, not a place to live.

Loop the sentence, shadow the speaker

The fastest gains come from looping one sentence and working it until it opens up.

Set the loop around a single sentence — the waveform shows pauses between sentences as visible gaps, so the boundaries are easy to place. Let it repeat at 75% until you can separate every word, then speak along with the recording, matching the speaker's timing and rises and falls. This shadowing works at slowed speed first, then at 100%. The lowered pitch does not hurt here either: you are copying the timing and the vowels, not the speaker's voice depth.

What to slow down

The material matters as much as the speed. The recordings worth slowing are the ones with real connected speech: interviews, conversational podcasts, radio shows — places where words run into each other the way they do in life. Scripted learner audio is already slow and already over-articulated, so slowing it further teaches you little. Songs are worth a special mention: sung lyrics compress and stretch words in ways textbooks never show, and at 75% you can finally hear what the singer does to fit the words into the melody. One warning applies to all of it: pick material where you already understand most of the words on paper. Slowing reveals sounds; it does not supply vocabulary.

A note on listening levels

Language study means long sessions, usually on headphones, often with the same sentence looping many times. Keep the volume at the low end of comfortable — comprehension comes from repetition and attention, not loudness. If a word stays buried, slow down another five percent instead of turning up; that is what the slider is for.

Start to finish: one listening session

1. Drop your recording into the tool above — a podcast episode, an interview, a lesson recording, any audio file.
2. Play it once at 100% and note where you lose the thread.
3. Pull the tempo to 75% — this page's default — and set the loop around the first sentence that beat you, using the pauses visible on the waveform.
4. Loop it until every word separates. If one spot stays welded together, drop briefly to 65% for that phrase alone.
5. Speak along with the loop, then take the tempo back to 100% and confirm the sentence now makes sense at full speed.
6. If you want a slowed copy for the commute, export it as MP3 or WAV — the free export covers the whole recording, with a short spoken note after the end that you can crop off.

For the daily-listening habit

You love learning a language by ear? There is a player for that.

Y2K Player is a music player for iPhone and iPad that treats any audio the way this page treats it — and it adds the thing the web tool honestly lacks: unlinked tempo and pitch, so slowed speech keeps the speaker's natural voice. It also saves each recording's settings, which matters for study material: episode four reopens at your speed, at your place, with your loop.

Loop and trim on a proper waveform make sentence work quick, playlists keep a course's episodes in order in your library, and everything runs offline with no account — the commute is where listening practice happens. Free to use. Pro is one payment — forever.