An audio pitch changer shifts the perceived pitch of a recorded sound up or down in semitone steps, and the Audio Pitch Changer does this locally in your browser for one file at a time, from -12 semitones (one octave down) to +12 semitones (one octave up), then exports a fresh uncompressed PCM16 WAV. Because the tool uses playback-rate resampling rather than a phase vocoder, pitch and duration move together: raising pitch makes the audio shorter and faster, lowering pitch makes it longer and slower. That single fact defines what the tool is good at, where it falls short, and how you should plan around it. The rest of this guide walks you through the terms, the math, the exact steps inside the Audio Pitch Changer page, and the limits to keep in mind before you click render. Nothing leaves your browser tab; decoding, resampling, and WAV encoding all run locally on the file you already have.
What an audio pitch changer actually does
Pitch shifting means raising or lowering the perceived frequency of a sound without rerecording it. Vocals, instruments, full mixes, and even spoken-word audio can all be shifted. A semitone is the smallest interval used in twelve-tone equal temperament — the same scale a piano keyboard is built on — and an octave is twelve stacked semitones. Each semitone raises or lowers a frequency by a factor of 2^(1/12), which works out to roughly 1.0595. McGill University's course material on MIDI and frequency conversion confirms the same scheme and notes that one semitone divides further into one hundred cents.
Whole-semitone shifts are the default in beginner tools because they land on musically meaningful intervals by design. +1 is a minor second up, +5 is a perfect fourth up, +7 is a perfect fifth up, +12 is a full octave up, and so on. Picking whole numbers tends to sound better than arbitrary decimal shifts, which is why the Audio Pitch Changer exposes a whole-number field rather than a continuous slider.
Why semitones map to one playback-rate number
The relationship between semitones and playback rate is a single short formula: rate = 2^(n/12), where n is the whole-semitone shift you choose. The MDN reference on AudioBufferSourceNode.playbackRate describes the value as a proportion of the original sampling rate: below one the audio slows, above one it speeds up, and any non-unit value triggers resampling. Underneath, the W3C Web Audio specification ties detune cents to a factor of 2^(detune/1200); twelve semitones equal twelve hundred cents, which collapses to the same 2^(n/12).
That single number drives everything the tool does. A reference set of points covers the range without surprises: 0 = 1× and unchanged, +12 = 2× rate and roughly half the original duration, -12 = 0.5× rate and roughly double the duration, with intermediate semitones filling in along the curve. A worked example shows the chain clearly:
- Pick +5 semitones, so n = 5.
- Rate = 2^(5/12) ≈ 1.3348.
- A 3:00 (180 s) source becomes roughly 180 / 1.3348 ≈ 134.85 seconds, or about 2:14.9.
Use the tool itself to lock in the exact figure for your file because decoded frame counts are rounded up at allocation time and will not always match an offline calculation precisely.
The pitch-and-duration tradeoff explained
This is the point most beginners trip on, so it deserves its own section. A playback-rate resampler treats speed and pitch as the same parameter: a faster rate pushes every frequency up, a slower rate pulls every frequency down. If you need to change pitch while keeping the beat grid or word timing in place, that is a time-stretching problem, not a pitch-shifting problem, and the right tool category is a phase vocoder, a SOLA-style engine, or some other dedicated time-stretch algorithm. The Audio Pitch Changer does not try to solve that problem and will not preserve tempo, speech cadence, beat positions, or the original duration.
The trade is honest for many real beginner tasks though. Slowing a song by lowering pitch helps with transcription practice, especially when the original key sits above your comfort range on guitar or piano. Speeding it up by raising pitch helps when you are learning a fast passage or trying to keep up with a recording. Voice effects like the chipmunk and the ogre work precisely because duration and pitch move together as a single parameter. Sound design often wants both pieces moving together, which is exactly the behavior this resampler produces.
How to shift pitch with Audio Pitch Changer
The flow is short, deliberate, and runs entirely in your current browser tab.
- Choose one audio file from your device. Common browser-decodable types — MP3, WAV, M4A, AAC, Ogg, WebM, and FLAC — are accepted, though codec support is ultimately your current browser's job, so a recognized extension does not guarantee every unusual codec profile will decode.
- Confirm the file is under 50 MiB encoded and that the decoded audio fits the limits shown in the table below.
- Choose a whole-number shift from -12 to +12 semitones. +12 is one octave up, -12 is one octave down, and 0 keeps the original pitch.
- Trigger the local render. The browser resamples the entire file through an offline processing pass at the source sample rate.
- Read the stated duration change the interface shows after rendering. The direction and magnitude follow the formula above.
- Download the result. The file is always a fresh uncompressed PCM16 RIFF/WAVE regardless of what you uploaded, and the page intentionally does not play the result inside the tool — open the WAV in a local audio player to hear it.
What you can upload and what you can download
The tool checks limits before it starts any expensive rendering, so a fast failure saves time. The values below are the exact disclosed thresholds, not estimates.
| Item | Limit |
|---|---|
| Encoded input size | up to 50 MiB |
| Decoded input budget | up to 30 million channel-samples |
| Decoded output budget | up to 30 million channel-samples (separate check) |
| Decoded duration | up to 5 minutes |
| Channel count | up to 8 |
| Sample rate | 8,000 Hz to 192,000 Hz |
| Pitch shift range | -12 to +12 whole semitones |
| Output container | PCM16 RIFF/WAVE, interleaved, little-endian |
Two budgets matter: the input sample budget and the output sample budget. Lowering pitch grows the output frame count, so a clip that fits comfortably going up by an octave may fail going down by an octave; the failure message says the requested shift is over budget and that nothing was truncated. Try a smaller shift or a shorter excerpt if that happens.
What to expect from the downloaded WAV
The download is freshly encoded and is not a re-wrap of your source. Output is uncompressed, interleaved, little-endian PCM16 inside a RIFF/WAVE container, with floating-point samples clipped to the legal -1 to +1 interval before conversion. Original MP3/AAC/FLAC compression quality, bitrate, ID3 or Vorbis tags, cover artwork, chapters, cue points, loop metadata, and any other container metadata are not preserved, by design.
Resampling quality is governed by the browser's own resampler and is implementation-defined. Aggressive upward shifts on bright material — cymbals, hi-hats, sibilant vocals — can reveal aliasing or codec artifacts that were quieter in the original. That is fine for practice tracks, sound effects, quick skits, and rough edits. It is not a transparent mastering process. When you need pitch change with the original tempo locked in, switch to a dedicated time-stretching workflow instead.
Useful first projects for beginners
A short list of beginner projects where the duration tradeoff is a feature rather than a bug:
- Transcribe a song by transposing it to a comfortable key. A -3 to -5 shift usually makes many pop or rock recordings more playable on guitar or piano.
- Build a chipmunk or ogre voice line for a skit. +12 reads as one full octave up; -12 as one full octave down.
- Match two short samples to a single key for a mashup. Whole-semitone shifts land close enough to blend for short snippets.
- Speed up a foreign-language recording to tease out the words, then slow it back down to confirm details.
- Preview a key change before committing to a re-record. A quick local render is faster than rebuilding the session.
Where this approach runs short
A few situations make the playback-rate approach the wrong tool, and it is worth knowing them up front:
- Beat-accurate repitching of a multitrack session where the underlying grid must hold. Use a phase vocoder or another time-stretch algorithm.
- Material at the five-minute duration cap or longer; the decoded limit will block the render before any shift is applied.
- Files where embedded metadata matters. The export is a clean, untagged PCM16 WAV by design.
- Very high upward shifts of bright cymbals, hats, or sibilant vocals where browser resampling artifacts dominate.
In those cases, start from a higher-quality source, shorten the excerpt, shift by a smaller number of semitones, or chain the result through a separate time-stretch workflow. If you want a deeper walkthrough that focuses specifically on the local-first side of the pipeline, see Change Audio Pitch in Browser Locally. For context on every step — from file reading through decoding to resampling and WAV encoding — running entirely inside your current tab, the result lives on your device from the moment the download finishes.