Pitch Shift

Some recordings are tuned slightly above or below standard concert pitch while staying internally consistent. Pitch Shift lets you measure that tuning offset and apply a song-local correction so playback, the detected pitch lines, and melody alignment all line up with the chromatic grid.

Opening the Dialog

  • Open a song in the editor.
  • Choose Extras -> Pitch Shift….

The dialog is available for any open song. Buttons that cannot run in the current context stay visible but disabled, with a tooltip explaining why, so the feature is never hidden.

Units

  • The precise unit is cents; 100 cents = 1 semitone.
  • A friendly semitone value is shown alongside, for example +18 cents (+0.18 semitones).
  • Negative values mean the recording is flat (below pitch).
  • Positive values mean the recording is sharp (above pitch).

Analyzing a Song

  • Analyze is available when Aubio is configured and an audio track is loaded.
  • Analysis is always an explicit action — loading audio never scans automatically.
  • Yass prefers the #INSTRUMENTAL track when the song declares one, because a fixed-pitch instrument holds tuning far better than a sung vocal.
  • The dialog shows the detected source offset and, when the measurement is marginal, a low-confidence note.
  • Offsets below 5 cents are reported as “no correction needed” — they are not meaningfully audible.

If a recording’s tuning drifts over time rather than sitting at one consistent offset, Yass reports that no single reliable correction exists. A global pitch shift can only fix a uniform offset.

Applying a Correction

  • Apply recommended stores the suggested correction.
  • You can also type a manual cents value.
  • Remove clears the stored correction.

The correction is stored with the song so it travels with the .txt file. Other song comment data is preserved.

Quick Adjust from the Toolbar

Once a song has a correction, a small cents box appears next to the key icon in the editor toolbar (for example +23 ct). It is a fast way to nudge the correction without reopening the dialog.

  • The box is only shown when the song has a non-zero correction. It appears after you apply a value in the dialog and disappears when the correction is cleared.
  • Adjust the value with the up/down arrows or by typing.
  • The change is applied when you press Enter or click away (focus loss). Each apply re-converts the playback audio, so adjusting is “set, then hear” rather than continuous — this matches the dialog’s live preview.
  • Setting the value to 0 removes the correction; the box then hides itself.
  • Values follow the same convention as the dialog: negative is flat, positive is sharp, in cents (100 cents = 1 semitone).

The box and the dialog edit the same stored value, so changes made in one are reflected in the other.

Hearing the Correction

Hearing the correction happens in two stages.

Live preview (automatic, non-destructive)

  • Playback is automatically heard already corrected.
  • Your original audio files are never touched — the shift lives only in a throwaway temporary file Yass uses for playback.
  • Changing the value re-converts the playback audio so the next playback reflects it. This is “apply then hear”, not real-time scrubbing; an unchanged value replays instantly from cache.

Make permanent (optional)

  • Render to Files… bakes the current correction into new audio files. It is available only when a non-zero correction is stored and an FFmpeg backend is available.
  • Yass renders shifted copies of every audio file referenced by the song (#AUDIO, #VOCALS, #INSTRUMENTAL) with collision-safe names and never overwrites the originals.
  • On full success the song points at the new files and the stored correction is removed (the shift now lives in the files, so it is not applied twice).
  • If you cancel or any render fails, your files, tags, and stored correction are left untouched.

How the Pitch Lines Stay Correct

When the pitch lines are shown from detected audio, the correction is applied to the audio first and the pitch is detected from that corrected audio. The pitch lines therefore already reflect the correction and are not shifted a second time — audio and visuals stay in agreement.