How to use Fugue State
Fugue State turns an audio file into an interactive spectrogram, constant-Q transform, and live spectrum — then lets you slow it down, loop it, re-tune it, and play along with live microphone feedback. The analysis runs on our servers; everything else happens in your browser. This guide walks through the whole workflow.
New here? Try a demo in your browser first.
1. Get started
Sign in. Create an account and log in. You land in the workspace: your project list, and where a new one is made. Nothing to install.
Create a project. Press Create New Project, give it a name and choose a file (drop it in the dialog or Browse Files) — or just drop an audio file anywhere on the window. Common formats work (MP3, WAV, FLAC, and more).
Wait for the analysis. The file uploads to our storage and is analysed on our servers — FFT, constant-Q transform and onset data. The project’s card shows the upload, then the processing progress; a song can take several minutes. When it finishes, press Open. The analysis is kept with the project, so re-opening it later is quick.
Just looking? The demos run the same displays on pre-analysed tracks, no account needed.
2. Read the displays
The workspace is a grid of resizable panes. Right-click a pane to split it, change its view, pick a channel (left, right, or both), or close it — so you can watch several representations of the same moment at once.
- Spectrogram — energy across frequency (vertical, perceptually/mel-scaled) over time (horizontal). Bright = loud. The playhead stays centered as audio scrolls past.
- Constant-Q transform (CQT) — like the spectrogram but with a log-frequency axis aligned to musical pitch, so octaves are evenly spaced and notes line up. Best for hearing harmony and melody.
- Spectrum (Fourier) — the live frequency spectrum of the moment currently playing, updating in real time.
- Waveform & timeline — the amplitude envelope and a zoomed-out overview for scrubbing across the whole track.
4. Control playback
The transport bar at the bottom drives playback:
- Play / pause with the transport button.
- Speed — slow a passage down to study it. Pitch is preserved, so it stays in tune at any speed.
- Loop — set a loop region to repeat a tricky bar over and over. Loop handles appear on the timeline.
- Volume — adjust output level without affecting the analysis.
5. Tune & shape the sound
Open the audio-controls panel for finer control over what you hear. The visual displays follow these changes so the coloring always matches the sound:
- Tuning — pitch-shift up or down by up to an octave (±1200 cents, in semitone steps) without changing speed, via a real-time phase vocoder. The spectrogram/CQT shift their frequency axis to match.
- Channel mode — listen to Stereo, Left only, Right only, Mid (L+R), or Side (L−R) to isolate what sits in the center vs. the sides of a mix.
- 10-band EQ — boost or cut bands from 32 Hz to 16 kHz (±10 dB) to emphasize an instrument or tame a frequency.
6. Practice mode
Practice mode lets you play or sing along with the track and capture takes to compare against it. Toggle it with the microphone button in the toolbar. The view splits into the reference spectrogram (top) and your take (bottom), aligned to the same timeline so you can see exactly where you land.
- Arm & record — the record dot in the take selector arms the mic; press play and it captures a new take. Selecting the “new take” slot auto-arms; loading a past take switches to playback.
- Punch in — select an existing take, arm, and play to re-record over it from the playhead, keeping the rest.
- Loop & stack — while looping, each pass finalizes a take and rolls over to a fresh one, so you build up attempts bar by bar.
- Compare — play a take back alongside the song. Takes stay in tune and in time even when you change speed or tuning, so the comparison is always fair. Pitch detection overlays show how close your notes are.
Takes live in the current session. Before you close the project, open File → Export in the toolbar: Export Takes downloads one file per take, and Export Project Audio downloads the project’s audio as a WAV.
7. Vocals and stems
Separating a track into stems — a “vocals removed” source you can switch a pane (or the global track) to — is not available in the hosted app yet. When a project does have stems, they appear as extra sources in the pane’s source menu; a project with only the original mix shows no such menu.
8. Tips
- Slow a fast run to half speed and zoom the time axis to read individual notes.
- Switch a pane to CQT when you care about pitch/harmony; the spectrogram when you care about timbre and noise.
- Narrow the frequency minimap to a single instrument’s range to see it clearly.
- Use Mid/Side mode to pull a centered vocal forward or push it back.
- Loop a hard bar, drop the speed, then bring it back up as it gets comfortable.