Singing Recorder — Mic, Monitor, Upload, Playback & Download

Record yourself singing in the browser: one strip for mic, optional headphone monitoring, record and review, timeline scrub, and export the Original capture, or convert to MP3, WAV, OGG, or M4A — files stay on your device, and processing uses the same local Web Audio pipeline as the rest of MusicalBoard (no upload of your takes to our servers). The recorder uses a format your browser supports: most non-Safari browsers prefer WebM with Opus, while Safari is first allowed to pick its own default format and only falls back to a tried list — MP4/AAC, then AAC, WAV, or WebM — if that default fails. The other export formats are decoded and re-encoded in the browser from whichever one your device actually captured. Use this page when you want a dedicated recorder workflow; pair with the pitch monitor or spectrum tools from the nav when you also want visual feedback.

Record your vocal practice and hear yourself objectively

Playback reveals habits you might not notice while singing — breathiness, pitch drift, and phrase endings — so you can focus the next run on what actually needs work.

Real-time monitoring to catch pitch issues while you sing

Open monitoring when you want latency-safe foldback (headphones strongly recommended). Adjust the monitor gain from the strip’s volume icon menu; pair with the Vocal Pitch page when you want a visual trace at the same time.

How to export your vocal practice session for review

Open the download menu, pick a filled slot (A–D), then Original, MP3, WAV, OGG, or M4A. Each take is its own file. Filenames include the slot letter and recording time. Everything stays on your device.

Compare independent takes from the session tray

The Takes and backing panel always shows four slots (A–D). Select an empty slot, then record or upload into it. Auto-Tune and Vocal Isolator do not overwrite a take by themselves — Save to take asks which slot to write, then a name. Each slot has an × to delete that take, plus Delete all.

Upload a file and analyze it with the same playback tools

Use the upload button in the strip to bring in an existing take and analyze it with the same playback-driven pitch, spectrum, and range pipeline. Upload analysis runs locally in the browser and then behaves like a finished take: press Play to drive charts and gauges from the uploaded history.

Practice with a backing track and choose whether it is written into the take

Load a local accompaniment from the Takes and backing panel, set the start point on the backing gauge, then turn on Play under take so the file starts from that cue when you hit record. Keep Voice only if the saved take should stay dry, or switch to Mix into take to bake the accompaniment into the recording. Pitch analysis still uses the microphone only.

Low—High
Pitch gate9%
Listening — sing any note

Sing with the microphone,
or upload an audio file.

How to use the Controller (shared strip)

This strip is the shared transport for the microphone graph tools: one activation feeds Vocal Pitch, Level & Waveform, and dashboard pitch widgets with the same buffers and timing. Metronome and Scale Generator do not use the mic path, but their cards sit in the same app shell—each saves its own settings in local storage separately from this strip.

Getting started

  • Click the microphone button and allow browser access when prompted. The icon switches between mic-on and mic-off artwork while capture is live.
  • Recording is tied to the audio pipeline: capture a take, then use Play to review it or open other tool pages—the last recording stays available until you record again or reset.
  • Tooltips describe each icon on hover (desktop). On mobile, open the Controller entry in the header menu if the strip is collapsed.

Main controls

  • Microphone — Starts or stops capture from your mic. Disabled while a recording is playing or when the host page temporarily locks controls.
  • Monitor — Toggles live monitoring of your input (speaker vs muted-speaker icon). Prefer headphones and a conservative monitor level to avoid feedback.
  • Monitor output level — Click to open a small panel with a range slider (0 up to the app’s maximum gain) and a percentage readout. Dragging updates foldback immediately; the chosen level is saved in browser local storage for next time.
  • Play / Pause — Plays or pauses the last captured take. Disabled while the mic is recording, until a take exists, or while controls are locked.
  • Jump to start — Moves the playhead to 0:00. If playback was running, it keeps playing from the beginning. Disabled under the same rules as Play when there is no take.
  • A↔B loop — Toggles section loop mode (see A↔B section loop below). On desktop it sits before Upload; on the mobile Controller grid row 3 runs Loop, Upload, Download, then Takes and backing.
  • Upload — Imports an audio file and converts it into playback history for the same analyzer views. During processing, transport and edit controls are temporarily locked; use Cancel in the processing overlay to abort (see Upload & restore processing below).
  • Download — Opens a menu: pick a filled slot (A–D), then format (Original, MP3, WAV, OGG, M4A). Each take is its own file. Disabled when there is nothing to export or while the mic is live.
  • Takes and backing — Opens the session tray. Four slots stay visible. Select an empty slot and record into it, or save a processed file from Auto-Tune / Vocal Isolator. × deletes one take; Delete all clears the bank. Rename is under the selected filled slot.

Skip ±10 seconds (transport icons)

After you finish a recording or upload a file, two extra transport buttons appear on either side of Play. They use the same enable rules as Play (not available while the mic is capturing, during upload analysis, or when controls are locked).

  • Back 10 seconds — Double left-pointing triangles (skip-back / “rewind” style). Moves the playhead 10 seconds earlier, clamped to the start of the take (0:00).
  • Forward 10 seconds — Double right-pointing triangles (skip-forward style). Moves the playhead 10 seconds later, clamped to the end of the take.
  • During playback, skipping does not stop audio: the engine seeks to the new time and keeps playing, and the pitch graph, spectrum, and Range Map LIVE readout follow the new position.
  • While paused, only the timeline position changes until you press Play again.
  • The progress bar, current time label, and chart playheads update immediately—same mechanism as dragging the scrubber.
  • Desktop layout — Order is: … volume, devices, Jump to start, −10, Play, +10, loop, upload, download, takes and backing, then the progress bar.

Backing track

A backing track is a local accompaniment you hear while you sing. It is not sent to MusicalBoard servers. Open the overlapping-card icon (Takes and backing) on the Controller strip; Backing is the lower block. During a mic recording every backing control is disabled, and stopping the take also stops the backing and rewinds it to the gauge cue.

  • Load backing — First row. With nothing loaded, tap to pick an audio/* file from this device (the same formats your browser can decode: typically WAV, MP3, M4A/AAC, OGG, WebM). After load, the button keeps a fixed width; a long file name scrolls right-to-left like a music player.
  • Remove backing — When a file is loaded, tap that same first-row button (the file name) to stop playback, drop the file, and return the label to Load backing.
  • Play backing / Pause backing — Second row. Previews the file on a loop so you can check level and arrangement before you record. The label switches to Pause backing while it is playing. Disabled until a file is loaded or while you are recording.
  • Backing gauge — Same style as the take progress bar. Drag it to set the cue; the next recording starts the accompaniment from that time. While the backing is playing, the fill follows the playhead.
  • Play under take — On by default after you load a file. When it is highlighted and you start a new mic recording, the backing seeks to the gauge cue and plays under you. Turn it off if you only want to preview and do not want the file to start with the next take.
  • Voice only / Mix into take — Voice only (default) keeps the saved take dry. Mix into take bakes the accompaniment into the recorded file from the same gauge cue after you stop. Pitch, spectrum, and range still use the microphone only either way.
  • What is analyzed and stored — Pitch, spectrum, range, and the A–D take queue use the microphone (or an uploaded take) only. Headphones are strongly recommended so the accompaniment does not leak into the mic and pull the pitch trace.
  • What is saved — The backing file, cue, and mix choices are stored in IndexedDB with the A–D takes (expiry checked on restore about 24 hours after the last save). Refreshing or opening another tool and coming back restores them. A take saved with Mix into take keeps the mixed audio.

Use this when you want a piano track, a minus-one, or a metronome-free groove while you record a phrase. For click-only practice, the Metronome tool is still the better fit.

Take bank (A–D)

Open the overlapping-cards icon to see four slots. Takes use IndexedDB with expiry checked on restore about 24 hours after the last save; nothing is uploaded.

  • Slots A–D — Always visible. Tap an empty slot, then record or upload into it. Tap a filled slot to play that take. Each slot shows its name and length.
  • Name — After you record or save, a dialog asks for a name. The default is the date and time. Rename from the panel any time.
  • Overwrite — Recording, uploading, or Save to take into a filled slot asks first, and shows the take name that would be replaced.
  • Delete — The × on a slot removes that take after confirm. Delete all clears every slot.
  • Save to take — Auto-Tune and Vocal Isolator keep the processed file on the tool page until you press Save Auto-Tuned as Take or Save Vocal as Take, choose a slot, and name it. That writes a new independent take, not a version of the original.

A↔B section loop

Use the circular-arrow loop button to repeat a slice of the take without manually scrubbing back each time. Loop mode is designed around the waveform minimap (canvas under the progress bar) so the main gauge does not shift or break layout.

  • Turn on — Tap the loop button after a take exists. The minimap stays open, shows the waveform overview, and draws A (orange) and B (teal) markers with a highlighted band between them. Default range: start at the current playhead, end about five seconds later (clamped to the take length).
  • Adjust A and B — Drag the A or B badge on the minimap (about 14px hit target). Markers cannot move closer than ~0.5s. Timestamps appear on the badges while dragging.
  • Scrub while looping — With loop on, the progress range input is non-interactive (display only). Click or drag on empty minimap canvas to seek; the playhead line and charts update the same way as gauge scrubbing.
  • Automatic wrap — During playback, when the playhead reaches B, the engine seeks to A and continues if audio was playing—same seek event as manual scrubbing, so pitch/spectrum views stay in sync.
  • Turn off — Tap loop again. The minimap closes unless you are hovering or dragging the gauge on desktop. Looping stops; playback is not paused if it was already running.

On desktop, hovering the progress bar still shows a transient minimap when loop is off. At 930px width and below, open the header Controller drawer for the same controls; row 2 of the grid holds Jump to start, ±10s, Play, and Loop across five equal-width slots (see Mobile layout).

Mobile layout (narrow screens)

  • At 930px width and below (header Controller drawer and this detail page), buttons use a grid instead of a single crowded row:
  • Row 1 (4) — Microphone, Monitor, Monitor volume, audio devices.
  • Row 2 (4) — Jump to start, Back 10s, Play/Pause, Forward 10s.
  • Row 3 (4) — A↔B loop, Upload, Download, Takes and backing.
  • Row 4 — Full-width playback progress bar with time labels.
  • Touch targets stay at least ~40px tall; icon artwork scales down slightly so the strip fits without horizontal scrolling.

Progress bar & waveform minimap

  • Shows current time and total duration (mm:ss.cc — centiseconds, two digits after the decimal) next to a wide range control. Drag, click along the track, use touch to seek, or use the ±10s buttons beside Play when the mic is not recording.
  • Seeking dispatches playback position to the analyzer so pitch and spectrum views jump to that moment in the take.
  • The bar is disabled while the microphone is actively capturing so you do not scrub mid-take by mistake.
  • A compact waveform minimap appears below the bar on hover (desktop) or while scrubbing. It mirrors the take overview and the current playhead position.
  • When A↔B loop is on, the minimap stays visible, draws the loop region and A/B markers on the canvas, and receives all seek and marker-drag input; the range input above is dimmed and ignores pointer events.

Plot settings (on pitch pages)

  • On Vocal Pitch and similar views, a separate bar above the chart still holds pitch plot options (MIDI range, overlays, sensitivity). Those values are stored per page, not inside the Controller strip.

Upload & restore processing (overlay)

A full-screen overlay appears while the browser decodes audio and rebuilds pitch/spectrum history. Progress is shown as a percentage. Transport and most strip actions stay locked until processing finishes or you cancel.

  • Fresh upload — Message: Analyzing uploaded file…. Runs the same offline analysis pipeline as a finished mic recording. On success, the take behaves like a normal recorded session (Play, scrub, charts). Cancel aborts the job and dismisses the overlay; the previous IndexedDB cache (if any) is left unchanged.
  • Restore after refresh or revisit — Message: Restoring your recording — rebuilding analysis…. Shown when you reload the page or open a pitch tool again and a non-expired take is still in IndexedDB. The app reads the stored audio bytes and re-runs analysis locally (history is not stored on disk, only audio + light metadata).
  • Cancel on restore — Stops the rebuild immediately, closes the overlay, and deletes the cached take from IndexedDB plus any in-memory session snapshot. The next full page load will not prompt you to rebuild the same file again. Use this if you do not want the old take restored.
  • Route changes without a full reload may still pass a take through an in-memory snapshot (no overlay); only the disk-restore path uses the rebuild overlay and Cancel-to-clear-cache behavior.

Recording cache and delete behavior

  • After a recording finishes or an upload completes analysis, the audio bytes plus a small metadata record (sensitivity, playback time, MIME type, timestamps) are written to your browser's IndexedDB on this device. Nothing is uploaded to a server, and other browsers or other devices won't see the take.
  • The cache lets takes A–D and the backing file survive a page refresh or a move to another tool: when you return, each take's audio is restored and analysis is rebuilt locally. The selected slot is analyzed first; the others fill in in the background.
  • Retention is capped at roughly 1 day (24 hours) from when the take or backing was saved. Once an entry is older than that, the next page load drops it.
  • Closing the tab does not wipe the cache immediately. The 1-day TTL is the cleanup guarantee. Delete still removes the selected take right away; a full reset clears the bank and backing.
  • Recording or uploading writes into the slot you selected. If that slot already has a take, you must confirm overwrite first.
  • The × on a slot removes that take after you confirm its name. Delete all clears every slot.
  • Plot settings (MIDI range, overlays, sensitivity, monitor gain, etc.) live in `localStorage` and are not affected by Delete, restore Cancel, or cache expiry — they always persist between visits.

Tips

  • Use wired headphones when monitoring to reduce feedback; on phones, keep monitor volume low.
  • Safari / iOS may require a tap to resume audio after the page loads—if meters stay flat, tap the mic or reload once with permission granted.
  • Upload guide: max duration is 10 minutes and max file size is 50MB. If either limit is exceeded, the app blocks analysis and shows a warning.
  • Extension policy: there is no strict hardcoded extension whitelist in the UI; the picker accepts `audio/*` and actual support depends on your browser’s decoder. In practice, WAV, MP3, M4A/AAC, OGG, and WebM are usually safe choices.
  • If a file extension or codec is not supported by your browser, decoding can fail during analysis. Convert the file to a common format (for example WAV or MP3) and upload again.
  • After upload starts, wait for analysis to finish before interacting with playback controls; once complete, Play uses the uploaded timeline just like a recorded take.
  • This Controller detail page keeps the analyzer hidden while the strip runs against a hidden PitchAnalyzer instance; on the dashboard, the same controls sit above the cards.
  • Metronome BPM/meter/subdivision preferences and Scale Generator root/interval/BPM/direction preferences each have their own local storage keys—adjust them on those cards without affecting monitor gain here.

Workflows worth setting up once

  • Load a backing, set the gauge cue, enable Play under take, pick Voice only or Mix into take, record into slot A, then save an Auto-Tune or isolated pass into B with Save to take and compare playback.
  • Enable A↔B loop on a shaky high note or melisma, drag A/B on the minimap, and let playback cycle while you watch the pitch graph at that moment.
  • Use Jump to start before a full listen-through, then loop a problem bar for deliberate practice.
  • Monitor while you vocalize to catch pitch or tension issues before they become habits.
  • Export a warmup or lesson clip for a coach, or archive monthly to hear long-term progress.

Recording live versus uploading a file

Every tool that uses the Controller — this recorder, the pitch monitor, spectrum, range test, Auto-Tune, and Vocal Isolator — reads from the same slot system, so the choice between recording and uploading is really about where the audio comes from, not which tools can use it afterward.

Record liveUpload a file
When it fitsChecking a warmup or take right now, with the mic you haveReviewing something recorded elsewhere — a phone voice memo, a lesson recording, a mixed song
Length limitStops automatically at 10 minutesUp to 10 minutes, and the file also has to be under 50 MB
What you hear while it runsOptional headphone monitoring of your own voiceNothing — the file is already finished audio
Room and mic qualityWhatever your current room and microphone pick upWhatever quality the file already has — a phone-recorded voice memo stays a phone-recorded voice memo
Backing trackCan play underneath while you record, voice-only or baked inNot available — the file is loaded as-is

A practical split: use recording for anything you want to check in the moment, and upload when you already have a file — a lesson recording to run through the pitch monitor, or a short song clip you want to try through Vocal Isolator before deciding whether to record over it.

Compare two short vocal takes without losing the original

Use the same phrase, microphone position, and input gain for both takes. Change one practice choice at a time so you can hear what changed. This is a suggested workflow, not a measured claim that a particular exercise improves singing.

  1. Open Takes and backing, select empty slot A, and record a comfortable 10–20 second phrase. Stop and play it back before moving on.
  2. Select empty slot B and repeat the same phrase. Keep the first take in A so the second recording does not replace your reference.
  3. Select A, listen to the phrase ending, then select B and listen to the same part. Compare timing, clarity, and audible pitch drift; a louder take is not automatically a better one.
  4. Open Download, select A, and choose Original. Repeat for B. Keep both files outside the browser before clearing takes or site data.

For a visual check, open the Vocal Pitch Monitor and play the selected take. If the recorded level clips or quiet passages are buried in noise, run Mic Room Diagnostics before repeating the comparison. Those displays cannot identify the physiological cause of a pitch error.

Choose the file to keep or share

  • Keep the source: Original preserves the selected slot’s source bytes and uses the detected container extension. An uploaded MP3 stays an MP3; a microphone capture keeps the format that browser recorded.
  • Share with a player that needs another format: choose MP3, WAV, OGG, or M4A. Conversion re-encodes the audio; it cannot recover detail already lost in a compressed source.
  • Keep a correction separately: use Download Auto-Tune on the Auto-Tune page, or Save Auto-Tuned as Take into another slot. Exporting Original from the dry source slot does not export an unsaved correction.

The audio export format guide explains when to keep Original, when to convert for sharing, and what conversion changes. Keep the source alongside any converted copy so another format choice does not become your only archive.

Read the recording walkthrough for setup and playback, or the browser storage guide for what survives a refresh. Cached takes are temporary; a downloaded file is the copy you control outside the site.

Questions

Does the singing recorder upload my voice?

No. Capture, playback, and export stay in the tab. A take is written to IndexedDB on this device so a refresh can restore it. MusicalBoard does not receive the audio.

Which file is the recording, and which files are converted?

The recorder uses a format your browser supports. Most non-Safari browsers prefer WebM with Opus. Safari is first allowed to pick its own default format, and only falls back to a tried list — MP4/AAC, then AAC, WAV, or WebM — if that default fails. Choose Original if you want the capture exactly as your browser made it; MP3, WAV, OGG, and M4A are encoded afterward in the tab from whichever one that was. WAV here is 16-bit PCM mono, so it is not a bit-for-bit copy. Filenames include the slot letter and the recording time.

How do slots A–D relate to Auto-Tune and the isolator?

Open Takes and backing. Record or upload into an empty slot, or tap a filled slot to play it. Auto-Tune and Vocal Isolator do not replace the source slot on their own. Press Save to take on those pages, pick a slot, and name the file. × deletes one slot. Delete all clears the bank.

What does the backing-track switch do?

Load a local file, set a cue on the backing gauge, and turn on Play under take before you record. Voice only keeps the backing out of the saved file. Mix into take bakes it in. Pitch analysis still follows the microphone, not the backing.

Can I upload a file instead of recording?

Yes. Upload in the Controller imports a file you have the right to use, up to 10 minutes or 50 MB. Play then drives the pitch, spectrum, and range views from that file. Nothing is sent to a server.

What happens if I close the tab?

Takes are checked for expiry on restore about 24 hours after their last save. Saving can renew that time and page exit attempts to save the bank. Clear site data to remove the cache; download files you need to keep.

I denied the microphone prompt — how do I record now?

The record button stays inactive until the browser's own permission for this site is granted. Look for a camera/mic icon or a padlock in the address bar, open the site's permissions there, and switch microphone to allow, then reload the page. The exact wording and icon differ by browser.