Full specification

Files on the device

Play the music already on your phone, with no account. Aurora notices new, moved and deleted files by itself, and a full rescan reports what it added or removed.

• 15 file types: mp3, flac, m4a, aac, ogg, opus, wav, aiff, aif, wma, alac, ape, dsf, dff, mka
• Library rebuilt about 2.5 s after files change; pull to refresh
• Full rescan times out after 5 min, a quick one after 60 s
• Android 8 and 9: folder walk of up to 20,000 files, 12 levels deep
• Albums saved as one file split into their tracks with .cue sheets: Add Cue Sheet Folder (since V3.2.0)

• Needs the Music and audio permission on Android 13 and newer, or storage access before that.

Also in Local folders on the desktop · Lyrics sources

Navidrome, Subsonic and OpenSubsonic

Sign in to Navidrome or any Subsonic-compatible server for your albums, artists, playlists and starred songs, plus server radio, folders and synced lyrics. Requests are signed with a salted token instead of sending your password.

• 12 items per Home shelf
• 30 tracks per similar-songs radio
• Up to 500 albums in the Albums tab; songs load as you scroll

• Tags can't be edited on a Subsonic server from Aurora; server tag editing is Jellyfin only.

Also in Navidrome, Jellyfin and Plex on the desktop · Merged library

Jellyfin

Connect with your Jellyfin user name and password. Favourites, similar-songs radio, folders, word-timed lyrics and metadata editing all work against the server.

• 12 items per Home shelf; up to 50 favourites
• 30 tracks per similar-songs radio
• Up to 500 albums in the Albums tab
• Rename playlists and edit their descriptions

• Metadata edits need an account with edit permission, and cover art isn't changed on the server.

Also in Navidrome, Jellyfin and Plex on the desktop · Built-in tag editor

Plex Media Server

Connect straight to your Plex server with a Plex token. Every music library the token can see appears, with Plex's own recommendation shelves, station radio, lyrics and editable playlists, and a 4 or 5 star rating counts as a like.

• Port 32400 by default
• 12 items per shelf; 5 s limit per recommendation section
• Up to 30 tracks per radio

• Sonic matches need Plex Pass and a finished sonic analysis on the server.
• No folder browsing or tag editing for Plex.
• Shared-library permissions still apply to playlist changes.

Also in Navidrome, Jellyfin and Plex on the desktop · Sonic radio and Auto DJ

Best-source playback

When the same recording exists in several places, Aurora plays the copy you prefer: by default a file on the device, then a download, then a stream. You can reorder the three tiers.

• Lookups take up to 3 s per source
• Results cached for 60 s, up to 128 entries
• ±4 s duration tolerance
• Up to 6 copies tried within 6 s for a streaming track
• Three settings under Streaming tracks: Play from your library, Match by Title and artist or Title and length, and Lossless copies only

• With Prefer local copies off, the stream plays instead.

Also in Merged library · Format and source badges

Extension libraries

Install a compatible extension app to add a read-only music library alongside your other sources.

• Up to 10,000 tracks, 100 per page
• Downloads up to 512 MiB

• Read-only, and kept out of the merged library in API 1.
• Found under Settings → Advanced audio → Extensions.
• Android only.

Also in Extensions · Extension SDK, API 1

Saved accounts and quick switching

Keep several accounts signed in, across servers and services, and switch from Settings → Accounts or the sign-in screen without typing credentials again.

• Switching stops playback from the previous account; each account keeps its own queue.

Also in Queue, saved per account · Protected server passwords on the desktop

Home feeds

Home opens with the newest albums, then shelves for recently played, your playlists, favourites, most played, recommended albums, artists and new releases. Plex, YouTube Music, Green Music App and SoundCloud add views of their own.

• 5 newest albums at the top
• 8 sections, each can be hidden
• Offline, with downloads from two or more kinds of source, Home switches between them
• On a phone, hold Home in the tab bar to switch feeds when there's more than one
• On a streaming service's Home, a shelf of 8 or more songs shows as a 4-row grid you play from, like YouTube Music's Quick picks

• With no source yet, Home offers Add Source.

Also in Player and Home customisation

Search

Search artists, albums, songs and playlists as you type, narrow with All, Songs, Albums, Artists or Playlists, and reopen recent searches. With YouTube Music, Green Music App or SoundCloud merged in, you choose between your library and their catalogues.

• Results 300 ms after you stop typing
• Last 12 searches kept
• Voice search on Android: the mic button, or hold Search in the phone's tab bar

• Voice search depends on the device's speech service.

Library tabs, likes and pins

Your Library has tabs for playlists, albums, artists, songs and downloads, quick tiles, a pinned shelf, and a Liked Songs collection that follows each source's own likes.

• 6 tabs, 4 sorts
• Shuffle all takes up to 10,000 songs
• Single: up to 2 tracks under 15 min; EP: up to 6 tracks under 35 min
• Likes are server stars or favourites; playlist likes stay on the device
• Merged: hold the heart (right-click it on Windows and Linux (beta)) to like the song on any source with a match (since V3.2.0)
• Liked Songs has a SoundCloud view; offline, it lists the downloads you've liked

Album, artist and playlist pages

Play, shuffle, like, download or pin any collection, or start a Mix or a radio from it. Artist pages separate albums from EPs and singles, list popular tracks and can show a short biography.

Also in Colours from the artwork · Artist biographies and photos · Mix

Folder browsing

Browse by folder instead of tags, then play or shuffle a whole folder.

• Not for Plex, YouTube Music, Green Music App or SoundCloud, or while offline.

Playlists

Create playlists and add or remove tracks; changes are saved on the source itself, whether that's Navidrome, Jellyfin, Plex, YouTube Music, Green Music App, SoundCloud or the device.

• SoundCloud: create (private), rename, describe, add, remove and delete; only playlists you own take new tracks
• Local playlists keep songs from every merged source, and a new playlist can be created on the library source you pick

• Changes depend on your account's permissions. Plex smart playlists can't be edited.

Also in Smart playlists · M3U import and export

Smart playlists

Build playlists from rules that update themselves as your library and listening change, such as genre is jazz and last played more than 30 days ago, in random order, limit 50.

• 11 fields, 9 operators, 7 sort keys

• Rules are checked against the first 2,000 songs of the current library.
• Stored on the device, not on the server.

M3U import and export

Export any album or playlist as an .m3u file, or import one into a new playlist with each line matched against your library.

• A line matches at 3 points or more
• 5 s length tolerance

• Import needs a connection, so it isn't available offline.

Built-in tag editor

Edit title, artist, album, album artist, genre, year and track number. Changes are written into files on the device, or to the server for Jellyfin, and you can pull in suggestions and cover art before saving.

• 7 fields
• Suggestions from MusicBrainz, and on Android from extensions

• Navidrome, Plex, YouTube Music, Green Music App and SoundCloud tracks can't be edited in Aurora.
• On the desktop, close the file in other apps before writing tags.

Also in Identify tracks with AcoustID and MusicBrainz · Jellyfin

Identify tracks with AcoustID and MusicBrainz

Identify an unnamed or mistagged file by its sound. Aurora fingerprints it on the device, asks AcoustID for matches, fetches recording details from MusicBrainz and a cover from the Cover Art Archive, and lets you review them before applying.

• First 120 s fingerprinted with Chromaprint
• Cover Art Archive fronts at 500 px

• Needs your own free AcoustID API key, under Settings → Integrations.
• Files on the device only.

Also in Built-in tag editor · LRCLIB, MusicBrainz and Wikipedia lookups

Duplicate finder

Scan the library for likely duplicate tracks and compare each copy's album, format and bitrate side by side, so you know which one to keep.

• Up to 5,000 tracks per scan
• ±4 s duration tolerance

• It finds duplicates; it doesn't delete anything.

Artist separators

Choose which characters or words split multi-artist credits, so a collaboration appears on each artist's page while the full credit stays on the track. File tags aren't changed.

• 7 separators by default
• Up to 24, each 1–32 characters

• Applies to the device library's artist index; servers keep their own artist split.

Artist biographies and photos

Artist pages can show a short biography, a photo, genre tags, country and active years, looked up from MusicBrainz and Wikipedia.

• On by default, under Integrations → Artist Information; artist names are sent to MusicBrainz and Wikipedia.

Also in LRCLIB, MusicBrainz and Wikipedia lookups · Album, artist and playlist pages

Missing covers

Songs and albums without artwork get covers from MusicBrainz and the Cover Art Archive, and Navidrome's generic placeholder is swapped for the real cover. Artwork you already have always wins.

• Cache up to 64 MB or 2,500 covers
• Images up to 8 MB and 12,000 px a side
• 3 lookups at a time

• On by default; artist, album and track names are sent to MusicBrainz and the Cover Art Archive.

Also in LRCLIB, MusicBrainz and Wikipedia lookups

Animated covers

Play looping video or animated covers on Now Playing and album pages from your own files: name them after the album in a folder you choose, or put cover.mp4 next to the tracks.

• 5 formats: mp4, m4v, webm, animated WebP and GIF
• Names like Artist - Album.mp4 or Album.webp

• Nothing plays until you pick a covers folder; Aurora doesn't download animated artwork.

Sonic radio and Auto DJ

Aurora analyses the sound of each track on the device (timbre, harmony, energy and tempo) and builds radio from tracks that sound alike, not from tags. Auto DJ chains a set chosen for similar sound, a compatible key and a close tempo, and the player shows each analysed track's key, Camelot code and BPM.

• 51 values per track, from a 30 s window; tempo 60–180 BPM
• Auto DJ: 40 tracks scored 45% sound, 35% key, 20% tempo, within ±16 BPM
• Server radio first where the source is better at it, such as Plex
• Autoplay radio keeps similar tracks coming when the queue ends
• Start radio from any song's menu, or an album or playlist's, for every source; without enough analysed tracks, the source's own radio plays

• Needs at least 2 analysed tracks; analysis on launch is off by default, and a large library takes a while.
• Autoplay radio is off by default.

Also in Mix · Queue, saved per account

Internet radio

Browse live stations from the open radio-browser directory (popular stations worldwide, genre shortcuts or a search), save favourites, or add any stream URL yourself.

• Top 80 popular stations; up to 80 results per search
• 15 genre shortcuts

• Under More and the Library Radio tile.
• Stations aren't scrobbled or added to history.

Also in Live streams and radio

Podcasts

Search podcasts or pick a topic, subscribe, and stream episodes straight from each show's RSS feed. No account is needed.

• 40 search results
• 10 topics

• Episodes stream only: no downloads and no saved resume position.

Save songs from YouTube Music

Search YouTube Music, then save songs as tagged audio files with cover art in Music/Aurora, where they join your local library. The audio is saved without re-encoding.

• Opus or M4A; Best, 128 or 64 kbps
• Named Artist - Title, up to 120 characters

• Opus needs Android 10 or newer; older versions save M4A.
• The files are ordinary files in Music/Aurora; manage them with your file manager.
• The source streams are lossy.

Also in YouTube Music · Files on the device

Stereo SACD ISO import

Import a local SACD .iso and its stereo tracks appear in your library as separate songs, played straight from the image without extracting it.

• Up to 256 images
• Tracks listed as DSD64: 2,822,400 Hz, 1-bit

• Experimental, tested with synthetic images.
• Stereo only; the image file must stay where it is.
• Android only.

Also in DSD playback

Offline downloads

Download tracks, albums and playlists from any connected source for offline listening, as lossless originals or at a smaller bitrate.

• 3 downloads at a time
• Lossless, 128, 192, 256 or 320 kbps
• Downloads are kept per source; offline, with two or more kinds, Home and Library switch between Library, YouTube Music, Green Music App and SoundCloud
• Offline Liked Songs, artists and downloaded playlists; singles without an album gather under Downloaded songs
• SoundCloud downloads save one AAC file; previews and protected tracks save their YouTube match
• YouTube downloads fetch in 1 MiB ranges, so YouTube doesn't slow them to playback speed

• On servers, lower bitrates are converted to MP3, so quality depends on the source.

Also in Offline cache, offline mode and streaming quality · Downloads and offline mode on the desktop

Offline cache, offline mode and streaming quality

Aurora keeps audio as you stream it, and fully cached tracks stay playable when the connection drops. An Offline Mode switch, separate Wi-Fi and mobile quality settings and a data saver come with it.

• Cache on by default: 1 GB, from 256 MB to 5 GB
• Oldest cache evicted first; downloads never are
• Lossless, 128, 192, 256 or 320 kbps, set separately for Wi-Fi and mobile data
• Data saver: about 96 kbps on mobile data
• Offline lists show downloads only; cached tracks still play, but don't appear in them

• Only fully cached tracks play offline.
• Quality settings apply to server sources; files on the device always play as they are.
• The desktop has one streaming quality and no data saver.

Also in Offline downloads · Downloads and offline mode on the desktop

Listening-history sync and Private Session

Aurora reports what you play to the source it came from: now playing, progress and a counted play for Navidrome, Jellyfin, Plex and YouTube Music, and your SoundCloud history for SoundCloud streams. One Private Session switch pauses all of it, along with Aurora's own history, detailed statistics and scrobbling.

• A play counts after 30 s, or half of a shorter song
• Progress every 30 s, and on pause, seek or speed change
• Merged libraries report to the source that actually played

• Offline plays aren't uploaded later.
• On Android, Private Session doesn't hide Discord presence, which has its own switch; on the desktop it does.
• Green Music App plays aren't reported to it.

Also in Listening statistics · Last.fm scrobbling · ListenBrainz scrobbling

Crossfade

Blend the end of one track into the next with a Smooth, Linear or Equal power curve. Overlap headroom keeps the join from getting louder than either track.

• 0–30 s in 1 s steps; 0 is off
• Next track prepared on a second player up to 15 s ahead
• Fade capped at half the track; ReplayGain applied per deck

• Not used for live streams.
• Processed USB output disables crossfade. Direct USB crossfades only local FLAC files with the same rate, channels and bit depth.
• Unavailable while casting or sending to DLNA or another Aurora.

Also in Mix · Gapless playback and two-deck crossfade on the desktop · USB DAC output: Direct and Processed

Gapless playback

Albums play back to back with no silence between tracks.

• The Gapless switch is saved but not read; playback is gapless either way.
• Gapless transitions aren't available while casting or sending to DLNA or another Aurora.

Also in Gapless playback and two-deck crossfade on the desktop

Shuffle and repeat that restore your order

Shuffle reorders the queue around the song you're on, and turning it off puts everything back in the original order. Repeat cycles Off, All and One.

• The current track stays first; the rest are shuffled
• Shuffle and the original order survive restarts and handoff to Cast, DLNA or another Aurora
• Notification and lock-screen buttons use the same commands

• Shuffle isn't available while a Mix plays.

Also in Queue, saved per account · Notification, lock-screen and headset controls

Queue, saved per account

Play next, add to the queue, drag to reorder, remove or clear what's coming up. Each account keeps its own queue and position, so it's still there after a restart or an account switch.

• Position saved about every 15 s and on pause
• Large collections load in 120-track batches
• Previously played above Now Playing; Save queue as playlist
• Previous restarts the track after 4 s

• Radio stations and podcasts are dropped when the queue is saved as a playlist.

Also in Swipe actions on tracks · Lyrics and queue panel on the desktop · Sonic radio and Auto DJ

Swipe actions on tracks

Swipe a song row to act on it without a menu: a short swipe right plays it next, a short swipe left queues it, and longer swipes like it or open playlists.

• Short: min(72 dp, 22% of the row)
• Long: min(180 dp, 55% of the row)
• Haptic tick on commit

• Editing a playlist needs a connection to that track's source and edit permission.

Also in Queue, saved per account

Player gestures and haptics

Swipe the artwork to change track, double-tap it to play or pause, and pull down to close player sheets. Each gesture has its own switch under Gestures & behaviour.

• Swipe, pull down and double-tap are on by default
• Haptics on play, skip and navigation, off by default

Also in Music Haptics

Full-art view

Tap the cover art in the player (click it on Windows and Linux (beta)) to let the artwork fill the screen; tap again to bring the controls back.

Music Haptics

Your phone taps along with the drums and bass of the music playing on it, at the strength of the phone's media vibration setting.

• Beat, Rhythm (default) or Detailed
• Intensity 20–100%, 80% by default
• Bass Texture adds a soft rumble under held bass notes
• Timing shift ±150 ms in 10 ms steps, for Bluetooth delay

• Only for music playing on this phone.
• Needs a vibration motor; Bass Texture needs one with amplitude control.
• Under Gestures & behaviour and in the player menu. Android only.

Also in Player gestures and haptics

Speed and pitch

Change speed from 0.5× to 2×. Pitch either follows the speed like a turntable or stays put, with its own shift of up to ±6 semitones.

• 0.5–2.0× in 0.05× steps; quick picks 0.5, 0.75, 1, 1.25, 1.5 and 2×
• Pitch ±6 semitones in 1-semitone steps
• Direct USB uses the USB engine's own varispeed, and is bit-perfect by design again at 1.0×

• Unavailable on Processed USB output and while casting or sending to DLNA or another Aurora.
• On the desktop, speed also changes pitch; there's no separate pitch control.

Also in Speed, sleep timer, ReplayGain and volume on the desktop

Skip silence and mono

Cut silent stretches inside tracks, or fold left and right into one channel for single-earbud or accessibility listening.

• Mono runs in the 64-bit path even with the EQ off

• Silence skipping isn't available on Processed USB or network outputs.
• The desktop has mono, but no silence skipping.

ReplayGain and loudness scanning

Even out the volume between songs with Track or Album ReplayGain, and measure local files without gain tags on the phone.

• Off, Track or Album
• Scanner: BS.1770 loudness to the ReplayGain 2.0 reference of −18 LUFS, gated at −70 LUFS
• Gains only turn the volume down, by up to 20 dB
• Scanned gains are cached; files aren't rewritten

• Not applied on Direct USB; Signal Path shows when levelling is active or bypassed.
• YouTube streams carry no ReplayGain data. Loudness scanning runs on Android.

Also in Speed, sleep timer, ReplayGain and volume on the desktop · Mix

Mix

Tap Mix on an album or playlist and Aurora plays it as one continuous mix in the original order: close tempos are matched, cues snap to the beat, joins are timed by key similarity and loudness is levelled.

• Tempo matched within ±6%
• 16-beat joins when keys are similar (chroma > 0.7), otherwise 8; bass exchange on beat-matched joins
• 3 s fade when tempos differ; 5 s before a track is analysed
• Silent intros and outros trimmed, up to 12 s; levelled by cutting up to 12 dB, never boosting

• Needs at least 2 tracks.
• On Android, not available with exclusive USB output or while casting; on the desktop a mix plays through the shared output and pauses the main player.
• Each track is analysed once, up to 3 minutes of it.

Also in Mix Studio · Crossfade

Mix Studio

Open any mix as a multitrack arrangement: change every entrance and exit, layer tracks, adjust tone, tempo and pitch per clip, and save mixes to reopen later.

• 8 overlapping layers; up to 10,000 tracks per mix
• Per clip: cue, fade, shape and bend 0.25–4, gain, pan, bass, mid and treble −24 to +6 dB
• Tempo 0.5–2×, pitch ±12 semitones; joins aligned to 4, 8, 16 bars
• 60 steps of undo

• Mix EQ and pitch send processed audio to the selected output, so they aren't bit-perfect.

Also in Mix · Vocal and backing separation

Vocal and backing separation

In Mix Studio, switch any track to just the vocals or just the instrumental. A neural network splits it on the device after a one-time model download.

• UVR-MDX-NET Voc_FT on ONNX Runtime
• 64 MB download, checked against SHA-256
• Tracks up to 30 minutes; stems cached per account and track

• Can take several minutes per track and needs free storage.
• Separated stems can carry artefacts.

Also in Mix Studio

Lyrics sources

Aurora asks your music server for lyrics first and falls back to LRCLIB. A time-synced version always wins over plain text.

• Order: server synced, LRCLIB synced, server plain, LRCLIB plain
• Navidrome and OpenSubsonic structured lyrics, Jellyfin per-word cues, Plex LRC
• Enhanced LRC word tags give true word timing
• With Green Music App merged, songs from every other source can use its synced lyrics: the song's own source first, then Green Music App, then LRCLIB
• Settings → Lyrics → Lyrics sources lets you reorder them
• Merge options can switch the borrowed lyrics off for each kind of source

• Lyrics embedded in local files and .lrc sidecar files aren't read yet; local tracks use LRCLIB.
• YouTube Music and SoundCloud bring no lyrics of their own; Green Music App brings its own synced lyrics.
• LRCLIB can be switched off; it receives title, artist, album and duration.

Also in Word-by-word lyrics · LRCLIB, MusicBrainz and Wikipedia lookups

Word-by-word lyrics

Synced lyrics light up word by word as they're sung. When a source only has line timestamps, Aurora estimates each word from its syllables and the song's own singing pace.

• Line by line, Word by word (default) or Follow vocals
• Syllables counted for Latin, Greek, Cyrillic, Han, kana and Hangul
• Pace 110–700 ms per syllable, smoothed over ±3 lines
• Lines stretch by at most 1.25×; the fill runs 80 ms ahead

• Estimated timing is approximate unless the source has word timestamps.

Also in Follow vocals · Synced lyrics with estimated word timing on the desktop · Lyrics sources

Follow vocals

Aurora listens to the track on the phone to find where the singing starts, pauses and stops, then fits each word to it, so breaths and new notes become the word breaks.

• 20 ms frames; 512-point FFT at about 11 kHz
• Centre-panned energy in a 180–3,800 Hz band; a 9-frame median drops drum hits
• Otsu threshold; below a separation of 1.6 it keeps the estimate
• Pauses of 350 ms or more favour a break; results cached for 300 songs

• Choose it under Settings → Lyrics; Word by word is the default.
• Only for line-synced lyrics without word timing.
• Tracks up to 20 minutes; DSF, DFF and SACD ISO are skipped. Streams are analysed on Wi-Fi only.
• Android only; the desktop estimates word timing.

Also in Word-by-word lyrics · Synced lyrics with estimated word timing on the desktop

Full-screen lyrics

Lyrics over a slowly drifting backdrop coloured from the cover. The current line stays sharp while the others recede, and tapping a line seeks there.

• Past lines at 35%, upcoming at 55%
• Blur grows 1.6 dp per line, up to 7 dp
• Back to the current line 4.5 s after you browse

• Plain lyrics show when there's no timing.

Also in Word-by-word lyrics · Tablet Now Playing panels · Synced lyrics with estimated word timing on the desktop

Lyric translation

A small translation into the app's language under each lyric line, made on the phone, so lyrics never go to a translation service.

• Google ML Kit language identification and translation
• About 30 MB per language pack, downloaded once on Wi-Fi
• Lines already in your language are left alone, so mixed-language songs translate only the foreign lines

• Translates into the app's language only.
• Android only.

Also in Word-by-word lyrics · 17 interface languages

Music videos for your own songs

Tap Video on a local or server track and Aurora finds the official YouTube video, then plays it muted in step with your own file, so you hear your copy through your settings.

• Chromaprint lines up 45 s of the song with 65 s of the video
• Re-syncs after a seek or more than 3 s of drift (1.5 s on the desktop); follows speed
• Auto, or 360p up to 4320p on Android and 2160p on the desktop; 720p by default
• YouTube Music tracks switch between audio and video at the same position
• Desktop: a full-screen video view; click the video to pause, double-click to leave
• On Android, Merge options can switch YouTube music videos off overall or for each kind of source; with videos off, a tap opens the Green Music App Canvas instead

• Uses unofficial YouTube stream extraction, and some songs have no match.
• Not offered for radio, podcasts or Mixes.

Also in YouTube Music

Live streams and radio

YouTube live streams and internet radio play as live audio: a LIVE badge replaces the seek bar, and on Android, Aurora rejoins the live edge if playback falls behind.

• HLS or DASH
• Up to 5 unplayable YouTube tracks skipped in a row

• Crossfade is skipped for live items.
• Radio and podcasts aren't scrobbled or added to history.
• YouTube streams are lossy.
• The desktop doesn't rejoin the live edge.

Also in Internet radio

Sleep timer

Stop the music after 5, 15, 30 or 45 minutes, an hour, or at the end of the track. The last 6 seconds fade out instead of cutting off.

• 6 s fade-out
• Works while a Mix plays

• The timed fade isn't available while casting or sending to DLNA or another Aurora; End of track still works.

Also in Speed, sleep timer, ReplayGain and volume on the desktop

Music alarm

Wake up to your liked songs, shuffled, fading in over 30 seconds, with a full-screen wake-up screen over the lock screen.

• One daily time, 07:00 by default
• Falls back to downloaded songs
• Exact system alarm, re-armed after reboots, updates and time-zone changes

• No snooze and no per-weekday schedule.
• Needs notification, lock-screen and battery permissions.
• Android only.

Android Auto

Browse Liked Songs, playlists, albums, artists and downloads on the car screen, or search your library, with full media controls.

• 5 top-level categories
• Sound presets can switch while connected, through a preset rule

Also in Preset rules

Voice requests

Ask your voice assistant to play a song, artist, album, genre or playlist on Aurora, or just to resume.

• Songs narrowed by artist and album
• Waits up to 15 s for the player to be ready

Home-screen widget and Quick Settings tile

A resizable widget shows the cover, title and artist with previous, play/pause and next. A Quick Settings tile toggles playback and shows what's on.

• 4 × 1 cells, resizable
• Tile subtitle Playing, Paused or Aurora (Android 10+)

Notification, lock-screen and headset controls

The media notification and lock-screen player carry Aurora's own shuffle and repeat buttons. Unplugging headphones pauses, and swiping the app away stops playback.

• Aurora sends no push notifications; recaps wait in the in-app bell inbox.

Also in Shuffle and repeat that restore your order · Recap inbox

Output picker and Independent Output

Choose where Aurora plays: speaker, wired, USB DAC, Bluetooth, hearing aid or dock. Independent Output lets other apps keep playing while Aurora plays on its own device, and calls don't pause it.

• 7 choices, including Automatic
• Applies to the current session; Signal Path shows the actual route

• Independent Output is off by default.

Also in Hi-res float output on Android · Signal Path

Chromecast

Send playback to a Chromecast as the original file (Direct) or with your Aurora effects applied first (Processed). The queue, position, shuffle and repeat carry over.

• Processed: rendered to 48 kHz 16-bit WAV
• Files served with a per-file token, to the receiver's address only, for 4 hours

• Tracks are prepared first, so starting can take a moment.
• No crossfade, gapless, speed changes or silence skipping while casting.

Also in DLNA and UPnP receivers · Play on another Aurora device

DLNA and UPnP receivers

Find DLNA and UPnP renderers on your network and play to them with volume control, as the original file or a processed version.

• SSDP discovery; AVTransport and RenderingControl
• Volume 0–100
• Format matched to what the renderer advertises; WAV for processed audio

• Compatibility depends on the receiver.
• No crossfade, gapless, speed changes or silence skipping.

Also in Chromecast

Play on another Aurora device

Turn a second phone or tablet into a paired Aurora player. It copies your queue and plays with its own sound settings and output, such as its own USB DAC, even if the controlling phone goes offline.

• 6-digit pairing code, valid for 2 minutes; 5 wrong tries close it
• TLS with a key held in Android Keystore; the certificate fingerprint is pinned
• Up to 8 controllers per receiver; copies up to 100 tracks or 512 MB before playing

• Turn on Allow Remote Control on the receiving device first.
• Speed changes and crossfade are unavailable.

Also in Chromecast · USB DAC output: Direct and Processed

Aurora Connect

Control another Aurora from this one: its queue, playback, volume, output and sound preset, while songs play from that device's own sources. Phones, tablets and computers can control each other on the same network, or over Tailscale by address.

• 6-digit pairing code, valid for 2 minutes
• Found automatically on the local network; Add by address for Tailscale or another network, after comparing the device key
• Port 47801 by default
• When switching devices: Keep both playing, Pause here (default) or Move playback

• Turn on Allow Remote Control on the device to be controlled: Settings → Network Audio on Android, Settings → Aurora Connect on the desktop, where it's off by default.
• Only song ids cross the network, never stream links, so a song plays only if that device finds it in its own sources.
• Both devices need a compatible Aurora version.

Also in Play on another Aurora device · Output picker and Independent Output

Tablet Now Playing panels

On a tablet, the player sits beside a lyrics or queue panel; drag the divider between them, or open the lyrics full screen.

• Split 35–65%, 50% by default

• Phones stay in portrait; landscape layouts are for tablets.

Also in Tablet layout · Full-screen lyrics

Format and source badges

Under the title, the player shows where the track plays from (Local, Downloaded or Streaming), a Lossless chip, and a line with codec, sample rate, bit depth and bitrate. Tapping it opens Signal Path.

• Format line: codec · kHz · bit depth · kbps
• LIVE or STREAMING replaces the seek bar for streams of unknown length

• Every .m4a file gets the Lossless chip, even when it holds AAC.
• The values come from metadata; Signal Path reports what actually reaches the output.

Also in Signal Path · Best-source playback

64-bit sound engine

All of Aurora's own processing runs on 64-bit floating-point samples, whatever the output. Changing a setting mid-song swaps the effect graph at a block boundary with a matched crossfade, instead of resetting the filters.

• binary64 samples, 53-bit significand
• Accepts 8–768 kHz
• Blocks of up to 8,192 frames
• Bypass copies the original bytes unchanged

• Turning a 32-bit float decoder output into 64-bit doesn't restore precision the decoder never had.
• Direct USB and native DSD go around the engine.

Also in Desktop sound engine · Processing rack

Tone engine and quick Custom DSP

Choose System effects (Android's own EQ, bass boost, virtualizer and loudness enhancer), Aurora's Custom DSP, or Off. Custom DSP gives quick controls in a fixed order: EQ, preamp, balance, width, crossfeed, tube saturation, channel delay and trim, compressor and limiter.

• Quick EQ: 12 parametric bands
• Preamp −12 to +12 dB, with a headroom readout and Auto preamp
• Limiter ceiling −6 to 0 dB, −0.3 dB by default
• Channel delay 0–20 ms; trim −12 to 0 dB

• System effects is the default, and how it sounds depends on the device.
• Any Custom DSP setting modifies samples, and Signal Path says so.
• The desktop offers Custom DSP or Off.

Also in Parametric EQ with 14 filter types · Graphic equaliser

Graphic equaliser

A fixed-band graphic EQ in 10, 15 or 31 bands, alongside the parametric bands, with one graph showing the combined result.

• 10 bands: 31 Hz–16 kHz, octaves
• 15 bands: 25 Hz–16 kHz, ⅔ octave
• 31 bands: 20 Hz–20 kHz, ⅓ octave
• Simple sliders or the Advanced draggable graph

Also in Parametric EQ with 14 filter types

Parametric EQ with 14 filter types

Each band uses one of 14 filter types, from peaks and shelves to Butterworth and Linkwitz–Riley crossovers or raw biquad coefficients. Drag points on the graph or double-tap to add one; each band can target both channels, left or right.

• Quick EQ 12 bands; each rack Equalizer stage 64; 256 across the rack; 512 biquad sections
• 10 Hz–24 kHz, ±30 dB, Q 0.01–100
• Butterworth orders 2, 4, 6 and 8 (12–48 dB/oct); Linkwitz–Riley 4 and 8
• Custom biquad: 5 normalised coefficients, checked for stability

Also in Processing rack · Response graphs · EQ import and export

Response graphs

Every EQ stage can show its calculated response at the decoder's current sample rate as magnitude, unwrapped phase or group delay; tap any frequency for its exact value.

• Log-frequency axis
• Includes the stage's wet/dry and bypass
• 48 kHz preview when the rate is unknown

• Group delay is the EQ's own, not device or playback latency.

Processing rack

Build a chain of stages that run top to bottom; reorder, duplicate, rename, bypass or blend any stage with the dry signal. Start empty, from your standard settings, or from a recommended order that ends with a limiter after convolution.

• 16 stages, 18 stage types, 256 bands
• Gain, Equalizer, Saturation, Stereo & trim, Crossfeed, Compressor, Limiter, Channel delay, Convolution, Channel utility, Dynamic equalizer, Multiband compressor, Adaptive loudness, Alignment delay, Gate/expander/de-esser, Tone, Delay & reverb, Modulation
• Wet/dry 0–100% and bypass per stage
• Saves automatically; an empty rack passes audio through

• Inactive until you choose Rack mode; Standard mode is the default.

Also in Parallel routing and mid/side · Saved subchains · EQ, rack and convolution on the desktop

Parallel routing and mid/side

Any stage can take inputs from the original signal or earlier stages (stereo, left, right, mid or side, M/S encoded or decoded), each with its own gain. Parallel branches are delay-aligned automatically.

• 1–4 inputs per stage
• Route gain −60 to +12 dB
• Feed-forward only, no loops
• Branch latency aligned up to 2 s

Also in Processing rack

Saved subchains

Save any group of stages as a named subchain, with the impulse responses it uses, and append it to any rack later.

• Up to 32 subchains

Also in Processing rack · Convolution and impulse responses

EQ import and export

Import parametric EQ from Equalizer APO text, AutoEq PEQ JSON or Aurora EQ JSON as new Gain and Equalizer stages, and export any EQ stage as APO text, or as Aurora EQ JSON when it uses filter types APO can't express.

• Files up to 256 KiB
• Export writes Preamp: 0 dB and warns when graphic EQ settings are left out

• Exports carry parametric bands only; processing presets carry the whole rack.

Also in Processing presets · Measured headphone and speaker presets

Channel utility and alignment delay

A 2 × 2 channel matrix with polarity inversion, a DC blocker, mono bass below a frequency you choose, and a sample-accurate alignment delay.

• Matrix gains −2 to +2 per path
• Mono bass off or 20–500 Hz
• Alignment delay 0–100 ms

Compressor, multiband and dynamic EQ

A single-band compressor and limiter, a 3-band multiband compressor with phase-compensated crossovers, and a dynamic EQ band that only acts when its detector band crosses a threshold.

• Threshold −80 to 0 dB; ratio 1:1–20:1; knee 0–24 dB
• Attack 0.1–200 ms; release 5–3,000 ms
• Crossovers 40–2,000 Hz and 400–16,000 Hz, at least an octave apart
• Peak or RMS detector, stereo-linked; solo, mute and listen to the detector

Gate, expander and de-esser

One dynamics stage with three modes: the expander makes quiet passages quieter, the gate closes below a threshold with a hold time, and the de-esser tames sibilance above a frequency you choose.

• De-esser 1–16 kHz
• Gate hold 0–500 ms

Tone and saturation

Tape and tube saturation, an exciter that adds harmonics to the highs, and a bass shelf. A separate saturation stage has selectable oversampling.

• Drive 0–24 dB; 30 Hz–16 kHz
• Tape, Tube and Exciter at 2× oversampling
• Saturation stage Off, 2×, 4× or 8×; 16× at most across the rack

• Bass is a low-shelf boost; the harmonics come from Tape, Tube and Exciter.

Delay, reverb and modulation

Stereo and ping-pong delay and a damped reverb with pre-delay, plus chorus, flanger, a 6-stage phaser, tremolo and vibrato with adjustable stereo phase.

• Delay and pre-delay 1–750 ms; feedback up to 0.65
• Reverb decay 0.1–4 s
• Modulation 0.05–10 Hz; stereo phase 0–180°
• Up to 4 space stages per rack
• The Reverb stage runs on the reverb's 8-line feedback delay network; pre-delay above 200 ms and decay below 0.2 s are clamped

Reverb

One reverb over everything you play, from a small room to a cathedral. Pick a preset and an amount in the player's speed and pitch sheet, or shape the space in its own settings screen.

• Room, Hall (default), Plate, Cathedral, Ambience or Lush, or your own
• Amount 0–100%, 25% by default
• Decay 0.2–12 s, pre-delay up to 200 ms, room size, early reflections and damping
• Low cut 20–1,000 Hz, high cut 1.5–20 kHz, stereo width up to 150%, modulation
• An 8-line feedback delay network; the rack's Reverb stage uses the same engine
• Concert mode (since V3.2.0): Club, Concert hall, Arena, Stadium, Festival or Amphitheatre; front row, middle, back or side; a sound per genre, audience from empty to packed, stage level, subwoofer layout, temperature and humidity

• Off by default.
• Works in every equalizer mode and with speed and pitch; skipped for bit-perfect USB output and direct casting.
• Signal Path lists it as a processing stage.

Also in Delay, reverb and modulation · Speed and pitch · Signal Path

Stereo width, balance and crossfeed

Narrow or widen the stereo image, shift balance and trim, and add crossfeed: a delayed, low-passed share of each channel blended into the other, for a more speaker-like image on headphones.

• Width 0–2×, 0 is mono
• Crossfeed 0–100%
• Balance −100% L to +100% R

Adaptive loudness

Adds bass and treble that scale with media volume, so music keeps its body when you listen quietly.

• Bass shelf at 100 Hz, up to +12 dB
• Treble shelf near 5 kHz, up to +6 dB
• Reference volume 5–100%

• It follows media volume; it isn't a sound-level estimate.

Also in Listening level estimates

Automatic headroom and limiter

Aurora estimates the rack's worst-case boost across more than 2,000 frequencies, convolution included, and lowers the input to make room. A sample-peak limiter at the end protects the output.

• Readout: Auto headroom x dB at y kHz
• Limiter ceiling −6 to 0 dB, −0.3 dB by default

• Saturation and dynamics stages aren't in the estimate.
• Sample-peak, not true-peak.

Convolution and impulse responses

Import WAV impulse responses (mono, stereo or true-stereo four-channel) into a library, then trim, delay, normalise or convert them to minimum phase. Several can run in one rack.

• Up to 32 in the library
• WAV up to 64 MiB, 1,048,576 frames a channel
• True stereo: LL, LR, RL and RR
• Peak-normalised to −1 dB; saved as 32-bit float

• Convolution adds latency and a tail; long responses may be resampled for each output rate.

Also in Saved subchains · Processing presets

Measurement projects

Import left and right frequency-response measurements, choose a target, and Aurora fits a peaking-filter correction within limits you set. It previews the measured, target, fitted and predicted curves, then appends the result to the rack.

• REW and Squig text, AutoEq JSON and CSV, Wavelet GraphicEQ
• Up to 16,384 points or 512 KiB
• 1–128 bands in all, up to 64 a channel; boost 0–24 dB, cut 0–30 dB
• 32 saved projects

• Magnitude only; phase isn't corrected.
• Peaking filters only.

Also in Target library · Live AutoEQ from squig.link

Target library

Pick a published target that downloads once and then works offline, import a custom one, or save any measurement as a target. Aurora warns when the target's measurement rig may not match yours.

• Harman in-ear 2019, Diffuse field 5128, JM-1 with Harman filters, Harman in-room 2013
• 32 saved targets
• Downloads checked against SHA-256

Also in Measurement projects

Measured headphone and speaker presets

Search a bundled offline index of measured corrections for headphones, in-ears, open earbuds and speakers. Applying one fetches its parametric EQ and preamp.

• 5,057 headphone, in-ear and earbud profiles
• 1,083 speaker corrections, Bluetooth speakers included
• Up to 12 filters plus preamp per profile

• The index is offline; applying a profile downloads its EQ file from GitHub.
• Speaker corrections don't account for your room.

Also in Live AutoEQ from squig.link · Presets bound to outputs

Live AutoEQ from squig.link

Search live IEM measurements on squig.link or Precog, choose Harman IE 2019 or Diffuse Field, and Aurora generates the correction on the device with the same biquads the engine runs, plus a preamp that keeps it from clipping.

• Up to 10 filters: a low shelf at 105 Hz (±12 dB), a high shelf at 10 kHz (±10 dB), then peaks
• Treble above about 9 kHz trusted less

• Needs a connection; search needs at least 2 characters.

Also in Measured headphone and speaker presets · EQ, rack and convolution on the desktop

Presets bound to outputs

Bind a processing preset or a device EQ profile to an output, and optionally to a named pair of headphones on it; it applies whenever that output appears. A USB DAC is recognised by vendor, product and serial.

• Speaker, earpiece, wired, Bluetooth, USB audio or HDMI
• Device identity stored as a SHA-256 hash
• Keep Current Sound holds your choice across reconnects

• Direct USB bypasses processing, so a DAC binding applies in Processed mode.
• On the desktop, a DAC is recognised by its output device, not its USB serial.

Also in Preset rules · USB DAC output: Direct and Processed

Preset rules

An ordered rule engine switches processing presets from what's playing and where: output, headphones, source, album, genre, playlist, sample rate, codec, container, or Android Auto and Cast. An Explain view shows why each rule matches.

• Up to 64 rules, 1–16 conditions each, All or Any
• 13 condition fields
• Priority −1,000 to 1,000

• The previous sound returns when no rule matches; rules pause during A/B tests.
• Android Auto and Cast conditions never match on the desktop.

Also in Presets bound to outputs · Processing presets

Processing presets

Save the whole sound setup (EQ, effects, the rack, loudness, ReplayGain, output and USB choices) as a preset, then duplicate, apply or export it as a ZIP that carries its impulse responses.

• Up to 100 presets
• ZIP: up to 8 impulse responses of 64 MB each, 512 MiB in all

• Applying may ask for an app restart when output or engine settings change.

Also in Preset rules · Level-matched A/B and blind ABX · Backup and restore

Level-matched A/B and blind ABX

Render the same passage through two saved presets, match their loudness, and switch A and B instantly, or run a blind ABX test that reports the chance probability.

• Up to 15 s; RMS matched within 0.1 dB
• About 20 ms crossfade between A and B
• 16 trials, one-sided binomial probability

• Compares saved processing presets only, and stops if playback, output or volume changes.

Also in Processing presets

Listening level estimates

Calibrate an output with headphone sensitivity, impedance, DAC voltage and gain, and Aurora estimates the left and right sound level with an explicit ± uncertainty, optionally keeping a history.

• Sensitivity in dB/V or dB/mW
• Uncertainty ±1 to ±40 dB
• 40 calibrations; history every 30 s, up to 2,880 samples

• An electrical estimate, not an acoustic measurement; uncalibrated outputs show none.
• On the desktop, calibrating needs playback on an exclusive output.

Signal Path

A live report of the whole chain, from source through decoder, processing, resampling and output to the device and its latency. It ends in a verdict with reasons: Samples preserved, Samples modified or Sample preservation unknown.

• 7 stages; peak and RMS in 100 ms windows, before and after processing
• Per-stage peak and gain change; underrun count
• USB: completed and pending frames, packet errors, timeouts, clock readback
• Copyable report without URLs, account or device IDs, or titles

• Not a true-peak, loudness or acoustic measurement.
• Sample preservation unknown means that path can't report the detail.

Also in Signal Path on the desktop · Format and source badges · USB DAC output: Direct and Processed

Output sample rate, resampling and dither

Choose how processed output is clocked: follow the source, a fixed rate, or the highest compatible rate, optionally keeping the 44.1 or 48 kHz family. A windowed-sinc resampler converts when needed, and integer output can be dithered.

• 44.1, 48, 88.2, 96, 176.4, 192, 352.8 and 384 kHz
• Dither Off, TPDF or noise-shaped (first order; TPDF below 44.1 kHz)
• Resampler delay shown in Signal Path

• Android: applies to hi-res output and Processed USB once the output reopens. Desktop: applies in exclusive mode; shared mode follows the system mixer.
• Noise shaping moves conversion noise; it doesn't add detail.

Also in Hi-res float output on Android · USB DAC output: Direct and Processed · WASAPI exclusive output on the desktop

Hi-res float output on Android

Prefer hi-res Android output switches the pipeline to 32-bit float when the device supports it, with an FFmpeg decoder for full-depth decoding, and lists the rates the device reports.

• 32-bit float output when supported
• Preferred device: Automatic, the phone speaker or a connected device

• Needs an app restart.
• Android owns the clock on this path and may still resample or mix; the DAC's format isn't verified.

Also in Android 14 bit-perfect mixer request · Output picker and Independent Output

Android 14 bit-perfect mixer request

When Aurora's own USB driver isn't playing, Android 14 and newer can be asked to use its bit-perfect USB mixer at the track's exact format. Aurora asks when hi-res output is preferred, or when exclusive USB falls back to Android.

• Audio Output → Android Capabilities lists bit-perfect, standard or no mixer modes, and the device's rates

• Android 14 or newer.

Also in Hi-res float output on Android · USB DAC output: Direct and Processed

USB DAC output: Direct and Processed

A built-in user-space USB Audio driver sends audio straight to a USB DAC, around Android's mixer. Direct mode skips effects and software volume, so it's bit-perfect by design; Processed mode keeps your processing and volume.

• USB Audio Class 1 and 2, asynchronous feedback
• PCM 16, 24 or 32-bit, mono or stereo, up to 384 kHz
• DAC clock read back and checked; packets up to 1,024 bytes
• If the DAC goes away: pause (default) or fall back to Android output

• Restart Aurora after changing USB settings.
• Signal Path reads Sample preservation unknown for Direct USB PCM: “USB transport conversion and downstream DAC behavior are not independently verified”.
• Processed USB disables crossfade, speed and pitch, and silence skipping. Direct USB crossfades only local FLAC files with matching rate, channels and bit depth.
• Compatibility varies by phone and DAC; the FiiO KA13 is the tested reference. Android may ask for USB permission again on reconnect.

Also in DAC hardware volume · DSD playback · Signal Path

DAC hardware volume

On Android in Direct USB mode, the phone's volume keys set the DAC's own volume through its USB feature unit, and turning the DAC's knob moves the phone's volume to match. In exclusive mode on Windows and Linux (beta), Aurora's volume sets the DAC's own volume when it has one, so the samples stay untouched either way.

• On by default: with Direct USB on Android, in exclusive mode on the desktop
• Android: volume 0 mutes the DAC; first contact settles on the quieter of the two levels
• Desktop: the DAC's top 60 dB, never above 0 dB; Signal Path shows the level

• Needs a DAC with a writable volume control. Without one, the desktop falls back to software volume, which leaves samples untouched only at 100%.
• On Android, Direct USB only.

Also in USB DAC output: Direct and Processed · WASAPI exclusive output on the desktop · ALSA exclusive output on the desktop

DSD playback

Plays mono and stereo DSF and DFF from DSD64 to DSD1024. By default Aurora converts DSD to 176.4 kHz PCM so your processing still applies; in Direct USB mode it can send the untouched DSD as DoP or native DSD.

• DSD64–DSD1024, 2.82–45.16 MHz
• PCM at 176.4 kHz through a 512-tap FIR at DSD64, up to 8,192 taps at DSD1024
• FiiO KA13 verified: native DSD64–256, DoP DSD64–128
• DST-compressed DFF decoded for DSD64

• Signal Path reads Samples preserved for DoP and native: “Source DSD bits retained through USB packing”.
• Raw DSD stops if its USB route goes away. No multichannel DSD, PCM-to-DSD or DSD over the network.
• On the desktop, DSF and DFF play as PCM.

Also in Experimental DSD for other DACs · Stereo SACD ISO import · USB DAC output: Direct and Processed

Experimental DSD for other DACs

An off-by-default switch that allows raw DSD (DoP or native) on DACs other than the tested reference, at rates up to DSD1024.

• Untested on real hardware above the KA13's verified rates.
• Disconnect headphones until the DAC confirms DSD mode, and start at a low volume.

Also in DSD playback

A visualizer on every output

The visualizer reads audio from inside Aurora's pipeline, so it works on every output, including bit-perfect USB and exclusive desktop output, without asking for the microphone.

Also in Visualizer · Full-window visualizer on the desktop

Listening statistics

Browse any day, week, month, year or all time: plays, minutes, artists, unique songs, albums and active days, with top artists, songs and albums ranked by plays or by minutes.

• A play counts after 30 s, or half of a shorter song
• Weeks start on Monday
• Up to 250,000 plays kept

• Private sessions aren't counted.
• Older entries without measured time use the track length, and the screen says so.

Also in Listening-history sync and Private Session · Recap inbox

Listening clock, calendars and comparison

Each recap shows your top artist's share of the time, a 24-hour listening clock, daily or monthly bars, and how many minutes more or fewer you listened than the period before.

• 24 hourly buckets
• 12 monthly bars for a year

Recap inbox

A bell on Home opens your finished recaps: the latest is featured, new ones wait under Ready to open, and opened ones move to Your archive.

• Daily, weekly, monthly and yearly
• Unread badge up to 99

• In the app only: recaps never post a system notification.

Also in Recap pictures

Recap pictures

Save any recap as a picture with your top five artists and songs, total minutes, plays, tracks and active days, in the Midnight, Paper or Aurora style.

• 1080 × 1600 px PNG

• Built from Aurora's own history; no account needed.

Also in Recap inbox

Listening history

Every play in a timeline grouped by day (Today, Yesterday, then weekday and date) with its time; tap a row to play the song again.

• The map pin on a play needs the listening map.

Also in Listening map

Detailed statistics

An opt-in mode that records more about each play (how long you listened, whether you finished or skipped, genre, source and shuffle) and, if you choose, the output device and audio path. A three-step setup starts it, and the data never leaves the device.

• Nothing recorded until setup finishes
• Kept out of cloud backups and Aurora backups
• Radio, podcasts and mixes aren't recorded

• Private Session pauses recording.

Also in When-you-listen heatmap · Places: Home, Work and Away · Listening map

When-you-listen heatmap

A weekday-by-hour heatmap of listening time. Select a block, such as Monday to Friday 08:00–09:59, to see its time, plays, average session, top artist, genres, places and outputs.

• 7 × 24 cells
• Sessions split after a 30-minute gap
• Top 3 genres per slice

• Needs Detailed statistics, and covers plays recorded after setup.

Also in Detailed statistics

Habits and badges

See what share of tracks you finished, skipped, shuffled or played in order, how many were repeats within a week, and your most common genre switch. Badges such as Night listener or Explorer unlock at set thresholds.

• 8 badges, after at least 10 counted plays

• Genre figures depend on the genre tags in your files or on the server.

Also in Detailed statistics

Output and audio-path shares

Listening time split by output (speaker, wired, Bluetooth, USB, HDMI or Cast) and by audio path: lossless, hi-res, AutoEQ, convolution, crossfeed, ReplayGain and bit-perfect shares, plus a format breakdown.

• The desktop guesses the output type from the device name.

Also in Detailed statistics · Signal Path

Places: Home, Work and Away

Optionally label each play Home, Work or Away. Plays keep only the label, and Insights shows minutes, top genre, top artist and peak hour per place.

• Radius 100, 250 or 500 m; 250 m by default
• Label refreshed at most every 10 minutes

• Needs location permission; on Android, plays started in the background need Allow all the time.
• On the desktop it uses the Windows location service or, on Linux (beta), the desktop portal.

Also in Listening map · Detailed statistics

Listening map

A separate opt-in that saves the exact spot of each play and shows it on an OpenStreetMap map; plays within about 30 m share a dot, and tapping one lists what you played there.

• A fix about every 2 minutes while playing, every 8 when you stay put
• Dots about 30 m across

• Off by default and separate from Places; exact coordinates stay on the device until you delete them.
• Viewing the map fetches tiles from OpenStreetMap's servers, which see the area you look at.

Also in Places: Home, Work and Away · Local first, no tracking

Theme modes and accents

System, Light, Dark or AMOLED black, with an accent from 10 presets, a custom hue, saturation and brightness picker, or Material You wallpaper colours. The whole app recolours live.

• 4 modes; Dark by default
• Rose, Coral, Magenta, Violet, Amber, Blue, Teal, Green, Sky and Mono

• Material You needs Android 12; the desktop offers the system accent instead.
• Accents apply to the Aurora style; the other styles bring their own palettes.

Also in Four interface styles · Colours from the artwork

Colours from the artwork

In the Aurora style, the player, queue and player sheets take their colours from the current cover; album, artist and playlist pages tint to their artwork, and the progress bar blends between colours.

• Player gradient intensity 0–150%
• Desktop: Follow the playing cover tints the whole app

• Retro hi-fi, Aero and Liquid Glass keep their own palettes.
• Follow the playing cover is off by default.

Also in Theme modes and accents · Album, artist and playlist pages

Four interface styles

Aurora, Retro hi-fi, Aero and Liquid Glass are complete visual identities: each redraws controls, surfaces, type, corners and backdrop, with its own colours. Your Aurora customisations are kept when you switch.

• 4 styles, each with a matching launcher icon on Android

• On the desktop, Glass is a static glass-style palette.

Also in Liquid Glass live refraction · Launcher icons

Liquid Glass live refraction

In the Liquid Glass style, the floating tab bar, mini player and other chrome bend, blur and tint whatever scrolls beneath them, and dragging across the tab bar moves a lens.

• Live glass on Android 12 (API 31) and newer; highlight shader on Android 13
• Frosted fallback below Android 12

• Android only; the desktop Glass style is a static palette.

Also in Four interface styles

Typefaces, text size and corners

Pick DM Sans, Plus Jakarta Sans, Manrope or the style's own face, scale text from 85% to 130%, and choose Sharp, Default, Rounded or Pill corners.

• 3 bundled typefaces
• Text 85–130%
• 4 corner styles

• Corner styles apply to the Aurora style, and some components keep fixed radii.

Player and Home customisation

Tune the Now Playing screen, the mini player, the library grid and which Home sections appear.

• Waveform seek bar 24–96 bars, 60 by default
• Artwork 60–100%, 86% by default; gradient intensity 0–150%
• 8 Home sections

• On the desktop, the waveform seek bar is decorative; it isn't drawn from the audio.

Also in Home feeds

Visualizer

31 real-time styles, from spectrum bars and radial waves to Strange attractor, Cymatics, Wormhole and Ink bloom, with control over colour, backdrop, bar count, frequency range, FFT size and frame rate.

• 16–160 bars, 64 by default
• FFT of 1,024 to 4,096 points; 30, 60, 90 or 120 fps
• Low cut 10–500 Hz, high cut 2–22 kHz
• Colour from the accent, a custom colour, a gradient or the album art

• Opens full screen from the player menu.
• It reads Aurora's own pipeline, so it needs no microphone permission.

Also in A visualizer on every output · Full-window visualizer on the desktop

Tablet layout

On large screens the bottom tabs become a sidebar, Settings and Library open in two panes, and Now Playing sits beside the content in landscape. A Tablet Layout section sizes the dock, panels, spacing, margins and cards.

• From a smallest side of 600 dp
• Player split 35–65%
• Dock 70–115%; page margins 0–120 dp

• Phones stay in portrait; tablets rotate freely.

Also in Tablet Now Playing panels

Navigation editor and Simple Mode

Choose which destinations sit in the main navigation, under More, or nowhere, and reorder them. Simple Mode hides the advanced audio destinations.

• 16 items, 3 of them fixed
• Main, More or Not shown
• Phone tab bar: 2 to 5 tabs, Home always one of them

• The custom phone tab bar is experimental.

Launcher icons

Pick a home-screen icon that matches a style: Aurora, Retro hi-fi, Aero or Liquid glass.

• 4 icons

• Some launchers drop the old icon when it changes, so you may need to add Aurora to the home screen again.
• Android only.

Also in Four interface styles

Local profile

Give your local library a display name, an avatar and a banner; the profile page shows them with your playlists and top artists.

• Name up to 80 characters
• Avatar 512 px, banner 1,280 px; source images up to 20 MB

• Server accounts show their server identity.

Guided first-run setup

Setup asks how you listen (Everyday or Advanced), what you use (USB DAC, IEMs, AutoEQ), where your music is and which extras you want, connects each source, and leaves a checklist that opens the matching settings.

• 4 questions; 6 source types
• Up to 9 setup tasks

• Android only; the desktop app has no first-run setup.

Also in Navigation editor and Simple Mode

17 interface languages

Aurora is fully translated into 17 languages; pick one in the app or with Android's per-app language setting. Song titles and artist names stay as they are.

• About 2,820 strings per language
• English, Русский, Türkçe, Español, Français, 简体中文, Português, हिन्दी, اردو, Deutsch, Italiano, Nederlands, Polski, Bahasa Indonesia, 日本語, 한국어, العربية
• The desktop app switches language without a restart

• Arabic and Urdu are translated, but layouts don't mirror right to left.

Also in Lyric translation

Downloads & Storage

One page that measures everything Aurora keeps on the device (downloads, caches, analysis files, models and history), with a clear button for each part and one for all caches at once.

• Download bitrate: Lossless, 128, 192, 256 or 320 kbps
• Cover cache capped at 64 MB or 2,500 covers

• On Android, vocal separation files in use are kept when Clear all caches runs.
• On Windows and Linux (beta), the full page arrived in V3.1.0.

Also in Offline cache, offline mode and streaming quality · Vocal and backing separation

Backup and restore

Export one aurora-backup.zip with your settings, racks and presets, the impulse responses they use, measurement projects, local playlists, likes, play history, profile and listening-level calibrations, and restore it on any device.

• Up to 512 MiB
• Up to 100,000 history entries

• Sign-ins are left out unless you switch on Include sign-ins, which saves login tokens and API keys in plain text; keep that file private. Restoring a backup without sign-ins keeps the accounts already on the device.
• Music files and downloads aren't included.

Also in Processing presets · Local profile

Permissions overview

One screen lists each optional permission with what it unlocks and a Grant button. Aurora works without any of them.

• Android can't grant USB access permanently, so the prompt can return on each reconnect.

Native desktop app

Aurora runs as a full desktop app on 64-bit Windows 10 and 11 and on x86_64 Linux (beta), with the same library, server and sound code as Android, in a layout made for mouse, keyboard and wide windows.

• Windows 10/11 x64; Linux beta x86_64 with glibc 2.35+
• Java 21 runtime bundled
• Window 960 × 600 at least; size, position and maximised state remembered

Also in Installers and packages

Local folders

Add folders from any drive; Aurora scans them for tags, embedded or folder cover art and ReplayGain, with no account, and you can browse them by folder.

• 21 file types, including wv, ape, mpc, tta and dsf
• Manual refresh with progress

• Local files get lyrics from LRCLIB, or from Green Music App when it's merged.
• DSF and DFF are decoded to PCM; there's no DSD output on the desktop.

Also in Files on the device

Navidrome, Jellyfin and Plex

Sign in to Navidrome or another Subsonic server, to Jellyfin, or to Plex with a token, and switch between saved accounts.

• Streaming quality: Lossless, 128, 192, 256 or 320 kbps

• YouTube Music, Green Music App and SoundCloud also work on the desktop, through a sign-in browser that Aurora downloads once, about 150 MB. Extension sources aren't supported.

Also in Navidrome, Subsonic and OpenSubsonic · Jellyfin · Plex Media Server

Merged library and best source

Merge local folders and every included server into one library; each track plays from the best copy in your order: a local file, a download, then a stream.

Also in Merged library · Best-source playback

Downloads and offline mode

Download tracks, albums and playlists from any connected source at the quality you choose, and switch to offline mode when you travel.

• Lossless, 128, 192, 256 or 320 kbps
• Streaming cache on by default: 1 GB, from 256 MB to 5 GB

• Only fully cached tracks play offline.

Also in Offline downloads · Offline cache, offline mode and streaming quality

Library and navigation

A sidebar with pinned items and playlists, grids that fill the window, side-by-side list and detail views, right-click menus, and M3U import and export.

• Split library from 960 dp; list pane 320–560 dp, remembered
• Two-pane settings from 840 dp
• Sidebar expands from 1,200 dp

Also in M3U import and export · Library tabs, likes and pins

Player bar

A player bar along the bottom with transport, like, shuffle and repeat, a seek bar, lyrics and queue toggles, an output-device button and a volume slider.

• Volume follows position³, closer to how loudness is heard
• Volume saved after 300 ms without changes

• There's no separate mini-player window.

Lyrics and queue panel

Open lyrics, the queue or related songs in a resizable panel beside whatever you're browsing; drag to reorder the queue or save it as a playlist.

• 360 dp wide by default, 300–560 dp
• Related appears with a YouTube Music account, active or merged

Also in Queue, saved per account · Synced lyrics with estimated word timing

Full-screen player

Now Playing slides up over the app with large artwork beside the lyrics or the queue; drag the divider to resize the halves, or use the full-screen button or F11 for true full screen.

• Artwork 60–100%; gradient 0–150%
• Optional row for speed, lyrics and queue
• Video button for the music video or the Green Music App Canvas

Also in Keyboard shortcuts

Synced lyrics with estimated word timing

Timed lyrics scroll with the song, and each word fills as it's sung, timed by estimates from the line timestamps. Click a line to jump there, or scroll freely and snap back.

• Line scroll 650 ms
• Inactive lines at 96% scale, blurred

• No Follow vocals and no lyric translation on the desktop.
• Local files get lyrics from LRCLIB, or from Green Music App when it's merged.

Also in Word-by-word lyrics · Follow vocals

Keyboard shortcuts

Control playback and move around without the mouse; Ctrl+/ shows every shortcut.

• Space: play or pause; Ctrl+← and Ctrl+→: previous and next
• Shift+← and Shift+→: seek 5 s; Ctrl+↑ and Ctrl+↓: volume; Ctrl+Shift+↓: mute
• Ctrl+S: shuffle; Ctrl+R: repeat; Ctrl+L: like; Ctrl+Q: queue
• Ctrl+F or Ctrl+K: search; Ctrl+,: settings; F11: full screen; Esc or Alt+←: back

• Seek, volume, mute, shuffle, repeat, settings, Ctrl+K and the Ctrl+/ list arrived in V3.1.0.

Windows media overlay and media keys

Aurora appears in the Windows media flyout with title, artist, album and cover art; media keys, seeking, shuffle and repeat work from outside the app.

• Timeline sent on change, every 5 s, or past 1.5 s of drift

MPRIS media controls

Desktop media widgets and media keys control Aurora through MPRIS, with track details, cover art, seeking, shuffle and loop.

• Volume and rate are reported but fixed at 1.0.

Tray and close to tray

A tray icon shows the current song and offers play/pause, next, previous, show and close; Close to tray keeps the music going after you close the window.

• Close to tray is off by default.
• On GNOME, the tray icon needs the AppIndicator extension.

Title bar and accent

Aurora draws its own title bar on Windows and Linux (beta), and a System title bar switch brings back your desktop's frame. The accent can follow the system's, live, or the cover of the song that's playing.

• Aurora's own title bar since V3.1.0
• Accent from the presets, a custom colour or the system
• In the Aurora style, hovering a cover tints the accent until the pointer leaves

• System title bar is under Settings → Appearance and applies after a restart; with it on Windows, Windows 11 takes Aurora's colours and Windows 10 shows the light or dark title bar.
• Following the playing cover applies to the Aurora style and is off by default.

Also in Colours from the artwork · Theme modes and accents

Keep-awake, one instance, data folders and logs

Aurora keeps the computer awake while music plays, runs as a single instance, keeps its data in the standard per-user folders and writes rotating logs for bug reports.

• Windows: installs to %LOCALAPPDATA%\Aurora Music
• Windows: %APPDATA%\Aurora and %LOCALAPPDATA%\Aurora
• Linux beta: $XDG_CONFIG_HOME/Aurora and $XDG_DATA_HOME/Aurora
• aurora.log, rotated at 2 MB, 3 kept

Protected server passwords

Saved credentials are encrypted: with your Windows account (DPAPI) on Windows, and on Linux (beta) with a key kept in GNOME Keyring, KWallet or another Secret Service provider.

• Linux beta vault: AES-256-GCM
• Also covers the YouTube Music, Green Music App and SoundCloud sign-ins

• Without a keyring on Linux (beta), passwords last until Aurora closes.

Also in Saved accounts and quick switching

Desktop sound engine

Every track is decoded by FFmpeg to 64-bit floating-point stereo and passed through the same processing chain as on Android; local files and streams take the same path.

• 16 stages, 64 bands per EQ stage, 256 in all

Also in 64-bit sound engine

Gapless playback and two-deck crossfade

The next track opens ahead of time so albums play without gaps, and an optional crossfade blends tracks with the curve you choose.

• Next track opened 20 s before the end, or crossfade length + 15 s
• Crossfade 0–30 s; Smooth, Linear or Equal power

• During a crossfade the output isn't bit-perfect, and Signal Path says so.

Also in Crossfade · Gapless playback

WASAPI exclusive output

Aurora takes sole control of the DAC and sends each track at its own sample rate and bit depth, around the Windows mixer. Signal Path reads Samples preserved when nothing processes the audio: volume at 100% (or any level when the DAC's own volume is in use), no ReplayGain, no crossfade, speed 1.0×.

• 44.1–384 kHz probed
• 24-in-32, 32, 24 or 16-bit integer, in that order

• If the device refuses, Aurora falls back to shared mode and says why.
• Its reason line: “Driver and DAC behavior beyond the exclusive WASAPI stream are not verified”.

Also in Signal Path on the desktop · Output sample rate, resampling and dither · DAC hardware volume

ALSA exclusive output

Shared mode plays through PipeWire or PulseAudio. Exclusive mode opens the sound card directly through ALSA, bit-perfect under the same conditions as WASAPI, and shows which formats the card accepts.

• 44.1–384 kHz; 16, 24, 24-in-32 or 32-bit
• Shared mode: 32-bit float at 48 kHz

• Pick the card itself, not the PipeWire or PulseAudio entry, and set the card's profile to Off if the sound server holds it.
• Its reason line: “Driver and DAC behavior beyond the exclusive ALSA stream are not verified”.

Also in Signal Path on the desktop · DAC hardware volume

Output device and hot-plug

Choose a specific output or follow the system default as it changes; Aurora recovers when a device is unplugged or replaced.

• Buffer 50–1,600 ms, 400 ms by default

Signal Path on the desktop

A live, stage-by-stage report from file to device that ends in Samples preserved, Samples modified or Sample preservation unknown, with the reasons.

• Meters on the latest 100 ms; refreshed every 500 ms

• What the driver and hardware do after the system stream is reported as not verified.

Also in Signal Path · WASAPI exclusive output · ALSA exclusive output

EQ, rack and convolution on the desktop

The same tone controls as on Android: graphic and parametric EQ, the processing rack, convolution with your own impulse responses, and live AutoEQ from squig.link.

• 16 stages, 64 bands per EQ stage, 256 in all, 512 biquad sections

• The bundled index of measured headphone and speaker presets isn't on the desktop; live AutoEQ from squig.link is.

Also in Processing rack · Parametric EQ with 14 filter types · Live AutoEQ from squig.link

Speed, sleep timer, ReplayGain and volume

Change speed, set a sleep timer that fades out, level tracks with ReplayGain, and use a volume control scaled the way loudness is heard.

• 0.5–2.0× in 0.05 steps
• 6 s sleep fade-out
• ReplayGain only lowers the volume, by up to 20 dB

• Speed changes pitch; there's no separate pitch control.
• No silence skipping on the desktop.

Also in Speed and pitch · ReplayGain and loudness scanning · Sleep timer

Full-window visualizer

A full-window visualizer with 31 styles; it works in exclusive bit-perfect mode too, because it taps the audio inside Aurora's engine.

• 64 bars by default

Also in Visualizer · A visualizer on every output

Installers and packages

An installer or an MSI on Windows; a .deb, .rpm or AppImage on Linux (beta). Java is bundled, so nothing else needs installing.

• Aurora-3.2.0.exe: 204.5 MB
• Aurora-3.2.0.msi: 203.6 MB
• aurora-deb.zip (aurora_3.2.0_amd64.deb): 187.0 MB beta
• aurora-rpm.zip (aurora-3.2.0-1.x86_64.rpm): 189.8 MB beta
• aurora-appimage.zip (Aurora-3.2.0-x86_64.AppImage): 199.8 MB beta

• The Windows installers aren't code-signed, so SmartScreen may ask before the first launch. The MSI replaces an older MSI install.
• Windows updates itself from Settings → About with the new MSI; on Linux (beta), install the new package.

Also in In-app updates · Native desktop app

Last.fm scrobbling

Send now-playing updates and scrobbles to Last.fm with your own API key and shared secret. Artist replacement rules can rename a credit before it's sent.

• A play qualifies after about 30 s
• Replacement rules match one exact credit, ignoring case, and stay if you disconnect

• Needs your own free Last.fm API account.
• Scrobbles that fail offline aren't retried later.

Also in ListenBrainz scrobbling · Listening-history sync and Private Session

ListenBrainz scrobbling

Submit playing-now updates and listens to ListenBrainz with your user token; it has its own switch.

• Listens that fail offline aren't sent later.
• Artist replacement rules apply to Last.fm only.

Also in Last.fm scrobbling · Listening-history sync and Private Session

Discord presence

Show what you're playing as a Discord Listening activity named after Aurora, the artist or the song, with optional album details.

• Text trimmed to 128 characters
• Desktop: through the Discord app on the same computer, with no sign-in; it reconnects when Discord restarts; without the app running, you can sign in under Without the Discord app

• On Windows and Linux (beta) it's on by default: whenever the Discord app is running, it shows what you play. Switch it off under Settings → Integrations → Discord; Private Session hides it there too.
• On Android it's unofficial and starts once you sign in: Aurora joins the Discord gateway with your account token, as your own presence, not a bot. Private Session doesn't hide it; use its own switch.
• Covers from public hosts such as YouTube show as they are; covers from your files or a private server need your own Imgur client ID and are then published on Imgur. On Android, cover art also needs your own Discord application ID.

Also in Listening-history sync and Private Session

LRCLIB, MusicBrainz and Wikipedia lookups

Optional online helpers: synced lyrics from LRCLIB when your server has none, covers from MusicBrainz and the Cover Art Archive, and artist biographies and photos from MusicBrainz and Wikipedia. Each has its own switch.

• LRCLIB receives title, artist, album and duration
• MusicBrainz receives artist, album and track names

• All three are on by default.

Also in Lyrics sources · Missing covers · Artist biographies and photos

Extensions

Install separate extension apps that add sound presets, read-only music libraries or tag suggestions. Nothing runs until you switch each capability on, and you decide when to load a preset, switch library or apply tags.

• Up to 32 extensions discovered
• Presets: 1–16 stages, up to 64 EQ bands a stage
• Tag suggestions: up to 4 providers, 20 each
• Access granted per capability and pinned to each extension's signing certificate

• Under Settings → Advanced audio → Extensions.
• Android only.

Also in Extension libraries · Extension SDK, API 1

Extension SDK, API 1

Build Aurora extensions as ordinary Android apps with the Apache-2.0 SDK and example app in the repository; they talk to Aurora over Android Messenger with a documented JSON contract.

• Messages up to 196,608 bytes
• 2 calls at a time, 5 s each
• 20 s to load a library; 8 MiB cached catalogue

• Built from the repository; not published to Maven.
• API 1 can't load native or arbitrary audio processors, only Aurora's own stages.

Also in Extensions

In-app updates

On Android, Aurora checks GitHub for a newer release, downloads the APK with progress, verifies it and its signing key, and hands it to Android's installer. On Windows it downloads the new MSI from Settings → About, checks it and runs it, and Aurora closes while it installs.

• Checks at most once every 60 minutes, when Settings opens (Settings → About on the desktop)
• Android: 15 s to connect, 30 s per request

• The check can't be switched off.
• Windows checks the installer's size, and its SHA-256 when GitHub lists one.
• On Linux (beta), Aurora says a release is out and links to it; install the new package yourself.

Also in Installers and packages on the desktop

Local first, no tracking

There's no Aurora account and no analytics or crash-reporting SDK. History, recaps and statistics are computed on the device, and detailed statistics, places, location history, and YouTube Music, Green Music App and SoundCloud sign-ins are kept out of cloud and Aurora backups.

• Cover, lyrics and artist lookups are on by default; opening Settings on Android, or Settings → About on the desktop, checks GitHub for updates.
• Viewing the listening map fetches OpenStreetMap tiles.
• Aurora backups leave sign-ins out unless you switch on Include sign-ins. Android's own device backup is on and includes app settings with saved server sign-ins and keys; on Android 12 and newer its cloud copy is made only when it can be encrypted end to end.
• On the desktop, Discord presence is on by default and shows what you play whenever the Discord app is running.

Also in LRCLIB, MusicBrainz and Wikipedia lookups · Backup and restore · Detailed statistics