Skip to content

Releases

Lull 1.13.5 — search works again

Choose a tag to compare

@PortableDiag PortableDiag released this 01 Nov 03:42
  • Search works again. Tapping the search icon closed the field the moment it opened, so nothing could be typed. Opening the field reported an empty query, which rebuilt the menu, which threw the open field away. The menu is now built once and updated in place.
  • Switching tab mid-search keeps the search, and the keyboard no longer closes on every letter.

Lull 1.13 — all seventeen house palettes

Choose a tag to compare

@PortableDiag PortableDiag released this 27 Oct 19:08
  • All seventeen house palettes. The theme button now opens a swatch list of every one — Material You (wallpaper colours), Ocean, Terminal, Trellis, Graphite, Blueprint, Futuristic, Neon, SynthWave, White Glow, Silkscreen, Sticky Notes, Amber Terminal, Phosphor Green, Phosphor P31, Zoro and Nebula — where before there was only light, dark and follow-system.
  • Material You, Ocean and Terminal follow light / dark / follow-system, behind the Light / dark button in the picker; every other palette is a single look and forces the night mode its own background calls for, so the status bar, dialogs and pickers come out the right way round for it rather than for the phone. Material You takes the wallpaper's colours on Android 12+ and is Lull's own blue below that; it is the default, as it is across the house set, and the mode still defaults to dark.
  • Amber Terminal, Phosphor Green, Phosphor P31 and Terminal set the type in monospace, emphasised text included.
  • A change shows at once on any screen that is open, including one changed from the API; a screen that was out of sight redraws when it comes back.
  • API: palette on GET/PATCH /api/settings, by key (white-glow) or by name in any case ("White Glow"); an unknown one is a 400 that lists the real ones.
  • Internal: the palettes are generated from ProjectForge's palettes.toml by scripts/palettes.py into theme overlays and a Palette enum with no Android imports, so the router validates against it and the tests run it. A fix found on the emulator: when AppCompat moves the night mode inside super.onCreate it rebuilds the theme from the manifest's, which kept a terminal palette's font but threw away its colours — the overlay is now applied again after it. 183 tests to 192.

Lull 1.12 — delete a track from the device

Choose a tag to compare

@PortableDiag PortableDiag released this 16 Sep 18:17

Delete a track from the device. The last entry on the selection bar — long-press a track, an album, a folder — and a button on the Now Playing top bar, because the moment you decide a track has to go is usually the moment you are listening to it.

  • Lull asks before the system does. On Android 11 and up the platform puts its own confirmation over the app and does the deleting itself, and Lull's dialog still comes first: the system's asks whether you meant to delete a file, while Lull's says which track, how many, and that the rating and the playlist entries go too. On Android 9 and below there is no system dialog at all, and Lull's is the only thing between a long-press and a file being gone for good.
  • All three of Android's answers to deleting are covered — MediaStore.createDeleteRequest for the whole batch on API 30+, a per-file RecoverableSecurityException consent on API 29, and a plain WRITE_EXTERNAL_STORAGE delete below that (which is why that permission is back, capped at maxSdkVersion="28").
  • Nothing trusts the result code. RESULT_OK says the user pressed a button, not that every file went. What actually happened is read back out of MediaStore afterwards, so the toast can say "Deleted 3 of 5" rather than claiming five.
  • Everything keyed to the track goes with it — the rating, the measured BPM, the loudness profile and its place in every playlist, each in one write. The deleted tracks come out of the playing queue too, highest position first.
  • The API deliberately cannot do this, and GET /api/docs now says so: from Android 10 up the system has to ask a person, and this is the only irreversible thing in the app.
  • 171 unit tests to 183.

Full notes in CHANGELOG.md.

Lull 1.11 — push through the quiet parts

Choose a tag to compare

@PortableDiag PortableDiag released this 15 Sep 02:29

The other half of the dead-spot problem. 1.10's Force Tempo sets a speed per track, and that is genuinely all it can do about a quiet bridge: the piano break in the middle of a 140 BPM song is also at 140 BPM, so there is no tempo there to correct. What is wrong with it is that the energy drops out — a different measurement, and a different fix.

Push through quiet parts finds those stretches and speeds up further while you are in one, then eases back out. Under Force tempo in the overflow menu:

  • How much faster in a quiet part — +10% to +150%, default +50%. It goes much further than the whole-track ceiling on purpose: this applies to the passage you specifically want to be over, where stretch artefacts matter far less than the seconds do.
  • How quiet counts as a dead spot — 3 to 24 dB below the track, default 9.

Relative to the track, never absolute. A quietly mastered song is not one long dead spot and a loud one is not free of them, so "full level" is a high percentile of that track's own seconds — not its peak, because one cymbal crash would raise the bar for the whole song and turn ordinary verses into dead spots.

Three guards against speeding up where it would be vandalism:

  • A stretch must last 6 seconds. The bar of held breath before a drop is the opposite of a dead spot, and speeding through it would wreck the one moment the track was building to.
  • Two quiet stretches split by a short fill are one dead spot, not two.
  • Every second of a reported section meets the threshold on its own.

It ramps, and it ramps early. ~800ms, starting 800ms before the quiet part — so it has finished arriving by the time you get there, rather than playing the first and most skippable seconds at the old speed.

One decode, two answers. Where a track goes quiet needs the whole file, unlike its tempo. TrackAnalysis decodes end to end once and produces both — once per track, on a background worker, only while the feature is on, and a track ahead of the one playing. What is stored is the loudness curve, not the conclusion, so moving the threshold slider re-reads every track already measured instead of invalidating a library, and the panel shows live what the current setting finds in what is playing.

API. GET /api/tracks/{id}/quiet lists what was found and distinguishes never analysed from analysed, nothing there; pushQuiet, quietBoost, quietThresholdDb on /api/settings; boosted on /api/player, whose speed is now read off the engine rather than recomputed — it moves during a track now.

Tests: 138 → 171. One is a regression caught on a device: a first cut widened the threshold on the way out of a section, in decibels, which let a run continue through seconds that did not qualify at all — a 4-second dip swallowed the remaining 160 seconds of the track. Gap-merging already gave the anti-flap property that was meant to buy, so the hysteresis is gone.

Verified on device: a generated track with a 30-second quiet section at 1:00 is found at exactly 60000–90000ms at every threshold, the speed ramps 1.181 → 1.772 across the boundary and back, a 4-second dip is correctly ignored at every threshold, and the same holds on the signed release APK.

See CHANGELOG.md for the full entry.

Lull 1.10 — Force Tempo

Choose a tag to compare

@PortableDiag PortableDiag released this 15 Sep 01:50

Set a BPM and every track plays at it. Under Force tempo in the overflow menu: pick a target anywhere from 60 to 200 and each track is sped up as far as it needs to reach it, pitch preserved — a 128 BPM track at 1.09x is still in the key it was recorded in.

It exists for one complaint: a workout playlist where one slow track is a slow patch in the workout. The defaults follow from that.

  • Never slow a track down (on by default) — anything already at or above the target is left completely alone.
  • A ceiling on the stretch (30% by default) — roughly where a stretch stops being unnoticeable. A track that can't reach the target inside the ceiling goes as far as the ceiling allows, not further.

Where a tempo comes from. The file's TBPM tag if it has one; otherwise Lull measures it — half a minute decoded from 30 seconds in, reduced to an onset envelope and autocorrelated. It's remembered per track, and the next track in the queue is analysed while the current one plays, so the only track ever heard at the wrong speed is the one that was already playing when you switched the feature on.

A track with no tempo is never touched. An audiobook or a drone is analysed, found to have no steady beat, and written down as having none. "Nothing known yet" and "nothing there" are two different states — a player that invented a tempo for a podcast and then played it 30% fast would be indefensible.

The measured number is editable, because it's a measurement rather than a preference and you can hear when it's wrong. Type over it in the panel and it sticks; clear it and Lull has another go.

Octaves. A tempo is only defined up to a factor of two, and that's handled twice over: a 70 BPM track counted in double time is 140 and is left alone, while a merely slow 95 BPM track is not excused as a secretly-fast one. The estimator searches lags at quarter-frame resolution because a 160 BPM beat lands every 37.5 envelope frames — a whole-frame search reports 80 for it, confidently.

API. GET/PUT /api/tracks/{id}/bpm; forceTempo, tempoBpm, tempoMaxChange, tempoOnlyFaster on /api/settings; speed and source_bpm on /api/player. GET /api/docs covers all of it.

Tests: 83 → 138. Verified on a device against generated click tracks at known tempos — 100, 128, 160 and a TBPM-tagged 95 all read within 1 BPM, pink noise correctly found to have none — with crossfade and A-B both still correct underneath a forced track.

See CHANGELOG.md for the full entry.

Lull 1.9

Choose a tag to compare

@PortableDiag PortableDiag released this 13 Sep 18:58

Share, and an API that can drive the whole app.

Share button

Send the playing track from the Now Playing top bar, or a whole selection from the library's contextual bar — long-press a track, an album, a folder.

The file itself travels, with its metadata alongside. Title, artist, album, genre, duration, size and your own rating go into the share text, because a content:// uri carries none of that and what arrives at the other end is otherwise a filename. Nothing is re-encoded and no tag is rewritten: the file that leaves is byte-for-byte the file on disk.

A local API

Turn it on under Local API in the overflow menu. Lull generates a key; a script or agent holding it can do what you can — browse the library, play, queue, seek, skip, set repeat and shuffle (including Favourites), rate, build and reorder playlists, change settings, arm the sleep timer, set an A–B region, and share.

KEY=<copied from the app>          # adb forward tcp:8787 tcp:8787, from a laptop
curl -s -H "X-API-Key: $KEY" 127.0.0.1:8787/api/docs
curl -s -H "X-API-Key: $KEY" -X POST 127.0.0.1:8787/api/player/play -d '{"group":"album:42"}'
curl -s -H "X-API-Key: $KEY" 127.0.0.1:8787/api/tracks/42/file -OJ    # the whole track
  • Off until you switch it on, bound to 127.0.0.1 unless you also turn on LAN access. The key is generated, never typed, compared in constant time, and can be reissued at any moment.
  • Its own foreground service, separate from playback — because "start playing something" is the request an agent most wants to make, and that is exactly when the playback service does not exist. The notification names the address, and carries a Stop.
  • POST /api/share pops the real chooser and says whether anyone can see it: Android discards a dialog thrown by an app that is not on screen, so the share text and the file urls come back either way.

The honest note

This version adds INTERNET to the manifest, because a listening socket needs it even to bind loopback. Lull still makes no outbound connection of any kind and has no analytics, ads or telemetry — but the README no longer claims "no network access".

Internals

Routing, auth, request parsing and every error shape are Android-free behind an ApiBackend interface and tested against a fake: 5 tests to 83, including one that walks the route list /api/docs advertises and fails if any of it is unrouted.

Lull 1.8 — star ratings and Favourites shuffle

Choose a tag to compare

@PortableDiag PortableDiag released this 11 Sep 16:28

Star ratings, and a shuffle that leans on them.

  • Rate a track 1–5 stars from the star row on Now Playing — tapping the star it already sits on clears the rating. Or long-press in the library and rate a whole selection at once.
  • Rated tracks show a compact star strip on their row, and get a seventh browse tab — Rated, five stars down to one. There is deliberately no "unrated" bucket: on a real library it would hold nearly everything and simply be a second copy of the Tracks tab.
  • Favourites shuffle is the new third state of the shuffle button (off → shuffle → favourites), and the point of the ratings: 3 stars and up come around early and often.
Rating Relative chance of landing early
★★★★★ 40
★★★★ 24
★★★ 12
unrated 3
★★ 2
★ 1

The step between 2 and 3 stars is the design: three stars is the lowest rating that still means "yes", so that is where the weight jumps rather than creeps. An explicit 1 star is rarer than no rating at all — a low rating is a judgement, no rating is just silence.

It is a reordering, not a filter: every track is still in the queue, so Next reaches all of it eventually. Tapping a specific track still plays that track first and weights the rest behind it. Switching it on mid-queue re-weights only what has not played yet and leaves the playing track alone.

Ratings are stored as MediaStore ids in SharedPreferences, like playlists — nothing is written to your audio files, and a rating survives a file moving.

Installs over 1.7.1; your shuffle setting carries across.

Full notes in CHANGELOG.md.

Lull 1.7.1

Choose a tag to compare

@PortableDiag PortableDiag released this 07 Sep 17:42

1.7.1 — 2026-09-07

  • Audiobooks are back in the library. 1.7 narrowed the MediaStore query to IS_MUSIC (plus podcasts), and an audiobook has IS_MUSIC = 0 — so on a device that is mostly spoken word most of the library disappeared. The query now excludes only what nobody wants in a player: ringtones, alarms and notification sounds. Everything else — music, podcasts, audiobooks, recordings — is listed, as it was before 1.7.

Lull 1.7

Choose a tag to compare

@PortableDiag PortableDiag released this 07 Sep 17:25

1.7 — 2026-09-07

  • Open from a file manager. Lull now appears in "Open with" for audio — audio/* plus the application/ogg/flac types some providers still send, over content:// and file://, with a matching extension filter for senders that give a bare uri and no type at all. It also accepts a share (SEND / SEND_MULTIPLE), so a selection can go straight to the player.
  • Opening one track offers its whole folder, so Next pages through it. Three routes in order: the ClipData the launching app attached (Sift does this — no permission needed, and the only route that works for a .nomedia folder), then the folder resolved out of Lull's own library, then the single file. Resolving the folder handles a file manager's private FileProvider uri by reading the real path off the open descriptor (/proc/self/fd/N), which answers neither a MediaStore id nor a DATA column.
  • A launch from another app never waits on media permission. Without it there is nothing to load, so the intent's own read grant carries the playback; with it, the library load is awaited first so the folder can be built.
  • Folder, Artist, Album and Genre views, alongside the flat track list and playlists — six tabs, each a list you drill into and Back out of. The tab and the folder/artist/album/genre/playlist you were on are both remembered.
  • Genre comes from MediaStore's genre membership tables rather than the GENRE column, which only exists from API 30; it is one query per genre, once per load, off the main thread.
  • Multi-selection. Long-press any row to start it, tap to add more. Play the selection, add it to the queue, add it to a playlist, remove it from the one you are in, or select all. It works on groups too — long-press an album or a folder and everything in it comes with it. Selection is held by track id, so it survives a search keystroke or a playlist edit.
  • A real playlist manager. Playlists are their own tab, with a per-row menu — play, rename, duplicate, delete — and a button to make a new one. Adding a selection is one write for the whole batch rather than one per track.
  • The selection bar now overlays the toolbar instead of stacking above it, which used to push the whole list down the moment a selection started.

Lull 1.6

Choose a tag to compare

@PortableDiag PortableDiag released this 04 Aug 01:28

Sleep timer

Play for 5–90 minutes, fade out over the last 30 seconds, then pause — the other half of falling asleep to a looping track.

  • Set it from the moon button on Now Playing (the title bar counts it down, Sleep in 24:31) or from Sleep timer in the library's overflow menu, where the entry shows the time left.
  • The countdown runs in PlaybackService, so it survives closing the app, and it is measured against elapsedRealtime so it counts through device doze rather than stopping with the CPU.
  • The fade is a raised cosine, which is flat at both ends: it eases in without an audible step and settles onto silence instead of arriving at it still dropping. On a short timer it is capped at half the total.
  • It ends on pause, not stop, so the queue and your place in it survive — one tap to carry on.
  • A running timer is deliberately not persisted; only the duration you last chose is.

Under the hood

Crossfade and the sleep fade both want the engine volume 25 times a second, so they now contribute independent gains that are multiplied in one place rather than overwriting each other. A crossfade that wouldn't finish before the deadline is skipped outright, so the track you hear last isn't one you never chose to end on.


APK is signed (CN=Lull). minSdk 26 · targetSdk 35 · versionCode 7.