Blueforcer/awtrix-ng v1.1.0

on GitHub

Added

• MP3s. Boards with a connected I2S DAC play your own MP3 files: drop them into the new Audio tab, then play one by name - sound:"ding" in a notification, from a script, or over the API.
• ESP32-S3 boards with quad PSRAM (N8R2, N16R2, N4R2) have their own image. The browser flasher recognises which PSRAM a board has and picks for you; the wrong image is refused instead of installed.
• Every sound output has its own volume. Buzzer, DFPlayer, stored MP3s and the radio are four separate sliders now, each 0-100.
• A DFPlayer no longer costs you the buzzer. The two are separate outputs and both stay live, so a board with an added MP3 module keeps its melodies, its melody editor and its soundRtttl in notifications.
• Two new keys on POST /api/v1/audio/play. sound takes a name and lets AWTRIX pick the output - a stored MP3, else a melody, else a DFPlayer track when the name is a plain number - which is the rule a notification's sound has always followed. track plays a DFPlayer track by number. The explicit keys - mp3, melody, track, rtttl - each name one output and never fall back to another.
• POST /api/v1/audio/stop takes an optional {"scope":"sounds"|"stream"|"all"}, so a script can silence its own chime without killing the radio someone is listening to.
• Scripts can ask whether the device is still making a sound, with sound.playing(), which output it has, with sound.sinks(), and name one directly with sound.mp3(), sound.melody() and sound.track().
• A converter for AWTRIX 3 flows on the documentation site turns an old configuration into the AWTRIX NG equivalent.

Changed

• The audio API has changed. Radio and sounds shared one speaker but had two addresses; everything now lives under /api/v1/audio, and MQTT under cmd/audio/*. Anything that drives sound from outside - Home Assistant, Node-RED, your own scripts - needs adjusting: HTTP API · MQTT.
• Sounds and Radio are one Audio tab: MP3s, Melodies, Radio.
• volume is gone, replaced by buzzerVolume, dfplayerVolume, mp3Volume and radioVolume, all on a 0-100 scale instead of 0-30. radioVolume keeps its name, its scale and your stored value; the old volume is dropped, so the buzzer, DFPlayer and MP3 volumes start at their defaults once - which are the old ones expressed in the new scale, so nothing gets quieter. PATCH {"volume":10} now answers 422: anything driving AWTRIX from Node-RED or Home Assistant needs the new key.
• A backup taken before 1.1.0 loses its settings on restore. Unknown keys are refused outright, so the whole settings category is skipped with a warning while icons, melodies, MP3s, palettes and scripts restore normally. Check your volumes afterwards.
• melody on a DFPlayer board no longer means a track number. It means a stored melody, as it does everywhere else, and {"track":7} is how you address the SD card. {"sound":"7"} still works and now prefers a /MP3/7.mp3 if you have uploaded one.
• soundEnabled no longer mutes the radio. It covers one-shot sounds - melodies, MP3s, tracks, a notification's own melody. A stream is something you started deliberately; use {"scope":"stream"} on /audio/stop to end it. sound.stop() in a script follows the same rule.
• GET /api/v1/capabilities reports one audio object - {"buzzer":…,"track":…,"mp3":…,"radio":…} - in place of the four loose radio/audio/melodies/sound flags, two of which could never disagree.
• A notification whose soundRtttl does not parse is now rejected with 422 instead of being accepted and then silently playing nothing.
• A melody or MP3 you ask for by name on a panel that has no such output answers 503 rather than 422: the request was fine, the hardware is not there.
• Scripts may be twice as long: scriptMaxBytes now defaults to 16 KB instead of 8 KB, still adjustable up to 32 KB.
• The USB install images come as a single usb-awtrix-ng.zip, so the update files are what you see first on this page.
• Both ESP32-S3 images now say which PSRAM they are for: firmware-awtrix-ng-s3-octal.bin for N8R8/N16R8 boards, firmware-awtrix-ng-s3-quad.bin for the quad ones. The plain -s3 name is gone.

Fixed

• The LookingEyes effect drew small square eyes instead of the full-size ones AWTRIX 3 has. They are back to size, look around properly and blink again.
• Uploading a script reserved memory for the largest script allowed rather than the one being sent, so a save could be refused on a busy device.
• An MP3 whose file name cannot be played back - spaces, brackets, accents, over 32 characters - is refused at upload instead of sitting there unplayable.
• Internet radio threw away pieces of every stream. A station that runs ahead of real time keeps the buffer at its limit, and getting back under it meant discarding audio the decoder had not seen yet - two or three gaps a second, for as long as the station played. Nothing is discarded now.
• A sound played over the radio stuttered as it ended. The output was left running unfed while the stream reconnected, so the tail of the sound repeated until the music came back.

Which file do I need?