Releases: cha0t1cnu3tral/pubsplash
Release list
v0.2.0
Additions
-
Added Spanish localization.
-
Added AAC support.
-
Added a per-source compressor.
-
Added a per-source noise suppressor.
Fixes
- Fixed bug with third-party SAPI voices not showing.
Changes
What's Changed
Full Changelog: v0.1.9...v0.2.0
Contributors
Assets 8
v0.1.9
Additions
- Added a Discord community link to the README.
Fixes
- Automatic updates now request administrator permission before running the installer. Users updating a per-machine installation from an earlier version must download this release's installer and run it as administrator manually.
Changes
- Updated the README's source-routing instructions.
Full Changelog: v0.1.8...v0.1.9
v0.1.8
Additions
-
Every source edit dialog now has an Effects tab, holding built-in effects that run on that source alone, before any effect on a bus it is sent to. The list has Add (ALT+A), Edit, Remove, Move up and Move down; DELETE removes the selected effect and CTRL with the up and down arrows reorders it, as on the Scenes and Sources lists. Changes on this tab take effect as they are made, so an effect can be set up by ear against live audio.
-
A new Auto-ducker effect turns another source down while a chosen one is making noise — music that gets out of the way when you talk. It is set up in percentages of full volume and times in milliseconds: which source to listen to, the ducked volume, the level that starts the ducking, the fade down and fade back up times, and how long it stays ducked.
Fixes
-
Google Translate TTS voices now populate the voice picker.
-
Render-device tests no longer interfere when they run in parallel.
Changes
-
Source settings are now one dialog per source with the kind's own settings on the first tab, replacing the five separate dialogs. Microphone sources have a real settings window for the first time.
-
Editing a source opens with the keyboard on the first control of its settings, rather than on the tab bar.
-
The Application source's list of running applications is now the settings tab itself, instead of a separate picker window opened from a button.
What's Changed
- Fix Google Translate TTS voice discovery by @Wasilewsk in #10
New Contributors
- @Wasilewsk made their first contribution in #10
Full Changelog: v0.1.7...v0.1.8
Contributors
Assets 8
v0.1.7
Additions
-
Streams can be scheduled to go live at a set time, from the new File > Schedule stream.... Simple mode connects at one time on whichever scene is active; Advanced connects early on a pre-stream scene and switches to a second scene when your content actually begins.
-
While a stream is scheduled, Home on the status list reads "Stream scheduled" with a spoken countdown, Start streaming becomes Cancel scheduled stream, and recording is unavailable. Once it connects, Home behaves exactly as it does for a stream you started yourself.
-
A schedule lasts for the current session only and is never saved, so Pubsplash can never start broadcasting at launch; closing the window asks before discarding one. If the computer was asleep when the time came round, the schedule is cancelled rather than going live more than two minutes late, and update questions now stand down while one is pending.
-
Pubsplash's own playback device can now be chosen, on the new Audio tab in Preferences. Monitoring, text-to-speech and sound cues all follow it and listeners are unaffected; Play a test sound checks the device without starting a stream, and a device that is unavailable is reported in the log rather than quietly swapped.
-
A Desktop Audio source can now capture one output device instead of all of them, through a new edit dialog. The device Pubsplash itself plays out of cannot be picked, since capturing it would send our own speech and cues to listeners, and pointing Pubsplash's output at a captured device moves that source back to capturing every device with Pubsplash excluded.
-
Preview sounds on the Sound packs tab of Preferences allows playback of every sound event for the chosen pack. Arrow to the event, hit SPACE to play or use the play button.
-
Sound packs can hold Opus as well as WAV, mixed freely in one pack, and existing packs keep working untouched. The Sound Pack Manager's Browse now accepts
.wavand.opus, and the new Encode sounds as Opus (96 kbps) checkbox converts WAV sources as they are saved, while an already-Opus source is copied rather than put through a second lossy pass. -
An Audiopub service's Icecast server and port can now be set, on the same two fields a direct Icecast service uses. They arrive filled in with the host Pubsplash used to derive from the site address, so existing setups are unaffected, and Connect now checks that the server can be found.
-
Pubsplash now reads MP3, FLAC, Ogg Vorbis and Ogg Opus wherever it reads audio, not just WAV. This covers what a speech service sends back as well as what a sound pack is built from.
Fixes
-
Pubsplash no longer says "Streaming" before Audio Pub has accepted the stream. Status now reads "Streaming (waiting for the server to accept the stream)" until the server confirms it, and "Streaming (the server has lost the source)" if the server later drops the source, with the spoken announcement following the same rule.
-
A stream still waiting after three quarters of a minute explains itself in the log, and the wait is dropped if the connection carrying the server's answer goes down. A direct Icecast mount has no such check and is unaffected.
-
The log now records how long each stage of starting a stream took, so a slow start can be told apart from a failing one.
-
The portable build is now genuinely portable, keeping settings, logs, crash dumps and caches in a
user_datafolder besidepubsplash.exeinstead of%LOCALAPPDATA%\pubsplash. Updates replace the program files and leaveuser_dataalone. -
An existing portable copy copies its settings across the first time it starts, minus logs and crash dumps, and leaves the old folder alone so an installed copy on the same machine keeps working. Saved passwords and API keys are encrypted for the Windows account that entered them, so they do not survive a move to a different machine or account.
-
The portable build's default recording folder is now
recordingsinsideuser_data, so recordings stay with the copy that made them. An installed copy still defaults to the Music library, and a folder you chose yourself is left alone in both. -
An Icecast server entered as
host:portnow works, instead of becominghost:8000:8000and failing with "unknown host (os error 11001)". A whole listen URL pasted into the field is understood too, and its parts are moved into the fields they belong in. -
Connecting to an Icecast service now checks that the address exists and reports it on the spot, rather than saying "Connected" and failing at the next Start streaming. The mount point is not touched by the check.
-
The Connect button now acts on the service you have highlighted, instead of disconnecting whichever service happened to be connected.
-
Disconnecting now says so out loud, the way connecting always has. The only sign it had worked was a button label change, which no screen reader reads out.
-
The service list keeps its "(connected)" marker up to date, instead of writing it once when the dialog opened.
-
Disconnecting from a service while a stream is live now ends the stream in Pubsplash too, rather than leaving it on "Streaming" and still encoding.
-
A stream that fails to start now says why in the log, not only in the modal that is gone as soon as it is dismissed.
-
Direct Icecast services now accept
/as the mount point, instead of refusing it and sending the request out asPUT //. -
STAR voices now populate the voice dropdown from the configured coagulator.
-
STAR now defaults to the coagulator's standard port 7774.
Changes
-
Sends is now one list of every destination. Master output and every bus are checkable rows toggled with SPACE, each announcing whether it is on; the separate Send directly to master checkbox and the Add and Remove send buttons are gone.
-
Each bus you check gets its own send level slider beneath the list, which disappears when you uncheck the bus and keeps its level if you switch it back on during the same visit.
-
Changes in the Sends dialog take effect immediately, so it closes with a single Close button. A source left sending to nowhere is pointed out as the dialog closes.
-
Editing a source's sends no longer restarts every capture source in the app, so it cannot glitch a live stream.
-
A release's page on GitHub now leads with that version's changelog entries, followed by the commit-level list it showed before. The same page is what "an update is available" links to.
-
Rewrote the README as a beginner-friendly setup and usage guide.
-
Removed the ALT mnemonics from every control in the app, so nothing has an underlined letter and no ALT+letter shortcut fires. TAB and arrow navigation is unchanged, as are the menu accelerators (
CTRL+,for Preferences,ALT+F4to exit). -
Test in the Sound Pack Manager now plays the file itself rather than handing it to Windows' built-in WAV player, so previewing an Opus file works and nothing briefly opens a background PowerShell.
-
Only one copy of Pubsplash can run at a time; a second one explains that Pubsplash is already running and closes, instead of fighting the first for the microphone, the settings file and the stream. An installed copy and a portable copy count as the same app, but other people signed in to the same computer are unaffected.
-
A fresh settings file no longer writes the default recording folder out as a fixed path, so the default follows a portable copy from one machine to the next. Preferences shows the folder that is actually in use either way.
What's Changed
- Keep the portable build's data in its own folder by @ogomez92 in #5
- Fix Icecast host parsing and the streaming-service manager by @ogomez92 in #6
- Configure Audiopub Icecast endpoints by @mad-gamer26 in #8
New Contributors
- @ogomez92 made their first contribution in #5
- @mad-gamer26 made their first contribution in #8
Full Changelog: v0.1.6...v0.1.7
