Fontelle 0.28.0-beta.1

A beta: offered only to machines with Settings > Updates > Beta updates
turned on. Everything in 0.27.0, plus better audio on Windows, for testing
before it comes to everyone.

Windows users reported clicky, crackly playback. Settings > Audio output >
Backend now offers, besides the shared WASAPI it had before:

• ASIO®, through your audio interface's own ASIO driver. It plays by way of
  fontelle-asio.exe, a separate GPLv3 program that comes in the Windows
  download (its source is attached to the release).
• WASAPI (low latency): shared with other programs, at the smallest period
  the driver allows instead of Windows's 10 ms.
• WASAPI (exclusive): Fontelle alone on the device, in its own format.

And on every backend, Settings > Audio output says how many dropouts there
were and whose they were (the device's, or a buffer Fontelle was late
with), and how busy the audio thread is; the transport bar marks a lost
buffer until you reset it.

Testing it: pick a backend, play a busy song, and tell us what Playing
through, Dropouts and Processing load said, with your interface and buffer
size. If ASIO will not open, Fontelle says why and plays through WASAPI.

---

The Windows download can play through ASIO® drivers by way of fontelle-asio.exe, a separate GPLv3 program whose source is attached below as fontelle-asio-source-0.28.0-beta.1.tar.gz. ASIO is a registered trademark of Steinberg Media Technologies GmbH.

Full Changelog: v0.26.3...v0.28.0-beta.1

Fontelle 0.27.2

(0.27.0 and 0.27.1 were tagged but never published: the checks before release caught problems first, all fixed here. The Linux download runs on Ubuntu 22.04, Debian 12, Mint 21 and newer, and singers work there too.)

Fontelle sings. A new vocal track holds notes with words on them, and the
Fontelle Singer, built in and free, sings them with classic UTAU voicebanks
or community DiffSinger voicebanks (the OpenUtau layout).

• Type lyrics straight onto notes, in English, Japanese, Chinese or Korean
  (with your input method); a line of words spreads a syllable per note.
• Draw pitch on the grid with the pitch pen, with Smooth, Snap to scale and
  Reset, and shape breath, voicing, tension and gender on their own lanes.
• Find singers installs voicebanks from inside Fontelle, with each author's
  terms shown before you install; .zip, .7z and .uar banks all install.
• A singer picker with portraits and search, and a singer panel that says
  plainly when a bank or its vocoder is missing and how to fix it.
• Singing that sounds sung: natural expression, smooth joins and breaths.
• Finished takes are kept, so a reopened song plays without singing again.
  Freeze a vocal track, render a selection to a clip, and export as usual.
• Vocal songs go in and out as files, and a shared song sends its singers
  to collaborators where the bank's terms allow.

Faster: the first phrase of a session is sung in about half a second, a
change to a phrase's dynamics is almost instant, and reopening a song with
saved takes no longer loads every voice up front.

New: Settings > Updates > Beta updates. Turn it on to be offered test
builds before they are released. Off by default; turning it off again never
downgrades you.

---

Full Changelog: v0.27.1...v0.27.2

Fontelle 0.26.3

Full Changelog: v0.26.2...v0.26.3

Fontelle 0.26.2

Full Changelog: v0.26.1...v0.26.2

Fontelle 0.26.1

Full Changelog: v0.26.0...v0.26.1

Fontelle 0.26.0

Full Changelog: v0.25.2...v0.26.0

Fontelle 0.25.2

Full Changelog: v0.25.1...v0.25.2

Fontelle 0.25.1

Full Changelog: v0.25.0...v0.25.1

Fontelle 0.25.0

Full Changelog: v0.24.0...v0.25.0

Fontelle 0.24.0

Full Changelog: v0.23.1...v0.24.0