craft-audio in craft-libs was built from audio-io and FilmCraft's codecs, so
each format has one implementation that every Craft app shares. Call sites use
its API directly: EncodeOptions::new(format).with_*(), EncodeFormat,
DECODE_EXTENSIONS, ENCODE_EXTENSIONS, encode_format_for, AudioInfo's container,
codec and layout.
soundcraft_model::FileFormat stays as the kind of file a session records for
each source: its variant names are part of the session format. It maps from
craft-audio's AudioInfo.
craft-audio and craft-media are pinned to the craft-libs branch of
storytold/craft-libs#824, which adds the extension tables SoundCraft had in
FileFormat and fixes the quadratic WAV chunk scan (#893).
Behaviour that changes: AAC, ALAC, Opus, AC-3/E-AC-3 and MP2 decode with
craft-audio's own decoders; MP4 and Matroska tracks import at their
presentation length; .opus, .ac3, .eac3, .ec3, .mov, .mka and .mpa are
importable names; .aifc is a writable name; FLAC files come from craft-audio's
encoder.
release.yml also runs nightly on main with craft-common-actions' dry-run: every release build
except macOS, with nothing signed, pushed or drafted. Breakage that only the release build hits
(riscv64, Flatpak, FreeBSD, packaging) then shows up the next morning instead of on release
day, as it did on 11 Nov 2026.
* fix(engine): create output files exclusively and check read access before probing
* address review: canonicalize through the checked io gateway, treat any occupied export name as taken
* address review: read group audio through the validated file handle
* address review: document that a write-only root writes a new numbered WAV per save
* address review: clip-group audio is opened without waiting for a FIFO writer
* Strip Silence: cut and remove pieces in one pass over the playlist
* address review: document the sorted-points order and test stretched audio offsets
* fix(ui): check movie read access and file type; private take folder without settings
* address review: make the fallback take folder exclusively under an unguessable name; qualify the owner-only claim to Unix
* address review: unique folder prefix must be one name; no fallback take folder for a saved session
* add tests for the take folder and the movie open checks
* fix(ui): bound MIDI grid lines and Video Universe thumbnails per repaint
* address review: grid fallback stays on snap positions and bars; thumbnail positions in f64
* address review: no thumbnails for an empty shown span
* fix(audio-io): bound packet samples and MP4 audio timing runs
* address review: decide MP4 audio timing bound from the stsd entry and bound PCM packets by their byte size
* address review: the last stsd in a stbl decides whether the stts bound applies
* fix(video): check MJPEG frame size from its header and budget H.264 frame requests
* address review: check the sample budget before reading, keep the decoder from skipping a refused sample, stop on a failed read
* address review: restore lenient MJPEG sizing and unreadable-sample skipping; per-request sample cap 2 GiB
* address review: an unreadable sample counts toward the no-picture budget
Adds step-security/harden-runner (audit, pinned to v2.22.4 like
craft-common-actions) as the first step of each job that did not have it,
so every workflow in the org's public repos reports its egress to
StepSecurity. Audit, not block: block-mode allow lists come from each
job's observed traffic.
* Read volume automation at every sample, not once per block
Track and master faders read the dB curve at a block's first and last
sample and ramped linearly in gain between them, so a breakpoint acted up
to a block early or late and a short crossfade came out differently
depending on where the block edges fell. Both now take the gain at each
sample: the curve is read at its breakpoints and interpolated in dB in
between, as the lane defines it. Blocks with no change keep one gain.
Fixes#1354.
* docs: volume breakpoints no longer depend on the render block (G-04)
* Step fader ramps by a ratio and scale channels in contiguous runs
Reading the fader at every sample converted dB to gain per sample on
every ramp, and applied it sample by sample across channels, even when
nothing moved. A stretch that is linear in dB is geometric in gain, so
each stretch now converts its two ends exactly and steps the samples in
between by one ratio (in f64, so it does not drift), with the -144 dB
floor and the cap checked per sample. Each channel is scaled over the
whole stretch at once, and a stretch with no change is a plain multiply.
Tests: stepped ramps against db_to_gain sample by sample (floor, cap,
NaN, one or two samples, a 16384-sample ramp), and summed volume, trim
and VCA lanes with a muted track, read from renders that start inside a
ramp or on block edges.
* fix(mix): recover extreme automation ramps from silence
* fix(mix): preserve extreme ramp endpoints and silence NaNs
---------
Co-authored-by: FredApps <FredApps@users.noreply.github.com>
* Compensate a hosted instrument's latency
Delay compensation summed only insert latency, so a synth reporting
latency played late against everything else in playback and bounces.
run_inserts now seeds the strip latency with the instrument's.
Fixes#1474. Added a mix test that fails without the change: a latent
hosted instrument now contributes its latency to the mix.
Signed-off-by: Vladimir Babin <vovababin@gmail.com>
* test(mix): verify rendered instrument latency alignment
---------
Signed-off-by: Vladimir Babin <vovababin@gmail.com>
Co-authored-by: FredApps <FredApps@users.noreply.github.com>
* CLAP host: a reset keeps the queued note-offs and releases held notes (#1464)
ClapPlugin::reset kept only parameter changes in its event queue and cleared
the held keys, so the note-offs all_notes_off had just queued on Stop, Play
and Locate never reached the plugin. A plugin whose own reset leaves its
voices sounding (Surge XT) played them on at the next Play, and All MIDI
Notes Off had nothing left to release.
Like the VST3 host since #441, reset now keeps the pending releases (CLAP
and MIDI note-offs, one per key, at the start of the next block) and
releases whatever is still held. The fixture synth's reset no longer
silences it, so the integration test sees what Surge XT does.
* CLAP host: keep queue room for the releases a reset sends (#1464)
A queue that distinct parameter changes filled had no room left for the
releases reset queues, so a held note could still sound on. The last 128
queue slots now take only note_off releases, and reset queues one release
per held key that has none yet, which always fits in that reserve.
* CLAP host: a release the full queue drops leaves its key held (#1464)
note_off cleared the key's held flag even when its release did not fit in
the queue, so the reset after all_notes_off had nothing left to release.
The flag is now cleared only when the release was queued; the reset then
sends it, after keep_releases has made room.
Behaviour-preserving: `[[ ]]` in place of `[ ]` for bash tests (SonarCloud shell:S7688). Every script has a bash shebang; no test uses `-a`/`-o`, escaped parentheses or an unquoted pattern on the right of `=`. `bash -n` passes on each script; the rest is left to CI.
- `packaging/linux/flatpak-bundle.sh` (4)
* Consolidate: run clip effects over the clip and a bounded tail
render_clips allocated and ran a clip-effects workspace spanning the
whole selection for every clip with effects, however short the clip.
Many tiny effect clips in a long selection multiplied allocation and
effect processing by the selection length.
Process each clip's effects in reused 4096-frame blocks from the clip's
start (the effects start from silence, so nothing comes out before it)
through its end, then keep rendering the tail until a block falls below
-180 dBFS or 10 s past the clip's end.
* address review: settle clip effect smoothers on a bounded silent lead-in
* address review: skip the lead-in for gain-only clip effects; keep the full-span chunk grid
Display labels only; command ids, catalog paths, session tokens and processor ids are unchanged.
- Avid-listed names and bundled plug-in products get our own names (Beat Analysis, Punch Recording, Room Reverb, Compressor/Limiter, Saturator, Bit Crusher, Rendered (High Quality), ...).
- Features we don't have stay in the menu as disabled placeholders with "Not available yet." (Tempo Sync, No ARA plug-ins, Analog Saturation).
- AudioSuite, Elastic Audio, TrackPunch, Pitch Shift Legacy and Time Compression Expansion keep their names until RFC-0029 settles.
- parity_floor and catalog dispatch unchanged; docs/feature-naming.md has the crosswalk.
Moves the craft-libs crates (craft-update, craft-docpath, craft-media,
craft-ui) from c5a1cbfb5f to a branch that merges storytold/craft-libs
#687 and #696.
- The result dialog shows whenever the updater reports a status, so the
first start after an update says whether it installed or failed.
Closing it dismisses the status; the startup prompt still takes over.
- Restart to Update in the dialog asks the app to close through the
usual close path.
- on_exit starts the helper of a ready update, since the system Quit
ends the process before the updater is dropped.
Address the review by clearing stale test scratch directories before setup and exercising an explicit supported WAV format through the real command, including decoded output format, frames and audio.
About > Contributors still showed an old snapshot, because nothing
regenerated contributors/contributors.json before a release. Opt in to
the shared refresh from craft-common-actions: a read-only credits job
regenerates the snapshot for the release commit, and the shared release
workflow installs it before every source build. The snapshot is not
committed back.