Skip to content

Seraph v0.1.0

Choose a tag to compare

View all tags
@yves-vogl yves-vogl released this 14 Jul 16:33
v0.1.0
aa81ae3
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Seraph is a choir and vocal processor built for operatic metal vocals. In a heavy-music mix chain it sits on lead or choir vocal tracks: it tames sibilance, adds airy openness above the vocal's natural top end, glues dynamics together with a gentle broadband compressor, and thickens a lead or choir line with a click-free four-voice doubler — all in one channel-strip-style plugin. Part of the "Basilica Audio" JUCE 8 plugin suite.

This is the first tagged release, covering the M1 milestone: DSP completion and test coverage.

What's in v0.1.0

Added

  • Project bootstrap: README, license, contributing guide, architecture and build docs, ADRs, and CI workflow.
  • DSP core: initial working Seraph signal path (De-Ess, Air, Doubler, Mix/Output) with unit tests.
  • Gentle Compressor DSP stage (Comp parameter, 0-100%): a hand-rolled, zero-latency, bit-exact-bypassable broadband downward compressor placed after Air and before the Doubler, for evening out dynamics before the signal is duplicated into doubled voices.
  • De-Ess Listen mode (DeEssListen parameter): solos the de-esser's detected sibilance band so DeEssFreq can be tuned by ear, independent of the current DeEss reduction amount.
  • Doubler extended from two to four voices, each with its own fixed per-voice pan position (a small-choir spread scaled by DoubleWidth, rather than a single symmetric L/R pair); gain-staging is compensated so the added level matches the original two-voice design at DoubleWidth == 0.
  • GUI: added a Comp knob and a De-Ess Listen toggle button to the v0.1 slider editor (now two rows of five controls).
  • Test suite broadened from 18 to 28 Catch2 tests: sample-rate sweeps (44.1-192 kHz) for the null test and for the full chain at hot settings, mono/stereo/rejected bus-layout coverage, a long-run (~10.7 s simulated) NaN/Inf and unbounded-growth stability sweep, prepareToPlay sample-rate-change robustness, and dedicated coverage for the new compressor, listen mode, and four-voice doubler.
  • docs/manual.md: a full user manual (what Seraph is, where it sits in a heavy-music vocal chain, signal flow, complete parameter reference, and mixing tips).

Deferred

  • True formant-preserving detune (LPC/cepstral spectral-envelope correction for the doubler) was requested for M1 but not implemented — it is a substantially larger DSP feature than fits safely alongside the rest of this milestone without risking the plugin's zero-latency and bit-exact-bypass invariants. See docs/architecture.md's Doubler section and the M1 "Complete and refine the DSP" issue, left open for this follow-up.

Formats

  • AU (Audio Unit) — macOS
  • VST3 — macOS and Windows
  • Standalone — macOS and Windows

Install

macOS

Format Path
AU (Component) ~/Library/Audio/Plug-Ins/Components/
VST3 ~/Library/Audio/Plug-Ins/VST3/

If your DAW (e.g. Logic Pro) doesn't pick up the plugin after installing, force a rescan by resetting the AU cache:

killall -9 AudioComponentRegistrar
auval -a

Windows

Format Path
VST3 C:\Program Files\Common Files\VST3\

Warning: unsigned binaries

These binaries are UNSIGNED and not notarized. macOS Gatekeeper and Windows SmartScreen will likely warn or block on first launch. Code-signing and notarization are tracked as roadmap milestone M4 and are not part of this release. Use at your own risk, and expect to manually approve the plugin in System Settings → Privacy & Security (macOS) or via "More info → Run anyway" (Windows) if prompted.

Assets 4
Loading