LYNCH AUDIO LAB — BROKE VOCAL 0.5.1 QUICK START Insert Broke Vocal on a vocal channel. The three large controls are TONE (warm to airy), DIRT (voice color and texture) and SPACE (the combined effect return, 0–200%). LEVELER evens the voice; AUTO ON/OFF beside it enables optional gain compensation. DOUBLE adds stereo micro-delay movement. OUTPUT trims the complete mix before the final limiters. Init is a clean, latency-aligned start. SHAPE SHAPE opens the controls inside this plugin window. Nine sections hold Voice, Tone / EQ, De-Esser, Tape Echo, Slap Delay, Spring, Plate, Galaxy and Dirt. Each section has a clear MODULE ON/OFF switch; OFF preserves its settings. Changing a page never loads a preset or changes the sound. Galaxy has two pages so its cloud, filters and pump stay readable. Close or Escape returns to the front. VOICE LEVELER is stereo-linked soft-knee compression, with Attack and Release under Shape. Increasing Leveler lowers the threshold and increases the ratio, giving a more even voice. Gate is a smooth downward expander; -80 dB is Off. Voice OFF bypasses the gate and Leveler, including Auto Gain, independently of the de-esser. AUTO GAIN compares the average voice energy before and after Leveler compression, then smoothly adds up to +18 dB to compensate the measured loss. The front AUTO ON/OFF button and Shape > Voice switch are the same saved, automatable setting. The AUTO +x.x dB readout shows the makeup gain currently applied. Compensation takes a little time to settle and approximately preserves the Leveler's average level; it does not guarantee exact perceived loudness or output peaks. The slower response retains the compressor's transient shaping. Auto Gain measures compression only, before Gate, De-Esser, Dirt and effects. It does not try to undo their sound or boost wet returns to a loudness target. Learning pauses during very quiet input or a closing/closed gate; already learned makeup gain is held instead of rising to chase silence. The estimator also follows compression while Auto is off, ready for a smooth switch on. Leveler at zero or Voice OFF fades compensation out. Auto defaults OFF in Init, factory presets and older sessions, preserving manual behavior. OUTPUT remains your final manual trim, and the Tube/Tape limiters still set the final ceiling when enabled. Wet Only removes the direct voice and keeps the five effect returns. Enable it for an aux/bus; leave it OFF for a normal vocal insert. Global Bypass always returns the original input. TONE / EQ TONE is neutral at 50%. Low and High are broad shelves; Mid has its own frequency control. Voice High Pass and Low Pass filter the main voice before dynamics, Dirt and effect sends. High Pass runs from 20 Hz (Off) to 20 kHz (Closed); Low Pass runs from 0 Hz (Closed) to 20 kHz (Off). Each filter has its own Q from 0.5 to 6.0: 0.707 is a gentle, non-peaking response; higher Q emphasizes the cutoff. At low sample rates the internal cutoff is safely bounded below Nyquist, while the Closed endpoint still fully attenuates the voice. Tone / EQ OFF bypasses all of this section. These voice filters are separate from the cuts on individual effects. They sit before Dirt: deliberately enabled Ambience/Crackle can still generate texture with a voice filter Closed. Set those texture controls to zero or switch Dirt OFF for silence. DE-ESSER Amount and Frequency control dynamic attenuation of the sibilant high band. The detector compares high-band energy with the voice's broadband energy, so quieter recordings can still trigger it. A quiet floor avoids reacting to silence. Start around 20–35%, then adjust for the voice. De-Ess Solo plays only what the de-esser removes, not the entire high band. It excludes Dirt, Double, effects and both final limiters; OUTPUT trims the audition. There is no hidden Solo gain boost. Solo is silent with De-Esser OFF or Amount at zero. Global Bypass takes precedence. The four-second graph shows actual high-band attenuation, with maximum reduction in each 50 ms bucket. Its scale expands when needed. The front warns when Solo is active. The graph is high-band gain reduction, not the change in full-vocal loudness. Dirt and strong final saturation can create new high frequencies after de-essing; use Dirt Low Pass or gentler distortion to control those separately. DIRT The large front DIRT knob is the master amount. Shape / Dirt provides independent Transistor distortion, Tube overdrive, Tape saturation, Ambience and Crackle, plus Dirt Tone. Start with one color, then blend in the others. The extra colors run Transistor → Tube → Tape after Base Color, inside the oversampled voice-color path. Zero on each new color removes that contribution. Dirt at zero or the Dirt module OFF removes all coloration and generated texture. Transistor adds harder bite; Tube adds asymmetric overdrive; Tape rounds and saturates. Ambience and Crackle are original generated noise textures, not room reverb or sampled records. They can continue in pauses; turn either amount to zero to remove that background. Use small amounts for age and larger amounts for an obvious damaged background. No commercial sample library is used. The former five large front buttons are now a compact Base Color menu here: Desk, Warm, Tube, Tape and Worn. This preserves the original voicings and older settings while freeing the front for larger knobs. Base Color only changes voice color; it does not load effect settings. These are original analog-inspired colors, not emulations of named hardware. DIRT HIGH PASS and LOW PASS shape the colored branch, including generated Ambience and Crackle, with a fixed smooth Q. High Pass spans 20 Hz (Off) to 20 kHz (Closed); Low Pass spans 0 Hz (Closed) to 20 kHz (Off). The filters act before the existing clean/Dirt blend. At lower Dirt settings some clean voice remains; Dirt at zero or the module OFF leaves the clean path unaffected by these cuts. MODULE SOLO / DELTA Each Shape page has a monitor button in the header. Only one module is monitored at a time; selecting another page does not change the sound until you press its monitor button. Press the active button again to return to the normal mix. The front and Shape header identify the active monitor; selection is saved and can be automated with Module Solo. The original De-Ess Solo button stays synchronized with this selector. Voice, Tone / EQ and Dirt use DELTA: the difference that the selected module adds or removes at its own stage. A neutral or bypassed module therefore gives silence. De-Esser SOLO gives the removed high-band signal. Tape Echo, Slap Delay, Spring, Plate and Galaxy SOLO give that effect's own filtered/ducked return, with its Send and master Space level. Upstream processing feeding that effect stays in place; other returns and the direct voice are excluded. OUTPUT trims every monitor. Both final limiters are excluded while monitoring, so their coloration does not disguise the selected module; MONITOR / NO LIMIT appears in the final strip. Global Bypass always returns the aligned original input. Solo is internal to Broke Vocal and does not solo the Logic track or mute other tracks in the session. Quiet difference signals can be much softer than the complete vocal. FIVE PARALLEL EFFECTS Tape Echo, Slap Delay, Spring, Plate and Galaxy can run together. Each has its own Send, Duck and recovery controls. SPACE is their combined 0–200% return level: 200% is twice the return amplitude at 100%. Send balances each effect; Duck lowers only that return while the voice is present. Duck at zero leaves it unducked. The dry voice remains independent. Each effect now has LOW CUT and HIGH CUT. Use Low Cut to remove low-frequency wash and High Cut to soften the top. These are effect filters, not filters on the dry voice. The new Low Cuts span 20 Hz–2 kHz with Off at 20 Hz; new High Cuts span 200 Hz–20 kHz with Off at 20 kHz. Galaxy keeps its original 20 Hz–1.2 kHz Low Cut, which filters its send and return. Tone remains the engine's original damping/voicing control, separate from the new cuts. Effect OFF stops new input and fades the return while the stored tail decays. SPACE at zero hides all returns; raising it can reveal a tail still decaying internally. Init, preset loading and host Reset clear the tails. TAPE ECHO Time sets free echo spacing. Sync and Division follow host tempo with straight, dotted and triplet choices. Feedback adds repeats; Wow adds motor drift/flutter. Tone softens repeats within the echo. Low Cut and High Cut shape the return separately. Ping Pong sends a mono sum into the left side and alternates feedback L/R/L/R. Feedback above zero gives alternating repeats. Use stereo output for movement; mono folds both sides together. With Ping Pong OFF, the original stereo echo behavior remains. SLAP DELAY A short 45–180 ms retro repeat, with Feedback, Width, Tone and separate ducking. Feedback at zero gives a single slap. Width offsets the left and right repeat times. Its own cuts keep the slap out of the low end or make it narrow and vintage. SPRING / PLATE Spring uses two dispersive feedback strings for metallic drip and splash. Plate uses a diffuse feedback network for a smoother tail. Both offer Size, Decay, Pre Delay, Tone, independent ducking, Low Cut and High Cut. Pre Delay separates the reverb onset from the voice. GALAXY Galaxy adds a long, wide diffuse echo cloud. Echo Time and its Sync/Division set spacing; Tail controls nominal mid-band decay. Diffusion smears echoes into a continuous cloud, while Motion adds floating delay-pitch movement. This is not an octave-shifting shimmer. Wide ranges from mono through natural stereo to an exaggerated spread; check it in mono. Tone damps the network. Warmth adds soft feedback saturation; Pre Delay separates the onset. Low Cut and High Cut control the cloud's frequency range. Very long tails suit pauses, throws and ambience. Duck follows the voice; 100–200% increases the attenuation depth beyond the original range. Duck Release controls how slowly the cloud blooms back. This detector works independently of the rhythmic pump. PUMP is a tempo-based volume envelope on Galaxy only. Amount controls depth; zero is Off. Rate sets its beat division, separately from echo spacing. Shape changes the recovery curve. Advancing host beat position locks the pump to the musical grid and follows loop/jump positions. Without host musical context it uses 120 BPM. When the host beat stops advancing, it runs freely at 120 BPM for audition. Pump does not need a kick sidechain. The dry voice and the other four returns are not pumped. Use Duck and Pump together for a cloud that stays behind words and breathes with the song. FINAL STAGE — TUBE THEN TAPE The bottom panel stays on the front. TUBE and TAPE have independent switches and amounts, always in that order. Use Tube alone for warm peak control, Tape alone for rounded saturation, or both for stronger color. They process the full vocal/effect mix after OUTPUT. MAX / CEILING sets the shared sample-peak ceiling from -18 to 0 dBFS; -1 dBFS is the default. Lowering it gives a genuinely lower maximum level when a limiter is active. RELEASE sets recovery from 20 to 800 ms. Both switches OFF removes final limiting; the ceiling then has no effect. Turning a stage on/off fades smoothly. Ceiling moves smoothly over roughly 20–25 ms; the selected maximum applies once the control has settled. Tube and Tape have separate real gain-reduction readings and four-second histories. These show the peak-control gain change, not an invented waveform or a measurement of all saturation. Saturation can be audible with little meter movement. No automatic loudness matching, lookahead, true-peak or LUFS targeting is claimed. PRESETS / SAVE / HELP 36 starting points: the previous 28 plus eight new Dirt, filtered-space and Galaxy-pump sounds. Init clears the processing for a clean start. Arrows browse presets; edits show Custom. Save and User use standard .aupreset files. Logic projects store all 105 sound/routing parameters. Original parameter IDs and AU identity are retained; missing new controls in older states restore neutral defaults. Existing De-Esser ON settings benefit from the corrected detector; use Amount to adjust its stronger response if needed. ? enables contextual help. Guide opens this document inside the plugin window; it scrolls and remains above the controls. Close or Escape returns to editing. Hover pauses while Guide or Shape covers the knobs. Normal drag/scroll, Shift for fine adjustment and double-click reset remain available. UI Size scales the whole interface within the screen. All sound controls support host automation. Use Logic Controller Assignments for MIDI hardware mapping; this audio effect does not have internal MIDI Learn or pitch correction. TECHNICAL Private Apple Silicon AUv2 effect, macOS 12+, aufx/Bv02/LyPl. Mono, Mono-to-Stereo and Stereo; noninterleaved float audio at 8–192 kHz. Processing and bypass remain aligned at 16 samples. No audio-thread allocation or disk reading. Noise textures are synthesized locally. Technical verification and musical listening approval are separate.