# CALM - User Instructions **Everworks** **Neural feedback control, de-reverb and vocal protection** **Version | May 2026** --- ## Index 1. What CALM is 2. Installation 3. Quick start 4. Processing modes 5. Main controls 6. SERENE and BLEED 7. Tonal Lock 8. Kalman Adaptive 9. Vocal Chain 10. Presets and recommended workflow 11. Use cases 12. Troubleshooting 13. Best practices --- ## 1. What CALM Is CALM is a plugin designed to protect and improve voices in real time. It combines neural anti-feedback, resonance control, de-reverb, de-bleed, noise reduction and a vocal DSP chain. The goal is not to crush the signal. The goal is to keep the voice stable, clear and usable when real problems appear: - Feedback. - Tonal resonances. - Room reverberation. - External bleed. - Background noise. - Sibilance. - Lack of body or presence. CALM works best when used as the first important processor on the vocal channel, before creative reverbs or mix effects. --- ## 2. Installation ### macOS The macOS package contains: - `CALM Universal.component` - `CALM Universal.vst3` - `CALM Universal.app` Recommended paths: ```text AU: ~/Library/Audio/Plug-Ins/Components/ VST3: ~/Library/Audio/Plug-Ins/VST3/ Standalone: Applications or any local folder ``` Steps: 1. Close Logic, Ableton, Reaper, Pro Tools or any open host. 2. Copy `CALM Universal.component` to `~/Library/Audio/Plug-Ins/Components/`. 3. Copy `CALM Universal.vst3` to `~/Library/Audio/Plug-Ins/VST3/`. 4. Open the DAW and force a rescan if needed. 5. In Logic, CALM appears as an Everworks Audio Unit. If Logic does not show it: - Close Logic. - Open Logic's Plug-in Manager. - Reset or revalidate the plugin. - Make sure there is no older duplicated version with another name. ### Windows The Windows package contains: - `CALM Windows.vst3` - `CALM Windows.exe` Recommended VST3 path: ```text C:\Program Files\Common Files\VST3\ ``` Steps: 1. Close the DAW. 2. Copy the `CALM Windows.vst3` folder to `C:\Program Files\Common Files\VST3\`. 3. Open the DAW and rescan plugins. 4. To test outside the DAW, open `CALM Windows.exe`. --- ## 3. Quick Start ### Safe starting point 1. Insert CALM on the vocal channel. 2. Leave the mode on **Studio** if you are mixing or recording. 3. Use **Balance** if you need less latency. 4. Use **Live** if latency is critical. 5. Keep **CALM** enabled. 6. Set the central knob between 30% and 50%. 7. Enable **Tonal Lock** if there are narrow tones, feedback or headset microphones. 8. Enable **SERENE** only if room/reverb is a problem. 9. Enable **BLEED** only if external ambience or instruments are entering the mic. 10. Adjust the Vocal Chain at the end if you need additional color or cleanup. ### Recommended live setup ```text Mode: Balance or Live CALM: On Tonal Lock: On if there is tone/feedback risk Turbo: Off at first, On only in critical situations SERENE: low or medium if there is room BLEED: low or medium if there is bleed DNS: gentle if there is constant noise Auto Gain: On Output Trim: adjust final output ``` ### Recommended studio setup ```text Mode: Studio CALM: On if there are resonances or simulated feedback Tonal Lock: On if there is a narrow tone SERENE: adjust by Amount BLEED: only if there is external ambience Clean / DeEss / Body / Glue: use as vocal chain Auto Gain: On Output Trim: adjust final output ``` --- ## 4. Processing Modes ### Studio 1024 Highest spectral resolution. Recommended for: - Mixing. - Recording. - Broadcast. - Podcast. - Post-production. - Voices where quality matters more than latency. ### Balance 512 Quality and latency compromise. Recommended for: - Live sound. - Conferences. - Headsets. - Lecterns. - Monitors. - Low-latency testing. ### Live 256 Minimum latency. Recommended for: - Real-time monitoring. - Latency-sensitive systems. - Live PA. - Computers with less CPU headroom. --- ## 5. Main Controls ### CALM Enables the main anti-feedback engine. Recommendation: - Leave it enabled almost always. - Turn it down only if you want to use CALM as a vocal channel without anti-feedback protection. ### Central Knob Controls the main intensity of the system. Quick guide: | Value | Use | |---|---| | 0-20% | Very gentle, almost transparent | | 30-50% | Recommended range | | 50-70% | Difficult live situations | | 70-100% | Emergency range, artifact risk | If the voice becomes too dull, lower the central knob first before changing other modules. ### Turbo Makes CALM more aggressive. Use it when: - The system is close to feedback. - A resonance keeps coming back. - You need more safety than transparency. Do not use it by default in studio if it is not needed. ### Ghost Experimental variation/perturbation mode. It may help in some extreme scenarios, but it should not be the first adjustment. ### Bypass Disables processing so you can compare input and output. --- ## 6. SERENE and BLEED ### SERENE SERENE reduces room tail and reverberation. In this version it uses the **SERENE Mamba V4 Elite** model. Main control: - On/Off. - Vertical Amount. Use it if: - The voice sounds distant. - The room masks words. - There is a tail after each phrase. - The microphone captures too much room. Do not push it to maximum by default. A medium setting is usually more natural. ### BLEED BLEED reduces external ambience and spill around the voice. Main control: - On/Off. - Vertical Amount. Use it if: - Drums, guitar, PA or audience enter the vocal mic. - Monitors are close. - The voice shares the space with other sources. BLEED should not be used like a hard gate. Raise the Amount until focus improves without cutting the voice. ### Difference Between SERENE and BLEED | Module | Attacks | Example | |---|---|---| | SERENE | Room/reverb tail | A voice recorded in a live room | | BLEED | External sound/spill | Drums entering a vocal microphone | --- ## 7. Tonal Lock Tonal Lock is designed for narrow and stable tones. It is very useful with: - Lecterns. - DPA-style headsets. - Lavalier microphones. - Theatre. - Conferences. - Voice with background music where a fixed tone appears. When enabled, CALM can attack a tonal resonance even while voice or music is playing. This avoids the case where the detector protects too much of the useful signal and lets a sustained tone through. Recommendation: - Enable it live if there is feedback risk. - Use it carefully on very tonal musical material. - Combine it with Resonance if there is harshness or annoying peaks. --- ## 8. Kalman Adaptive The Kalman matrix shows frequency risk zones. Colors: | Color | Meaning | |---|---| | Dark | No risk | | Green/cyan | Low risk | | Amber | Medium risk | | Red | High risk | In V1, the matrix is not decorative: it is fed by real confirmed detections. If it lights up, CALM has detected a risk area or the Kalman memory is reacting. What it is for: - Helps repeat the correction if the same frequency returns. - Reinforces Tonal Lock. - Reduces reaction time in problematic bands. --- ## 9. Vocal Chain Vocal Chain contains the secondary modules. You can keep it hidden for fast operation or expand it when you need fine adjustment. ### Expanse Expander for reducing noise between phrases. Controls: - Threshold. - Depth. - Speed. Tip: - Set Threshold until the noise drops between phrases. - Use moderate Depth to avoid abrupt cuts. ### Resonance Smooths dynamic resonances. Controls: - Sensitivity. - Depth. - On/Off. Tip: - If the voice becomes harsh, raise Sensitivity gradually. - If it loses life, lower Depth. ### DNS Reduces constant noise. Controls: - Bias. - Attenuation. Tip: - A small amount of well-adjusted DNS is better than aggressive DNS. - If watery artifacts appear, lower Attenuation. ### Clean Musical EQ for focus and clarity. Controls: - Focus. - Focus Gain. - Clarity. - Clarity Gain. Tip: - Use Focus to bring the voice forward. - Use Clarity moderately for air. ### Body Adds controlled low body. Controls: - Frequency. - Amount. - Slope. Tip: - Works well on thin voices. - Avoid too much if there is proximity effect or low room buildup. ### DeEss Controls sibilance. Controls: - Mode. - Threshold. - Range. - Frequency. - Sharpness. - Listen. Tip: - Use Listen to find the sibilance area. - Turn Listen off when finished. ### Vocal Finisher Final color and polish. Controls: - Drive. - Opto. - Air. Tip: - Drive for density. - Opto for smoothness. - Air for brightness. ### Glue Final compression to glue the voice. Controls: - Threshold. - Ratio. - Makeup. - Auto. Tip: - Use small gain reductions. - If the voice pumps, raise Threshold or lower Ratio. ### Auto Gain and Output Trim Auto Gain keeps the output level more stable. Output Trim allows manual final level adjustment. --- ## 10. Presets and Recommended Workflow ### Live workflow 1. Balance or Live mode. 2. CALM On. 3. Tonal Lock On if there is feedback risk. 4. Strength around 40%. 5. Raise microphone gain to the working level. 6. Adjust Resonance if there is harshness. 7. Enable DNS only if there is constant noise. 8. Enable SERENE or BLEED only if needed. ### Studio workflow 1. Studio mode. 2. CALM On gently. 3. SERENE if there is room. 4. DeEss if there is sibilance. 5. Clean for presence. 6. Body if the voice needs weight. 7. Glue/Vocal Finisher at the end. 8. Adjust Output Trim. --- ## 11. Use Cases ### Headset / DPA live Recommended: ```text Mode: Balance CALM: On Tonal Lock: On Turbo: Off at first Resonance: gentle SERENE: low if there is room BLEED: low if there is PA or audience ``` ### Lectern / conference Recommended: ```text Mode: Balance or Live CALM: On Tonal Lock: On DNS: gentle if there is room noise Expanse: gentle to clean pauses ``` ### Voice recorded in a live room Recommended: ```text Mode: Studio SERENE: On SERENE Amount: medium Clean: as needed DeEss: if needed ``` ### Voice with instrument bleed Recommended: ```text Mode: Balance BLEED: On BLEED Amount: low/medium SERENE: only if there is also room DNS: only if there is constant noise ``` ### Podcast in a domestic room Recommended: ```text Mode: Studio SERENE: low/medium DNS: gentle DeEss: if there is sibilance Body: gentle Glue: gentle ``` --- ## 12. Troubleshooting ### The plugin does not appear in Logic 1. Make sure the AU is here: ```text ~/Library/Audio/Plug-Ins/Components/ ``` 2. Close Logic. 3. Open the Plug-in Manager. 4. Revalidate CALM. 5. Make sure you are searching for the correct name: in this build it can appear as `CALM Universal`. ### The DAW does not find the VST3 on Windows Make sure the folder is here: ```text C:\Program Files\Common Files\VST3\CALM Windows.vst3 ``` If the DAW has a plugin cache, force a full rescan. ### The voice becomes dull Try this order: 1. Lower the central knob. 2. Turn Turbo off. 3. Lower SERENE Amount. 4. Lower BLEED Amount. 5. Lower DNS Attenuation. 6. Lower Resonance Depth. ### The tone/feedback appears only when voice or music is playing Enable Tonal Lock. That mode was created for narrow resonances that appear on top of useful signal. ### BLEED cuts the voice Lower BLEED Amount. BLEED should act as gradual reduction, not as a mute. ### SERENE does not make the voice completely dry This is normal. SERENE is tuned to preserve the direct voice. If you want more room reduction, raise Amount carefully and combine with Expanse or DNS only if the problem requires it. ### There is too much latency Switch to: - Balance for lower latency. - Live for minimum latency. --- ## 13. Best Practices - Do not enable every module by default. - Start with CALM and Tonal Lock. - Use SERENE only when there is real reverb. - Use BLEED only when there is real bleed. - Use DNS only when there is constant noise. - Adjust with the voice playing, not only with test tones. - A pure test tone does not always behave like real feedback. - Always compare with Bypass at the same level. - If an improvement sounds spectacular but dulls the voice, reduce intensity. --- ## 14. Quick Summary For live: ```text Balance / Live CALM On Tonal Lock On Strength 35-55% Turbo only when needed ``` For studio: ```text Studio CALM gentle SERENE if there is room Vocal Chain for finishing Auto Gain On ``` For bleed: ```text BLEED On Amount low/medium Raise only until focus improves ``` For reverb: ```text SERENE On Amount medium Avoid forcing it if the voice starts losing naturalness ```