Repository navigation
Releases
Releases · skttr87/getsu
Getsu v1.3.1
What's New in Getsu v1.3.1
- Windows CoreAudio PnP Hardware Bus Auto-Detection (
PKEY_Device_EnumeratorName):- Implemented direct Win32 COM querying of
IMMDeviceproperty stores via ctypes (PKEY_Device_EnumeratorName,{a45c254e-df1c-4efd-8020-67d146a850e0}, PID 24). - Categorizes audio hardware into canonical bus architectures:
BLUETOOTH_HFP(BTHHFENUM),BLUETOOTH_A2DP(BTHENUM),USB,INTERNAL_HD(HDAUDIO),INTEL_AUDIO(INTELAUDIO), andVIRTUAL(VB-Audio / cable drivers).
- Implemented direct Win32 COM querying of
- Native Active Playback Resolution:
- Direct Win32 CoreAudio query of active default audio endpoints (
eRenderacrosseConsole,eMultimedia, andeCommunications), completely removing reliance on brittle MMEsd.default.device[1]index lookups. - 4-tier output resolution in
get_physical_output_device()prioritizing active CoreAudio playback endpoint, hardware bus priority, keywords, and verified physical endpoints.
- Direct Win32 CoreAudio query of active default audio endpoints (
- Dynamic Bluetooth Hands-Free (HFP) Jitter Buffer Adaptation:
- Automatically identifies Bluetooth HFP headsets (
BTHHFENUM) and silently adapts PortAudio stream latency to0.035s(35ms RF packet smoothing buffer) while retaining ultra-lowlowlatency (10ms) for USB and HD Audio devices.
- Automatically identifies Bluetooth HFP headsets (
- Hardware-Aware Laptop Microphone Detection:
- Upgraded
is_laptop_microphone()to inspect hardware bus enumerators (INTERNAL_HD,INTEL_AUDIO) before falling back to localized name matching. - Harmonized gain compensation (1.2x) across live streaming (
_compute_current_gain) and in-memory Voice Test preview (_run_voice_test_worker).
- Upgraded
- Voice Test Phase 2 Live Output Rescan:
- Live Voice Test preview explicitly triggers
get_physical_output_device(force_rescan=True)immediately before playback, routing test audio to the current default output even if user plugged in headphones mid-recording.
- Live Voice Test preview explicitly triggers
- Hotplug Concurrency Synchronization & Rescan UI Error Recovery:
- Added state lock acquisition with timeout in
_handle_hardware_rescan(), preventing race conditions with_stop_workerfrom resurrecting orphaned background streams when hotplug events occur during user stop. - Implemented comprehensive error recovery: if stream restart fails after hardware changes, the UI automatically resets
btn_toggleto "START", status badge to[ STOPPED ], and re-enables Voice Test rather than locking the user out.
- Added state lock acquisition with timeout in
- Type-Safe Device Querying & Codebase Hygiene:
- Hardened
get_all_devices()against malformed mock data, non-dict elements, and string/non-integer host API indexes. - Cleaned dead code (
hide_to_tray) and pruned unused imports acrosssrc/gui.py,src/stream.py,src/cli_cleaner.py, andsrc/tray.py.
- Hardened
Verify Your Download
- File:
Getsu-v1.3.1-Setup.exe - SHA-256:
EA159395EA4C0869FAE064AC3714AF555272131D9AABD5BFB330AF9F530BAB39 - MD5:
115435FE89C03022DC8257F29A663A43 - Size:
38,328,459 bytes(~36.6 MB)
Getsu v1.3.0
What's New in Getsu v1.3.0
-
Dual-Binary Architecture & Batch CLI Engine (
getsu-cli.exe&getsu.exe):- Introduced dedicated console-subsystem binary (
getsu-cli.exe,console=True) alongside the GUI app (getsu.exe,console=False), ensuring full synchronous execution in PowerShell/cmd and reliable$LASTEXITCODEreturns. - Decoupled CLI batch cleaning from GUI single-instance mutex locks (
activate_existing_instance()), allowing real-time desktop filtering and offline batch file cleaning to run simultaneously without interference. - Production-grade batch cleaner with universal format decoding (
miniaudio), automatic 48 kHz resampling, sample-accurate duration trimming, 2-frame zero-flush latency extraction, and TPDF-dithered 16-bit PCM output. - Coupled Linked-Stereo Gating (
--stereo): Synchronized master gain envelope across multi-channel audio eliminating stereo panning image wobble.
- Introduced dedicated console-subsystem binary (
-
Adaptive Near-Field Voice Anchoring (
$L_{\text{user}}$ ):- Dynamic Voiced Speech Anchor tracking the 90th percentile of user speaking frames; classifies distant cubicle background chatter
$> 14\text{ dB}$ below the anchor as far-field bleed and routes it to downward expansion ($\le 0.15$ gain) rather than tripping open the gate. - Strictly preserves whispers and soft phrase starts with a 400ms near-field hangover.
- De-stacks manual mic boost and dynamic speech leveling (
SpeechLeveler), shifting target level bymic_boost_dbto protect headroom.
- Dynamic Voiced Speech Anchor tracking the 90th percentile of user speaking frames; classifies distant cubicle background chatter
-
In-Speech Transient De-Clicking & Pre-RNNoise Saturation Guard:
- Zero-latency
TransientSuppressorattenuates high-frequency ($> 2.5\text{ kHz}$ ) mechanical keyboard switch clicks and desk thumps by$6\text{–}10\text{ dB}$ using smooth S-curve / Hann windowing. - Pre-RNNoise
soft_preclip()saturation curve rounds off ADC rail flat-topping ("P-pops") to protect neural Bark-band features. - One-click Music & Performance Mode (
music_mode) bypassing RNNoise with an extended$800\text{ ms}$ natural hangover and$250\text{ ms}$ decay for singing vibrato and acoustic instruments.
- Zero-latency
-
Proportional Hysteresis & Dynamic Gate Synchronization:
- Coupled
vad_thresholdandvad_close_thresholdproportionally during live slider adjustments, eliminating inverted hysteresis anomalies where low sensitivity settings caused premature gate dropouts.
- Coupled
-
GUI Viewport & System Tray Parity:
- Expanded dynamic notice headroom to +26px in
src/gui.py, preventing multi-line status wrap pills from crowding the non-scrollable window layout. - Synchronized friendly device names alongside IDs in
src/tray.py, and isolated runtimemuteto prevent ephemeral state pollution inconfig.json.
- Expanded dynamic notice headroom to +26px in
-
Packaging & Windows PATH Integration:
- Inno Setup
setup.issintegration withChangesEnvironment=yes, optional PATH registration task (addtopath), and clean uninstall routines. - Updated PyInstaller spec and
scripts/publish.pypipeline auditing bothgetsu.exeandgetsu-cli.exebinaries before release.
- Inno Setup
Verify Your Download
- File:
Getsu-v1.3.0-Setup.exe - SHA-256:
AD7293208CE74F3A76EF3018AF0B327FA8B59E38C8DEFF58AE11836BF942E592 - MD5:
542C2C05571B8D567F814E54A9324C9D - Size:
38,324,363 bytes(~36.5 MB)
Getsu v1.2.9
What's New in Getsu v1.2.9
-
Actionable Hardware Jack Disconnect Diagnostics (
0x00000492): Dedicated, hedged error classification for Realtek/Windows open-circuit jack errors (ERROR_SET_NOT_FOUND). Replaced misleading audio path advice with clear troubleshooting instructions (checking 3.5mm cable, using a Y-splitter for combo headsets, or trying another microphone). -
Text Truncation Removal: Eliminated the artificial 82-character truncate in
gui.py:_format_audio_error(). The multi-sentence troubleshooting advice now wraps smoothly inside the status notice without clipping or overlapping the footer. -
Unified Core DSP Pipeline (
process_mono_frame): Consolidated frame-level DSP (High-Pass Filter$\to$ Metering$\to$ RNNoise$\to$ Gain Staging$\to$ AdaptiveNoiseGate$\to$ Soft Limiter) into a single authoritative processing function shared by both the live streaming engine and the Voice Test thread, permanently eliminating parameter divergence and volume desync. -
Acoustic Speech Leveler (
SpeechLeveler): Introduced dynamic target speech leveling for quiet microphones (Bluetooth TWS earbuds, USB headsets, and laptop mic arrays) targeting$-24.0\text{ dBFS}$ RMS with a$+9.0\text{ dB}$ maximum boost clamp and asymmetric slew rates. Staged safely behindconfig["auto_level"](default off for v1.2.9). -
Dead Code & Config Sanitization: Fully pruned dormant
vad_cold_start_gain(superseded by continuous downward expander floor0.06), deprecated contradictoryconfig["mic_gain"]in favor of authoritativemic_boost_db, removed unusedensure_physical_default_playback(), and purged unread metadata (mute,app_name) fromDEFAULT_CONFIGwith an automated migration purge on startup.
Verify Your Download
- File:
Getsu-v1.2.9-Setup.exe - SHA-256:
112FD4B25D771430B48BB4EE7ED14B982689C1F2102960BCDC4E8A1E1D68C97E - MD5:
D6DBB1BE3F8BDB0F2FFC916D32B56777 - Size:
37,471,484 bytes(~35.7 MB)
Getsu v1.2.8
What's New in Getsu v1.2.8
- Host-API Tiering & Deduplicated Safe Inputs (
list_selectable_inputs): Enumerate physical recording endpoints using modern Windows CoreAudio (WASAPI) as the authoritative hardware layer, completely pruning synthetic Win32 proxy mappers (Microsoft Sound Mapper,Primary Sound Capture Driver) across all Windows locales. Virtual cable bridges (CABLE Output) are strictly excluded to prevent infinite feedback loops. - Hardware Channel Count Fidelity: Capture authentic mono/stereo input channels (
SelectableInput.channels) directly from device capabilities rather than hardcoded layouts. - Dynamic Wrong-Jack Advisory (
selection_hint): Added contextual, real-time UX hints beneath the microphone dropdown ingui.py. If a user selectsLine InorStereo Mix, Getsu immediately warns them and points them to the pinkMicrophoneentry before stream initiation. - Granular PortAudio Error Diagnosis (
src/audio_errors.py): Replaced generic"driver conflict"error banners with actionable, plain-language diagnosis for specific PortAudio error codes (-9985busy/exclusive mode,-9997sample rate mismatch,-9993host mismatch,-9999host error) while preserving up to 400 characters of raw technical error text for logs. - Targeted Transient Stream Retry: Audio stream retry loop now specifically retries on transient
-9985(PaErrorCode -9985device busy / exclusive lock) rather than wasting 700ms on unrecoverable disconnected hardware errors. - Subscripting Compatibility & Hardening: Implemented dictionary subscripting (
item['index'],item['name']) onSelectableInputdataclass to ensure backwards compatibility acrossdevices.pyand UI lookups.
Verify Your Download
- File:
Getsu-v1.2.8-Setup.exe - SHA-256:
75DBB62F4410D29CBD09C33B41E9A44FB9278079023534175E591620CD4399D5 - MD5:
D9975B9182461AB4F260C2EFB550BB2F - Size:
37,467,448 bytes(~35.7 MB) - VirusTotal Report: 75DBB62F4410D29C...
Getsu v1.2.7
- Proactive Host API Resolver: Introduced
src/host_api_resolver.py(resolve_pairandresolve_single) to inspect and align duplex input/output endpoints onto the same safe Windows Host API (prioritizing WASAPI -> DirectSound -> MME) before stream initialization, eliminating PortAudio host mismatch crashes (PaErrorCode -9993). - Elimination of WDM-KS Driver Crashes: Proactively avoids PortAudio's unstable WDM-KS backend on Realtek onboard audio jacks (
GLE 0x492/ERROR_SET_NOT_FOUND/PaErrorCode -9999), transparently mapping selected endpoints to modern, stable WASAPI equivalents. - Intelligent Jack Role & Token Matching: Resolves hardware jack endpoints (e.g., WDM-KS
"Mic in at front panel (black)") to matching software endpoints (e.g., WASAPI"Microphone (Realtek(R) Audio)") using acoustic role dictionaries (mic,line,mix) and token similarity. - PortAudio Channel Desync Guard: Updates stream channel counts (
in_channels,out_channels) directly from resolved device metadata, preventing mono/stereo channel mismatch invalidations (PaErrorCode -9998). - Host-Guarded Stream Settings: Enforces
sd.WasapiSettings(auto_convert=True)strictly on WASAPI streams, preventing invalid parameter exceptions (PaErrorCode -9984) on DirectSound and MME fallbacks. - Synchronized GUI & Voice Test Engine: Updated
src/gui.pyto route both the live engine and the "Hear Myself" 11s voice test through the resolver. Resolved device indices are safely persisted back toconfig.json, with visual indicators if endpoint mapping was ambiguous. - Energy-Guarded Consonant Sustain (Dual-Condition Keep-Open): Added acoustic energy evaluation (
input_rms_db - noise_floor_db >= 6.0 dBandP >= 0.25) capped strictly to 400ms following genuine voiced frames (_frames_since_voiced <= 40). Unvoiced stops (/p/,/t/,/k/) and sibilants/fricatives (/s/,/sh/,/f/) are preserved with 100% fidelity without risk of perpetual open states in continuous ambient noise. - Calibrated Near-Field Cooldown Breakthrough: High-energy near-field speech (
SNR >= onset_snr_db + 8 dB,P >= 0.55) immediately breaks through post-close cooldown, preventing swallowed syllables while maintaining robust immunity against background chatter and mechanical transient clicks.
Getsu v1.2.6 — Voiced-Burst Flutter Detection, Calibration Profiler & Soft Limiter
Getsu v1.2.6 — Voiced-Burst Flutter Detection, Calibration Profiler & Soft Limiter
What's Changed
- Voiced-Burst Flutter Detection: Upgraded chatter flutter detection to track voiced speech frames per burst (
_burst_voiced < 15, speech P >= 0.52) instead of total open time. Three rapid bursts under 150ms within a 3s window smoothly extend the cooldown holdoff to 300ms, automatically reverting after 3s of calm. - Speech-Guarded Ambient Calibration Profiler: Intelligent non-speech median profiler over the initial 50 clean frames (~500ms), bounded within
[-75, -36] dBFSwith safe fallback to-60 dBFSon timeout or high ambient noise. - Noise Floor Tracking Clamps & Rise Cap: Clamped downward floor tracking to
FLOOR_MIN_DB = -75.0 dBFSto prevent baseline drift in long silence. Capped upward floor rise to a maximum of0.5 dB/frameto eliminate sudden acoustic transient distortions. - Clean State Reset on Mute: Implemented
gate.reset()andhpf.reset_state()on mute transitions, cleanly flushing delay buffers, high-pass filter memories, and flutter history to eliminate stale audio leaks on unmute. - Pre-RNNoise Headroom & Post-RNNoise Gain Staging: Removed pre-RNNoise hard clipping and microphone gain multiplier. RNNoise GRU now receives pristine, uncompressed microphone audio with combined gain staging (
output_gain * mic_gain) moved post-RNNoise. - Gentle Soft Limiter Peak Protection: Integrated hyperbolic tangent soft limiting (
soft_limit, threshold = 0.85) to smoothly compress loud acoustic peaks without harsh DAC clipping. - WASAPI Resilient Stream Support: Added
sd.WasapiSettings(auto_convert=True)host-guarded for Windows WASAPI endpoints, preventing stream initialization failures on devices running non-native 48kHz sample rates. - Parity in Voice Test Preview: Aligned GUI Voice Test DSP pipeline with stream engine settings (
close_threshold = 0.52,hangover_ms = 320.0ms, post-RNNoise gain staging, unclipped input, and soft limiter). - Safe Single-Run Version Migration: Added
migrate_v126()to automatically migrate stale defaults toclose_threshold = 0.52andhangover_ms = 320.0mswhile strictly honoring user customizations.
Package Checksums (SHA-256)
Getsu-v1.2.6-Setup.exe:217142CBF0B31358B716BC7197E244218BCDA56C32E54041000F2F9CF897D87C
Getsu v1.2.5 — Gate Re-Arm Cooldown, Smooth Onset & Anti-Chatter Hysteresis
Getsu v1.2.5 — Gate Re-Arm Cooldown, Smooth Onset & Anti-Chatter Hysteresis
🚀 Key Improvements & Bug Fixes
-
Gate Re-Arm Cooldown Timer (150ms Holdoff): Implemented an intelligent 15-frame (150ms) re-arm holdoff in
AdaptiveNoiseGate. When the gate closes upon finishing speech, speculative low-confidence onset triggers are inhibited during the decay window. This 100% eliminates tail clicks, flutter ticks, and decay-abort spikes caused by people talking in the background. High-confidence user speech ($P \ge 0.70$ ) breaks through immediately with zero latency penalty. -
Gentle Syllable Onset (
attack_ms = 28.0ms): Extended the onset envelope time constant to 28.0ms. Frame 1 (silence pre-roll) only reaches0.259($-11.7\text{ dB}$ ) instead of0.437($-7.2\text{ dB}$ ), dropping the initial ambient noise ramp surge by an additional$4.5\text{ dB}$ . Completely eliminates leading-edge harshness after long silence while delivering 90% consonant energy in Frame 2 via the$C^3$ Septic S-curve. -
Anti-Chatter Hysteresis Close (
vad_close_threshold = 0.55): Raised close threshold from 0.45 to 0.55, preventing distant background voices ($P \approx 0.40 - 0.52$ ) from resetting the hangover counter or keeping the gate trapped open. -
Calibrated Onset Discrimination (
vad_onset_snr_db = 10.0 dB,vad_onset_threshold = 0.35,vad_hangover_ms = 280.0ms): Tightened hangover from 360ms to 280ms for prompt sentence endings, while calibrating onset energy rise to require$+10.0\text{ dB}$ over ambient cooling pad noise, cleanly filtering far-field chatter while passing near-field user speech ($+30\text{ dB}$ to$+50\text{ dB}$ SNR). - Seamless Config Migration: Automatically upgrades legacy configurations to v1.2.5 parameters on launch.
📦 Checksums & Verification
- Installer:
Getsu-v1.2.5-Setup.exe(35.71 MB) - SHA-256:
753f71c0d8b2062580c7ffa6131feb5f869c3fe48642892387c6e86f5fe4648e
Getsu v1.2.4 — Septic Smootherstep C3 S-Curve, Noise Gate Floor & Playback Isolation
Getsu v1.2.4 — Septic Smootherstep C3 S-Curve, Noise Gate Floor & Playback Isolation
🚀 Key Improvements & Bug Fixes
-
Septic Smootherstep (
$C^3$ Continuous) Gain Ramping: Upgraded gate transition interpolation inAdaptiveNoiseGateto a 7th-order Septic Smootherstep ($-20t^7 + 70t^6 - 84t^5 + 35t^4$ ) precomputed lookup table. Guarantees zero velocity ($S'=0$ ), zero acceleration ($S''=0$ ), and zero jerk ($S'''=0$ ) at buffer boundaries, reducing boundary entry velocity by$100\times$ compared to Quintic and$15,000\times$ compared to Hann windowing, 100% eliminating transient clicks and pops. -
Continuous Tail Decay (Pop-Free Sentence Endings): Lowered the S-curve transition threshold from
0.005down to1e-5. Ensures late decay frames interpolate sample-by-sample all the way down tofloor_gain, completely eliminating the high-pitched tail tick/pop when you finish speaking. -
Optimized Dynamic Expander Floor (
floor_gain = 0.06/ -24.4 dB): Calibrated gate floor baseline to0.06(-24.4 dB). Keeps wireless headset DAC amplifiers (e.g. MPOW 2.4GHz) reliably energized with a solid$\approx 3\text{ LSB}$ pilot signal across prolonged silence (> 30s) so hardware squelch circuits never enter sleep/standby mode, while maintaining dead digital black silence at -89 dBFS to -94 dBFS (10 dB beneath human hearing and 44 dB beneath Discord's -45 dBFS VAD floor). -
Smooth Syllable Attack Curve (
attack_ms = 22.0ms): Paired the0.06floor with a smoothed22.0msattack envelope. Drops the Frame 1 onset step surge by$9.1\text{ dB}$ (from$+26.4\text{ dB}$ down to$+17.3\text{ dB}$ ), eliminating the leading-edge harshness after prolonged silence and allowing RNNoise recurrent GRU states to converge smoothly while consonants are fully preserved by the 10ms pre-roll lookahead buffer. -
Immunity to Background Chatter (
onset_snr_db = 8.0 dB): Raised onset threshold to require a$2.51\times$ amplitude jump ($6.31\times$ power increase) above ambient cooling pad noise, completely preventing distant room chatter and family members from false-triggering the gate. -
Zero Audio Output Interference: Eliminated all playback device tampering across the entire app lifecycle. Removed background playback guards (
_check_playback_guardinsrc/gui.py) and installer overrides (--ensure-physicalinsetup.iss). Getsu now strictly operates on audio input (eCapture), leaving whatever playback endpoint is chosen by Windows or the user (USB/Wireless headsets, Bluetooth, DACs, HDMI, speakers) 100% untouched during installation, launch, stream start, stream stop, and exit. -
Direct WASAPI Headset Playback for Voice Test: Fixed
get_physical_output_device()insrc/devices.pyto query the active Windows WASAPI default playback device directly. "Hear Myself (11s Voice Test)" now reliably plays clean audio into your active USB/Wireless headset without falling back to motherboard Realtek speakers. -
1-Frame (10ms) Pre-roll Lookahead Buffer: Added an ultra-low-latency 10ms pre-roll lookahead buffer to
AdaptiveNoiseGate. When speech begins after long silence, the gate ramps open smoothly across the pre-roll frame before the syllable peaks, passing your natural voice with 100% full fidelity and zero muffling, zero harshness, and zero perceptible delay. -
Synchronized Voice Test Pipeline: Synced all DSP parameters (
hangover_ms=360,decay_ms=80,attack_ms=22.0,floor_gain=0.06,onset_snr_db=8.0,lookahead=True) in the Voice Test worker with the live stream pipeline. -
Seamless Config Migration: Added automatic legacy config upgrade in
src/config.pythat gracefully migrates olderconfig.jsonfiles to the smooth v1.2.4 gate parameters.
📦 Checksums & Verification
- Installer:
Getsu-v1.2.4-Setup.exe(35.71 MB) - SHA-256:
635e20b0f3985dc72b562493c826be354f0a199ae1ff7cf0dc69580f5284a4d3
Getsu v1.2.3 — Sample-Accurate Slew Ramping & Headset Auto-Selection
What's New in Getsu v1.2.3
🎧 First-Launch Headset Auto-Selection
- Preserved Default Microphone Across Installation: Added capture endpoint backup (
--backup-captureand--restore-capture) in the Inno Setup pipeline andAudioRestore.cs. When VB-Cable virtual driver is installed, Windows no longer defaults to motherboard Realtek audio, safely preserving your physical headset microphone. - GUI Smart 4-Tier Hardware Discovery: Connected
src/gui.pydirectly toauto_select_input_device(), ensuring external USB and wireless gaming headsets (MPOW, HyperX, Razer, Logitech, Corsair, SteelSeries, Yeti, Rode, etc.) are accurately auto-selected on first launch.
🌊 Zero-Click Sample-Accurate Slew Ramping
- Sample-Accurate Slew-Rate Ramping: Replaced block gain stepping with per-sample linear interpolation (
np.linspace) across the 480-sample (10ms) frame during gate opening and closing transitions inAdaptiveNoiseGate. Completely eliminates micro-square-wave transient clicks and harsh digital edges when breaking silence. - Enhanced Consonant Sensitivity (
onset_snr_db = 4.0 dB): Lowered the energy jump threshold from 7.0 dB to 4.0 dB (2x more sensitive to speech power). Soft whispered speech and light unvoiced consonants ("h", "s", "p", "t", "k") open the gate immediately on the very first frame without muffled sound.
Verify Your Download
- File:
Getsu-v1.2.3-Setup.exe - SHA-256:
35F9AB61B7EF0AE93BB983268D76CD669044A15634B8521CD797297364335D37 - Size: 37,441,880 bytes (~35.7 MB)
Getsu v1.2.2 — Dual-Key Onset Trigger & Cold-Start Voice Clarity
What's New in Getsu v1.2.2
🎙️ Speech Onset Clarity & Dead Silence Recovery
- Dual-Key Dynamic Onset Trigger (
AdaptiveNoiseGate): Tracks ambient background noise floor (-60 dB to -20 dB) dynamically during silence. When speaking again after prolonged dead silence, if incoming audio rises+7.0 dBabove ambient room noise and speech probability exceeds0.30, the gate triggers open instantly on the very first 10ms frame. - Eliminates Muffled / Swallowed Consonants: Unvoiced consonant attacks ("h", "s", "p", "t", "k") are preserved with 100% clarity, resolving the "mumbled/swallowed" onset issue caused by neural VAD cold-state latency.
- Cold-Start Gain Snapping (
cold_start_gain = 0.70): Snaps gate gain directly to 0.70 (-3 dB) from absolute silence, guaranteeing full consonant volume while maintaining smooth, zero-click attack transitions. - Extended Natural Speech Hangover (
360.0ms): Broadened gate hold duration to 360ms to preserve soft conversational breathing, natural pauses between thoughts, and subtle word endings without premature gating. - Studio-Calibrated Exponential Decay (
80.0ms): Softly fades out to absolute zero silence within ~350ms, eliminating abrupt gating while keeping mechanical keyboard typing noise filtered out.
Verify Your Download
- File:
Getsu-v1.2.2-Setup.exe - SHA-256:
6370EB1BBA8BBF38CC66EBA5BA72FAF137E2B96ABB24AFB6861499827613C3BF - Size: 37,436,643 bytes (~35.7 MB)