Jivetalking 0.2.2

Changes since 0.2.1

• fix(processor): reject silence candidates with high crest factor (b1a95fb)
• feat(processor): add LA-2A hot-input protection (87ca1be)
• fix(processor): require speech profile for de-esser tuning (9e2b2b4)
• feat(logging): add high_crest_factor tip and improve level tips (4513878)

Installation

Download the appropriate binary for your platform below, make it executable, and move it to your PATH:

# Linux (amd64)
chmod +x jivetalking-linux-amd64
sudo mv jivetalking-linux-amd64 /usr/local/bin/jivetalking

# macOS (Apple Silicon)
chmod +x jivetalking-darwin-arm64
sudo mv jivetalking-darwin-arm64 /usr/local/bin/jivetalking

Checksums

SHA256 checksums are provided below for verification.

Jivetalking 0.2.1

Changes since 0.2.0

• fix(processor): guard against empty searchIntervals in silence detection (fa4c51a)
• perf(processor): reuse spectral metrics to avoid duplicate extraction (56450e0)
• perf(processor): use binary search to find start interval (ef13c5c)
• perf(processor): avoid redundant sorts in silence/room-tone detection (a370edf)
• perf(processor): measure silence and speech regions in single file open (7a765ce)
• perf(processor): reuse AVFrame to reduce allocations in loudnorm passes (7c8afe5)
• refactor(logging): extract three-column metric table helpers (5f940c7)
• refactor(processor): consolidate spectral metric accumulation in analyzer (a07f7dc)
• refactor(processor): extract shared subregion refinement into helper (90b146c)
• refactor(processor): extract astats metadata into helper (ff6a045)
• refactor(processor): extract spectral accumulation into helper method (2354b64)
• refactor: remove legacy helpers and unify numeric helpers (c2476ec)
• docs(readme): add analysis-only mode and clarify data flow (33d67de)
• feat(logging): normalise final spectral metrics for gain comparison (6d5e4ba)
• feat(logging): prefer speech crest factor in analysis display (a94ab74)
• feat(logging): make recording tips speech-aware and refine rules (581cc26)
• feat(processor): prefer speech-specific metrics when available (527363b)
• feat(logging): add recording tips to analysis output (97852aa)
• docs(readme): recommend 10-15s room tone at recording start (eadee28)
• fix(processor): use two-pass silence candidate selection (dc0f14f)
• fix(processor): reject silence candidates with large peak-RMS gap (e87d711)
• feat(processor): add stability scoring for silence candidates (17ca32b)

Installation

Download the appropriate binary for your platform below, make it executable, and move it to your PATH:

# Linux (amd64)
chmod +x jivetalking-linux-amd64
sudo mv jivetalking-linux-amd64 /usr/local/bin/jivetalking

# macOS (Apple Silicon)
chmod +x jivetalking-darwin-arm64
sudo mv jivetalking-darwin-arm64 /usr/local/bin/jivetalking

Checksums

SHA256 checksums are provided below for verification.

Jivetalking 0.2.0

Changes since 0.1.0

• docs(processor): update comments for anlmdn-based noise profile usage (#22) (64c4695)
• fix(cli): handle os.Stdout.Stat error in isTTY (132b8c9)
• feat(analysis): add --analysis-only mode to perform Pass 1 (3856188)
• feat(processor): add crest-factor penalty for silence candidates (905c110)
• chore: remove cruft from integration testing justfile (afe7758)
• fix(processor): average aspectralstats spectral metrics across frames (e0ff1b6)
• docs(readme): rewrite README for clarity and updated workflow (e986552)
• docs(analysis): add Levelator comparison and gap analysis (119a132)
• docs(processor): add spectral metrics reference to AGENTS.md (208d3c8)
• fix(processor): correct SNR penalty application in speech scoring (c710277)
• feat(processor): add voicing density and SNR-aware scoring (599afe3)
• feat(processor): add speech stability scoring (rolloff, flux, voicing) (84b20f5)
• fix(processor): restore la2aCentroidBright threshold (a97381d)
• fix: align spectral interpretation thresholds with reference document (4d1835f)
• refactor(logging): clarify spectral interpretation strings (4b24496)
• feat(processor): lower spectral thresholds for adaptive tuning (a2846a0)
• feat(processor): tune spectral thresholds; add metrics reference (8fb150e)
• feat(processor): add adeclick click/pop repair after loudnorm (7528ebe)
• fix(processor): enforce minimum 10ms attack for DS201 gate (a80d892)
• refactor(processor): migrate limiter from UREI 1176 to CBS Volumax style (c17f48a)
• feat(processor): implement aggression-based DS201 gate thresholding (f27adc0)
• refactor(processor): remove breath-reduction mode and CLI flag (f90180d)
• feat(processor): refine long speech regions to golden sub-window (a613feb)
• fix(analyzer): require significant regression before stopping search (6849db7)
• refactor(audio): remove DC1Declick filter from processing pipeline (bcd22aa)
• feat(processor): add breath reduction enhancement to DS201 gate filter (c0b8831)
• feat(processor): make NoiseRemove compand parameters adaptive to measured noise floor (832dfbc)
• docs(spectral): add authoritative references to spectral metric interpretations (7b7c26f)
• refactor(logging): calibrate spectral interpretation thresholds for speech-only analysis (81ddd5e)
• fix(processor): address nil pointer risks and formatting edge cases (0c539db)
• fix(analyzer): correct region measurement capture and display (22b739e)
• feat(logging): expand report tables with full three-stage region metrics (5a53bb6)
• feat(processor): wire region measurement capture into Pass 4 normalisation stage (17e9c2f)
• feat(processor): wire region measurement capture into Pass 2 filtering stage (cbd2135)
• feat(analyzer): implement region measurement capture for processing stages (f86e250)
• refactor(processor): upgrade silence measurement to full candidate metrics (eb1e238)
• fix: ensure interval start is found after start index (536273c)
• feat(analyzer): expand candidate metrics to include all interval measurements (f81a382)
• feat(analyzer): add speech candidate detection for adaptive tuning (08c8027)

Installation

Download the appropriate binary for your platform below, make it executable, and move it to your PATH:

# Linux (amd64)
chmod +x jivetalking-linux-amd64
sudo mv jivetalking-linux-amd64 /usr/local/bin/jivetalking

# macOS (Apple Silicon)
chmod +x jivetalking-darwin-arm64
sudo mv jivetalking-darwin-arm64 /usr/local/bin/jivetalking

Checksums

SHA256 checksums are provided below for verification.

Jivetalking 0.1.0

Initial Release

• docs(readme): add CLI usage, contributing, and license sections (ed3836a)
• feat(cli): add --debug flag to conditionally enable debug logging (ed04816)
• docs: consolidate AI instructions into AGENTS.md (1970029)
• chore: update ffmpeg-statigo to v8.0.1.3 (5733364)
• feat(release): add automated release workflow and version management (b92462d)
• refactor(processor): migrate noise removal to anlmdn with fixed compand (55ee600)
• feat(testdata): add audio analysis commands for before/after comparison (07fba86)
• feat(mains): add automatic mains frequency detection (4168567)
• refactor(processor): remove dead code and unused parameters (1cd85bc)
• refactor(analyzer): replace pre-scan with interval-based noise floor estimation (9fb399a)
• refactor(processor): remove DefaultFilterOrder alias (47478b4)
• refactor(processor): inline processWithStandardFilters into processWithFilters (d79add2)
• refactor(processor): remove unused FilterSilenceDetect infrastructure (f4d6b47)
• refactor(cli): remove unused --config flag (3d12f9f)
• refactor(cli): remove unused TUI-predecessor functions (3df4e94)
• fix(ui): improve Pass 3/4 progress bar update frequency (23d2c60)
• refactor(build): move presenter tests to separate gitignored justfile (879b9b6)
• docs: update README.md (57d955a)
• feat(logging): add noise reduction delta tracking to reports (31d1cb9)
• fix(normalise): increase limiter safety margin for ISP compliance (b700af0)
• perf(analyzer): eliminate file re-reads in silence detection (c7e2ee2)
• feat(analyzer): add golden sub-region refinement for noise profiling (7be787a)
• fix(dns1500): cap compand expansion at 15 dB to prevent pumping (2fc9f3e)
• docs: update documentation for four-pass pipeline and -18 LUFS (7916cbf)
• refactor(processor): change default LUFS target from -16 to -18 (ebd6bd8)
• fix(processor): clamp compand threshold for clean recordings (63beb4d)
• fix(dns1500): calculate compand from effective floor after afftdn (e60477a)
• feat(dns1500): add compand filter for residual noise suppression (e823249)
• refactor(justfile): add episode variable for test file paths (699b60a)
• fix(normalise): clamp limiter ceiling to alimiter minimum (5ce98ce)
• feat(normalise): add adaptive peak limiter for loudnorm linear mode (1361f92)
• refactor(processor): remove makeup gain from Pass 2 filters (642638c)
• refactor(logging): restructure report into consolidated tabular format (9f4271f)
• feat(processor): replace gain normalisation with loudnorm filter (2feac64)
• feat(denoise): improve arnndn denoising system with multiple models (8934501)
• feat(processor): add Pass 3 normalisation with UREI 1176 limiter (6433b07)
• feat(processor): implement UREI 1176-inspired adaptive limiter (1e956da)
• chore(deps): bump actions/upload-artifact from 5 to 6 (8a76416)
• refactor(processor): use consistent mono conversion filter (a02b5ca)
• refactor(processor): extract sample format switch into reusable function (5b41f01)
• refactor(processor): extract silence detection magic numbers to constants (25ba905)
• refactor(processor): consolidate dB/linear conversion functions (4fc8701)
• refactor(logging): consolidate spectral metrics formatting in reports (6de5325)
• refactor(processor): centralise filter builders via package-level registry (cac44d9)
• refactor(processor): remove unused calculateDS201GateRange function (e47839f)
• refactor(processor): consolidate metadataAccumulators via embedding (d31ead1)
• refactor(processor): consolidate AudioMeasurements and OutputMeasurements structs (6636748)
• refactor(analyzer): consolidate spectral metadata extraction (5e47456)
• refactor(processor): remove obsolete WAV extraction from silence detection (f0d4fa8)
• test(processor): replace testdata files with synthetic audio (0c868b2)
• feat(processor): add DNS-1500 noise reduction with voice protection (c3e1100)
• refactor(processor): remove speechnorm and dynaudnorm filters (139d1c2)
• refactor(processor): consolidate gain staging in LA2A compressor (7a17bf3)
• refactor(processor): improve silence detection with stop-on-regression selection (aa91c61)
• feat(analyzer): expand astats metrics and add interval-based silence detection (72b3cb7)
• feat(processor): add CEDAR DC-1 inspired declicker filter (e17188f)
• fix(gate): add gentle mode for very quiet low-LRA recordings (557a9ec)
• refactor(processor): migrate DolbySR filter from afftdn to mcompand (ea398a2)
• refactor(audio): replace compressor makeup gain calculation (757e403)
• feat(noise-gate): extend gate release time for low LRA audio (9394e68)
• docs: create comprehensive project documentation with detailed filter descriptions (424a676)
• feat(processor): enables DS201 adaptive gate makeup gain (4dfd7b7)
• fix(adaptive): correct calculation for adaptive noise reduction (3c129bf)
• feat(processor): add multi-band Dolby SR-inspired noise reduction (4f2a138)
• chore(processor): remove standalone afftdn filter (721a024)
• feat(analyzer): add adaptive silence detection with pre-scan and segmentation (3b40a34)
• refactor(processor): remove obsolete afftdn_simple filter (e973705)
• fix(processor): align DS201LowPass with spec, show reason when disabled (0aa3fc5)
• feat(processor): add Dolby SR-inspired single-stage denoise filter (51109d4)
• chore(processor): remove obsolete BleedGate filter (a933efb)
• feat(ds201-lowpass): enable ultrasonic cleanup for all speech content (da8ff35)
• feat(processor): add content-aware DS201 lowpass filter tuning (c3ce4f1)
• refactor: implement DS201-inspired frequency-conscious gating (a00a5fe)
• fix(processor): correct LA-2A compressor adaptive tuning (ebe527f)
• feat(processor): add adaptive noise type selection for afftdn_simple (ce6b7cb)
• refactor(arnndn): simplify adaptive tuning and disable by default (895a7cf)
• feat(processor): implement comprehensive adaptive gate tuning (4c5ef32)
• test(processor): decouple tests from DefaultFilterConfig (2f0fb9b)
• feat(processor): add highpass tunables for warm voice protection (c02202a)
• refactor(processor): replace Q-based bandreject with Hz-based width and TDII transform (41305eb)
• fix(processor): disable highpass for voices with skewness > 1.0 (94d1c82)
• feat(analyzer): add spectral analysis capabilities and detailed report generation (80a454d)
• feat(processor): add infrastructure filters to unified filter chain (b2bb46d)
• feat(report): enhance report formatting to show input-output comparisons (afb4385)
• feat(audio): enable true peak measurement in ebur128 filter (f4d7360)
• fix(processor): normalize silence region measurement between analysis passes (a8f4ec8)
• feat(report): add zero crossings rate and max difference metrics to pass 2 (1f0b099)
• fix(processor): reorder aformat filter after output analysis (5dff604)
• refactor(filters): disable most audio filters during recalibration (03b13d7)
• chore: remove deprecated ANLM denoiser filter (e3b1e66)
• refactor(logging): replace ad-hoc report sections with ordered filter chain output (4ce9901)
• wip(processor): add BleedGate to suppress amplified headphone bleed (44e6abe)
• feat(processor): implement profile-based afftdn noise removal (649f757)
• feat(processor): implement adaptive noise removal pipeline (5ff5594)
• feat(analyzer): add extended astats measurements for adaptive processing (b04cd52)
• refactor: simplify audio processor logic and main entry point (848c6a4)
• test(processor): add speechnorm denoise activation tests (7e3dfb0)
• test(processor): add config sanitization and clamp edge case tests (c62e20d)
• test(processor): add filter spec generation and dB conversion tests (bc14d9d)
• test(processor): add gate threshold tuning tests (cb1257d)
• test(processor): add de-esser intensity tuning tests (96d98a2)
• test(adaptive): add unit tests for noise reduction LUFS scaling (0ad1308)
• test(adaptive): add unit tests for highpass frequency tuning (55e1cf8)
• chore(processor): remove vestigial debug output comments (2d11a03)
• refactor(processor): extract Encoder type to separate file (8375104)
• refactor: standardise error messages with file path context (c06e386)
• refactor(logging): remove unused abs() function (2427911)
• refactor(processor): extract filter graph setup into shared helper (49ec79e)
• refactor(adaptive): consolidate magic numbers into named constants (ecedd87)
• refactor(analyzer): remove unused parameters from createAnalysisFilterGraph (be99231)
• refactor(analyzer): extract duplicated metadata extraction to helper (66b39ed)
• refactor(processor): extract adaptive filter tuning to dedicated module (4303d89)
• refactor(processor): modularise filter chain configuration (586db80)
• perf(processor): remove debug output from filter chain builder (5b6f67c)
• perf(analyzer): cache metadata key CStr allocations in hot loop (f9c7c4f)
• feat: add CLI application in cmd/jivetalking (433cafb)
• test(processor): improve analyzer test performance and feedback (0a68d67)
• test(processor): skip analyzer test when testdata unavailable (928dbbf)
• build(ffmpeg): migrate from ffmpeg-go to ffmpeg-statigo (31ff978)
• fix(processor): update tests for AudioMeasurements type rename (c345ee5)
• docs: add loudness normalization analysis and strategy documentation (071b488)
• feat: adds RNN and Non-Local Means denoise filters (a6b36f7)
• refactor(audio): replace loudnorm with dynaudnorm for normalization (301efbe)
• chore: simplify test tasks by reading log files directly (09aaae8)
• refactor: improve progress tracking and remove debug logging (e460bd3)
• feat(processor): add adaptive dynamic normalization with noise threshold control (ccb1bf8)
• feat(audio): improve noise floor detection with RMS_trough (0eb18c3)
• refactor(audio): replace loudnorm with ebur128 for analysis phase (203e5ab)
• feat(audio): add dynaudnorm as alternative to loudnorm (9545f60)
• chore(deps): bump actions/checkout from 5 to 6 (2765a77)
• chore(deps): bump github.com/alecthomas/kong from 1.12.1 to 1.13.0 (a1b321e)
• fix(processor): update audio...