pvx Algorithm Limitations and Applicability

Generated from commit 35e9761 (commit date: 2026-05-25T08:14:42-04:00).

This document summarizes assumptions, likely failure modes, and practical exclusion cases for each algorithm group and algorithm module.

Group-Level Summary

Group	Assumptions	Failure Modes	When Not To Use
analysis_qa_and_automation	Feature extraction settings align with domain (speech vs music etc.).	False positives/negatives under domain shift.	Avoid treating single metrics as absolute quality verdicts.
creative_spectral_effects	Spectral manipulations are desired even with timbral coloration.	Can introduce intentional but strong coloration or temporal artifacts.	Avoid for transparent restoration/mastering paths.
denoise_and_restoration	Noise/artifacts are distinguishable from desired signal statistics.	Over-reduction can remove detail and create modulation artifacts.	Avoid high reduction settings on sparse acoustic sources without auditioning.
dereverb_and_room_correction	Late reverberation is separable from direct content under chosen model.	Speech/music clarity can drop if early reflections are over-suppressed.	Avoid strong dereverb when room character is part of artistic intent.
dynamics_and_loudness	Program dynamics fit compressor/limiter time constants and thresholds.	Pumping, breathing, or overs if thresholds and release are mis-set.	Avoid applying multiple strong dynamics stages without gain staging checks.
granular_and_modulation	Grain and modulation rates are musically matched to source texture.	Incoherent grain scheduling can produce choppiness or blur.	Avoid dense granular settings on speech intelligibility-critical content.
pitch_detection_and_tracking	F0 evidence is strong in the selected analysis band and frame size.	Octave errors and voicing flips under heavy noise/polyphony.	Avoid as the sole control signal for dense polyphonic mixtures.
retune_and_intonation	Detected notes map cleanly to intended tonal center/scale.	Over-correction can flatten expressive vibrato or slides.	Avoid aggressive correction when preserving natural micro-intonation is required.
separation_and_decomposition	Sources have partially separable spectral or statistical structure.	Component bleeding and musical noise under overlap or model mismatch.	Avoid expecting perfect stems from strongly correlated or co-modulated sources.
spatial_and_multichannel	Channel geometry/order and timing metadata are correct.	Spatial collapse, combing, or localization bias from misalignment.	Avoid blind spatial processing when channel order/calibration is unknown.
spectral_time_frequency_transforms	Transform parameterization matches target time-frequency structure.	Incorrect parameterization can smear events or over-fragment spectra.	Avoid default settings for highly nonstationary signals without tuning.
time_scale_and_pitch_core	Frames are locally quasi-stationary and harmonic evolution is reasonably smooth.	High-ratio stretch can introduce phasiness and blurred transients.	Avoid for extreme percussive-only material when attack realism is critical.

analysis_qa_and_automation

Algorithm ID	Assumptions	Failure Modes	When Not To Use
analysis_qa_and_automation.auto_parameter_tuning_bayesian_optimization	Feature extraction settings align with domain (speech vs music etc.).	False positives/negatives under domain shift.	Avoid treating single metrics as absolute quality verdicts.
analysis_qa_and_automation.batch_preset_recommendation_based_on_source_features	Feature extraction settings align with domain (speech vs music etc.).	False positives/negatives under domain shift.	Avoid treating single metrics as absolute quality verdicts.
analysis_qa_and_automation.clip_hum_buzz_artifact_detection	Feature extraction settings align with domain (speech vs music etc.).	False positives/negatives under domain shift.	Avoid treating single metrics as absolute quality verdicts.
analysis_qa_and_automation.key_chord_detection	Feature extraction settings align with domain (speech vs music etc.).	False positives/negatives under domain shift.	Avoid treating single metrics as absolute quality verdicts.
analysis_qa_and_automation.onset_beat_downbeat_tracking	Feature extraction settings align with domain (speech vs music etc.).	False positives/negatives under domain shift.	Avoid treating single metrics as absolute quality verdicts.
analysis_qa_and_automation.pesq_stoi_visqol_quality_metrics	Feature extraction settings align with domain (speech vs music etc.).	False positives/negatives under domain shift.	Avoid treating single metrics as absolute quality verdicts.
analysis_qa_and_automation.silence_speech_music_classifiers	Feature extraction settings align with domain (speech vs music etc.).	False positives/negatives under domain shift.	Avoid treating single metrics as absolute quality verdicts.
analysis_qa_and_automation.structure_segmentation_verse_chorus_sections	Feature extraction settings align with domain (speech vs music etc.).	False positives/negatives under domain shift.	Avoid treating single metrics as absolute quality verdicts.

creative_spectral_effects

Algorithm ID	Assumptions	Failure Modes	When Not To Use
creative_spectral_effects.cross_synthesis_vocoder	Spectral manipulations are desired even with timbral coloration.	Can introduce intentional but strong coloration or temporal artifacts.	Avoid for transparent restoration/mastering paths.
creative_spectral_effects.formant_painting_warping	Spectral manipulations are desired even with timbral coloration.	Can introduce intentional but strong coloration or temporal artifacts.	Avoid for transparent restoration/mastering paths.
creative_spectral_effects.phase_randomization_textures	Spectral manipulations are desired even with timbral coloration.	Can introduce intentional but strong coloration or temporal artifacts.	Avoid for transparent restoration/mastering paths.
creative_spectral_effects.resonator_filterbank_morphing	Spectral manipulations are desired even with timbral coloration.	Can introduce intentional but strong coloration or temporal artifacts.	Avoid for transparent restoration/mastering paths.
creative_spectral_effects.spectral_blur_smear	Spectral manipulations are desired even with timbral coloration.	Can introduce intentional but strong coloration or temporal artifacts.	Avoid for transparent restoration/mastering paths.
creative_spectral_effects.spectral_contrast_exaggeration	Spectral manipulations are desired even with timbral coloration.	Can introduce intentional but strong coloration or temporal artifacts.	Avoid for transparent restoration/mastering paths.
creative_spectral_effects.spectral_convolution_effects	Spectral manipulations are desired even with timbral coloration.	Can introduce intentional but strong coloration or temporal artifacts.	Avoid for transparent restoration/mastering paths.
creative_spectral_effects.spectral_freeze_banks	Spectral manipulations are desired even with timbral coloration.	Can introduce intentional but strong coloration or temporal artifacts.	Avoid for transparent restoration/mastering paths.

denoise_and_restoration

Algorithm ID	Assumptions	Failure Modes	When Not To Use
denoise_and_restoration.declick_decrackle_median_wavelet_interpolation	Noise/artifacts are distinguishable from desired signal statistics. Noise model should be representative of observed noise floor.	Over-reduction can remove detail and create modulation artifacts. Mismatched noise estimate leaves residue or damages detail.	Avoid high reduction settings on sparse acoustic sources without auditioning. Avoid static settings on rapidly varying nonstationary noise.
denoise_and_restoration.declip_via_sparse_reconstruction	Noise/artifacts are distinguishable from desired signal statistics. Noise model should be representative of observed noise floor.	Over-reduction can remove detail and create modulation artifacts. Mismatched noise estimate leaves residue or damages detail.	Avoid high reduction settings on sparse acoustic sources without auditioning. Avoid static settings on rapidly varying nonstationary noise.
denoise_and_restoration.diffusion_based_speech_audio_denoise	Noise/artifacts are distinguishable from desired signal statistics. Noise model should be representative of observed noise floor.	Over-reduction can remove detail and create modulation artifacts. Mismatched noise estimate leaves residue or damages detail.	Avoid high reduction settings on sparse acoustic sources without auditioning. Avoid static settings on rapidly varying nonstationary noise.
denoise_and_restoration.log_mmse	Noise/artifacts are distinguishable from desired signal statistics. Noise model should be representative of observed noise floor.	Over-reduction can remove detail and create modulation artifacts. Mismatched noise estimate leaves residue or damages detail.	Avoid high reduction settings on sparse acoustic sources without auditioning. Avoid static settings on rapidly varying nonstationary noise.
denoise_and_restoration.minimum_statistics_noise_tracking	Noise/artifacts are distinguishable from desired signal statistics. Noise model should be representative of observed noise floor.	Over-reduction can remove detail and create modulation artifacts. Mismatched noise estimate leaves residue or damages detail.	Avoid high reduction settings on sparse acoustic sources without auditioning. Avoid static settings on rapidly varying nonstationary noise.
denoise_and_restoration.mmse_stsa	Noise/artifacts are distinguishable from desired signal statistics. Noise model should be representative of observed noise floor.	Over-reduction can remove detail and create modulation artifacts. Mismatched noise estimate leaves residue or damages detail.	Avoid high reduction settings on sparse acoustic sources without auditioning. Avoid static settings on rapidly varying nonstationary noise.
denoise_and_restoration.rnnoise_style_denoiser	Noise/artifacts are distinguishable from desired signal statistics. Noise model should be representative of observed noise floor.	Over-reduction can remove detail and create modulation artifacts. Mismatched noise estimate leaves residue or damages detail.	Avoid high reduction settings on sparse acoustic sources without auditioning. Avoid static settings on rapidly varying nonstationary noise.
denoise_and_restoration.wiener_denoising	Noise/artifacts are distinguishable from desired signal statistics. Noise model should be representative of observed noise floor.	Over-reduction can remove detail and create modulation artifacts. Mismatched noise estimate leaves residue or damages detail.	Avoid high reduction settings on sparse acoustic sources without auditioning. Avoid static settings on rapidly varying nonstationary noise.

dereverb_and_room_correction

Algorithm ID	Assumptions	Failure Modes	When Not To Use
dereverb_and_room_correction.blind_deconvolution_dereverb	Late reverberation is separable from direct content under chosen model. Reverberation tail is assumed more diffuse than direct content.	Speech/music clarity can drop if early reflections are over-suppressed. Over-suppression can thin tonal body and ambience.	Avoid strong dereverb when room character is part of artistic intent. Avoid for intentionally wet effects unless mix preservation is planned.
dereverb_and_room_correction.drr_guided_dereverb	Late reverberation is separable from direct content under chosen model. Reverberation tail is assumed more diffuse than direct content.	Speech/music clarity can drop if early reflections are over-suppressed. Over-suppression can thin tonal body and ambience.	Avoid strong dereverb when room character is part of artistic intent. Avoid for intentionally wet effects unless mix preservation is planned.
dereverb_and_room_correction.late_reverb_suppression_via_coherence	Late reverberation is separable from direct content under chosen model. Reverberation tail is assumed more diffuse than direct content.	Speech/music clarity can drop if early reflections are over-suppressed. Over-suppression can thin tonal body and ambience.	Avoid strong dereverb when room character is part of artistic intent. Avoid for intentionally wet effects unless mix preservation is planned.
dereverb_and_room_correction.multi_band_adaptive_deverb	Late reverberation is separable from direct content under chosen model. Reverberation tail is assumed more diffuse than direct content.	Speech/music clarity can drop if early reflections are over-suppressed. Over-suppression can thin tonal body and ambience.	Avoid strong dereverb when room character is part of artistic intent. Avoid for intentionally wet effects unless mix preservation is planned.
dereverb_and_room_correction.neural_dereverb_module	Late reverberation is separable from direct content under chosen model. Model priors assume training-like signal statistics. Reverberation tail is assumed more diffuse than direct content.	Speech/music clarity can drop if early reflections are over-suppressed. Generalization gaps can produce unstable artifacts. Over-suppression can thin tonal body and ambience.	Avoid strong dereverb when room character is part of artistic intent. Avoid fully unattended use on out-of-domain material. Avoid for intentionally wet effects unless mix preservation is planned.
dereverb_and_room_correction.room_impulse_inverse_filtering	Late reverberation is separable from direct content under chosen model. Reverberation tail is assumed more diffuse than direct content.	Speech/music clarity can drop if early reflections are over-suppressed. Over-suppression can thin tonal body and ambience.	Avoid strong dereverb when room character is part of artistic intent. Avoid for intentionally wet effects unless mix preservation is planned.
dereverb_and_room_correction.spectral_decay_subtraction	Late reverberation is separable from direct content under chosen model. Reverberation tail is assumed more diffuse than direct content.	Speech/music clarity can drop if early reflections are over-suppressed. Over-suppression can thin tonal body and ambience.	Avoid strong dereverb when room character is part of artistic intent. Avoid for intentionally wet effects unless mix preservation is planned.
dereverb_and_room_correction.wpe_dereverberation	Late reverberation is separable from direct content under chosen model. Reverberation tail is assumed more diffuse than direct content.	Speech/music clarity can drop if early reflections are over-suppressed. Over-suppression can thin tonal body and ambience.	Avoid strong dereverb when room character is part of artistic intent. Avoid for intentionally wet effects unless mix preservation is planned.

dynamics_and_loudness

Algorithm ID	Assumptions	Failure Modes	When Not To Use
dynamics_and_loudness.ebu_r128_normalization	Program dynamics fit compressor/limiter time constants and thresholds.	Pumping, breathing, or overs if thresholds and release are mis-set.	Avoid applying multiple strong dynamics stages without gain staging checks.
dynamics_and_loudness.itu_bs_1770_loudness_measurement_gating	Program dynamics fit compressor/limiter time constants and thresholds.	Pumping, breathing, or overs if thresholds and release are mis-set.	Avoid applying multiple strong dynamics stages without gain staging checks.
dynamics_and_loudness.lufs_target_mastering_chain	Program dynamics fit compressor/limiter time constants and thresholds.	Pumping, breathing, or overs if thresholds and release are mis-set.	Avoid applying multiple strong dynamics stages without gain staging checks.
dynamics_and_loudness.multi_band_compression	Program dynamics fit compressor/limiter time constants and thresholds.	Pumping, breathing, or overs if thresholds and release are mis-set.	Avoid applying multiple strong dynamics stages without gain staging checks.
dynamics_and_loudness.spectral_dynamics_bin_wise_compressor_expander	Program dynamics fit compressor/limiter time constants and thresholds.	Pumping, breathing, or overs if thresholds and release are mis-set.	Avoid applying multiple strong dynamics stages without gain staging checks.
dynamics_and_loudness.transient_shaping	Program dynamics fit compressor/limiter time constants and thresholds.	Pumping, breathing, or overs if thresholds and release are mis-set.	Avoid applying multiple strong dynamics stages without gain staging checks.
dynamics_and_loudness.true_peak_limiting	Program dynamics fit compressor/limiter time constants and thresholds.	Pumping, breathing, or overs if thresholds and release are mis-set.	Avoid applying multiple strong dynamics stages without gain staging checks.
dynamics_and_loudness.upward_compression	Program dynamics fit compressor/limiter time constants and thresholds.	Pumping, breathing, or overs if thresholds and release are mis-set.	Avoid applying multiple strong dynamics stages without gain staging checks.

granular_and_modulation

Algorithm ID	Assumptions	Failure Modes	When Not To Use
granular_and_modulation.am_fm_ring_modulation_blocks	Grain and modulation rates are musically matched to source texture.	Incoherent grain scheduling can produce choppiness or blur.	Avoid dense granular settings on speech intelligibility-critical content.
granular_and_modulation.envelope_followed_modulation_routing	Grain and modulation rates are musically matched to source texture.	Incoherent grain scheduling can produce choppiness or blur.	Avoid dense granular settings on speech intelligibility-critical content.
granular_and_modulation.formant_lfo_modulation	Grain and modulation rates are musically matched to source texture.	Incoherent grain scheduling can produce choppiness or blur.	Avoid dense granular settings on speech intelligibility-critical content.
granular_and_modulation.freeze_grain_morphing	Grain and modulation rates are musically matched to source texture.	Incoherent grain scheduling can produce choppiness or blur.	Avoid dense granular settings on speech intelligibility-critical content.
granular_and_modulation.grain_cloud_pitch_textures	Grain and modulation rates are musically matched to source texture.	Incoherent grain scheduling can produce choppiness or blur.	Avoid dense granular settings on speech intelligibility-critical content.
granular_and_modulation.granular_time_stretch_engine	Grain and modulation rates are musically matched to source texture.	Incoherent grain scheduling can produce choppiness or blur.	Avoid dense granular settings on speech intelligibility-critical content.
granular_and_modulation.rhythmic_gate_stutter_quantizer	Grain and modulation rates are musically matched to source texture.	Incoherent grain scheduling can produce choppiness or blur.	Avoid dense granular settings on speech intelligibility-critical content.
granular_and_modulation.spectral_tremolo	Grain and modulation rates are musically matched to source texture.	Incoherent grain scheduling can produce choppiness or blur.	Avoid dense granular settings on speech intelligibility-critical content.

pitch_detection_and_tracking

Algorithm ID	Assumptions	Failure Modes	When Not To Use
pitch_detection_and_tracking.crepe_style_neural_f0	F0 evidence is strong in the selected analysis band and frame size. Model priors assume training-like signal statistics.	Octave errors and voicing flips under heavy noise/polyphony. Generalization gaps can produce unstable artifacts.	Avoid as the sole control signal for dense polyphonic mixtures. Avoid fully unattended use on out-of-domain material.
pitch_detection_and_tracking.harmonic_product_spectrum_hps	F0 evidence is strong in the selected analysis band and frame size.	Octave errors and voicing flips under heavy noise/polyphony.	Avoid as the sole control signal for dense polyphonic mixtures.
pitch_detection_and_tracking.pyin	F0 evidence is strong in the selected analysis band and frame size.	Octave errors and voicing flips under heavy noise/polyphony.	Avoid as the sole control signal for dense polyphonic mixtures.
pitch_detection_and_tracking.rapt	F0 evidence is strong in the selected analysis band and frame size.	Octave errors and voicing flips under heavy noise/polyphony.	Avoid as the sole control signal for dense polyphonic mixtures.
pitch_detection_and_tracking.subharmonic_summation	F0 evidence is strong in the selected analysis band and frame size.	Octave errors and voicing flips under heavy noise/polyphony.	Avoid as the sole control signal for dense polyphonic mixtures.
pitch_detection_and_tracking.swipe	F0 evidence is strong in the selected analysis band and frame size.	Octave errors and voicing flips under heavy noise/polyphony.	Avoid as the sole control signal for dense polyphonic mixtures.
pitch_detection_and_tracking.viterbi_smoothed_pitch_contour_tracking	F0 evidence is strong in the selected analysis band and frame size.	Octave errors and voicing flips under heavy noise/polyphony.	Avoid as the sole control signal for dense polyphonic mixtures.
pitch_detection_and_tracking.yin	F0 evidence is strong in the selected analysis band and frame size.	Octave errors and voicing flips under heavy noise/polyphony.	Avoid as the sole control signal for dense polyphonic mixtures.

retune_and_intonation

Algorithm ID	Assumptions	Failure Modes	When Not To Use
retune_and_intonation.adaptive_intonation_context_sensitive_intervals	Detected notes map cleanly to intended tonal center/scale. Pitch trajectory estimates should be continuous enough for retuning.	Over-correction can flatten expressive vibrato or slides. Fast F0 jumps can cause audible stepping.	Avoid aggressive correction when preserving natural micro-intonation is required. Avoid high-strength retune on breath/noise segments.
retune_and_intonation.chord_aware_retuning	Detected notes map cleanly to intended tonal center/scale. Pitch trajectory estimates should be continuous enough for retuning.	Over-correction can flatten expressive vibrato or slides. Fast F0 jumps can cause audible stepping.	Avoid aggressive correction when preserving natural micro-intonation is required. Avoid high-strength retune on breath/noise segments.
retune_and_intonation.just_intonation_mapping_per_key_center	Detected notes map cleanly to intended tonal center/scale. Pitch trajectory estimates should be continuous enough for retuning.	Over-correction can flatten expressive vibrato or slides. Fast F0 jumps can cause audible stepping.	Avoid aggressive correction when preserving natural micro-intonation is required. Avoid high-strength retune on breath/noise segments.
retune_and_intonation.key_aware_retuning_with_confidence_weighting	Detected notes map cleanly to intended tonal center/scale. Pitch trajectory estimates should be continuous enough for retuning.	Over-correction can flatten expressive vibrato or slides. Fast F0 jumps can cause audible stepping.	Avoid aggressive correction when preserving natural micro-intonation is required. Avoid high-strength retune on breath/noise segments.
retune_and_intonation.portamento_aware_retune_curves	Detected notes map cleanly to intended tonal center/scale. Pitch trajectory estimates should be continuous enough for retuning.	Over-correction can flatten expressive vibrato or slides. Fast F0 jumps can cause audible stepping.	Avoid aggressive correction when preserving natural micro-intonation is required. Avoid high-strength retune on breath/noise segments.
retune_and_intonation.scala_mts_scale_import_and_quantization	Detected notes map cleanly to intended tonal center/scale. Pitch trajectory estimates should be continuous enough for retuning.	Over-correction can flatten expressive vibrato or slides. Fast F0 jumps can cause audible stepping.	Avoid aggressive correction when preserving natural micro-intonation is required. Avoid high-strength retune on breath/noise segments.
retune_and_intonation.time_varying_cents_maps	Detected notes map cleanly to intended tonal center/scale. Pitch trajectory estimates should be continuous enough for retuning.	Over-correction can flatten expressive vibrato or slides. Fast F0 jumps can cause audible stepping.	Avoid aggressive correction when preserving natural micro-intonation is required. Avoid high-strength retune on breath/noise segments.
retune_and_intonation.vibrato_preserving_correction	Detected notes map cleanly to intended tonal center/scale. Pitch trajectory estimates should be continuous enough for retuning.	Over-correction can flatten expressive vibrato or slides. Fast F0 jumps can cause audible stepping.	Avoid aggressive correction when preserving natural micro-intonation is required. Avoid high-strength retune on breath/noise segments.

separation_and_decomposition

Algorithm ID	Assumptions	Failure Modes	When Not To Use
separation_and_decomposition.demucs_style_stem_separation_backend	Sources have partially separable spectral or statistical structure.	Component bleeding and musical noise under overlap or model mismatch.	Avoid expecting perfect stems from strongly correlated or co-modulated sources.
separation_and_decomposition.ica_bss_for_multichannel_stems	Sources have partially separable spectral or statistical structure.	Component bleeding and musical noise under overlap or model mismatch.	Avoid expecting perfect stems from strongly correlated or co-modulated sources.
separation_and_decomposition.nmf_decomposition	Sources have partially separable spectral or statistical structure.	Component bleeding and musical noise under overlap or model mismatch.	Avoid expecting perfect stems from strongly correlated or co-modulated sources.
separation_and_decomposition.probabilistic_latent_component_separation	Sources have partially separable spectral or statistical structure.	Component bleeding and musical noise under overlap or model mismatch.	Avoid expecting perfect stems from strongly correlated or co-modulated sources.
separation_and_decomposition.rpca_hpss	Sources have partially separable spectral or statistical structure.	Component bleeding and musical noise under overlap or model mismatch.	Avoid expecting perfect stems from strongly correlated or co-modulated sources.
separation_and_decomposition.sinusoidal_residual_transient_decomposition	Sources have partially separable spectral or statistical structure.	Component bleeding and musical noise under overlap or model mismatch.	Avoid expecting perfect stems from strongly correlated or co-modulated sources.
separation_and_decomposition.tensor_decomposition_cp_tucker	Sources have partially separable spectral or statistical structure.	Component bleeding and musical noise under overlap or model mismatch.	Avoid expecting perfect stems from strongly correlated or co-modulated sources.
separation_and_decomposition.u_net_vocal_accompaniment_split	Sources have partially separable spectral or statistical structure.	Component bleeding and musical noise under overlap or model mismatch.	Avoid expecting perfect stems from strongly correlated or co-modulated sources.

spatial_and_multichannel

Algorithm ID	Assumptions	Failure Modes	When Not To Use
spatial_and_multichannel.binaural_itd_ild_synthesis	Channel geometry/order and timing metadata are correct.	Spatial collapse, combing, or localization bias from misalignment.	Avoid blind spatial processing when channel order/calibration is unknown.
spatial_and_multichannel.binaural_motion_trajectory_designer	Channel geometry/order and timing metadata are correct.	Spatial collapse, combing, or localization bias from misalignment.	Avoid blind spatial processing when channel order/calibration is unknown.
spatial_and_multichannel.coherence_based_dereverb_multichannel	Channel geometry/order and timing metadata are correct. Reverberation tail is assumed more diffuse than direct content.	Spatial collapse, combing, or localization bias from misalignment. Over-suppression can thin tonal body and ambience.	Avoid blind spatial processing when channel order/calibration is unknown. Avoid for intentionally wet effects unless mix preservation is planned.
spatial_and_multichannel.cross_channel_click_pop_repair	Channel geometry/order and timing metadata are correct.	Spatial collapse, combing, or localization bias from misalignment.	Avoid blind spatial processing when channel order/calibration is unknown.
spatial_and_multichannel.dbap_distance_based_amplitude_panning	Channel geometry/order and timing metadata are correct.	Spatial collapse, combing, or localization bias from misalignment.	Avoid blind spatial processing when channel order/calibration is unknown.
spatial_and_multichannel.decorrelated_reverb_upmix	Channel geometry/order and timing metadata are correct.	Spatial collapse, combing, or localization bias from misalignment.	Avoid blind spatial processing when channel order/calibration is unknown.
spatial_and_multichannel.microphone_array_calibration_tones	Channel geometry/order and timing metadata are correct.	Spatial collapse, combing, or localization bias from misalignment.	Avoid blind spatial processing when channel order/calibration is unknown.
spatial_and_multichannel.multichannel_noise_psd_tracking	Channel geometry/order and timing metadata are correct.	Spatial collapse, combing, or localization bias from misalignment.	Avoid blind spatial processing when channel order/calibration is unknown.
spatial_and_multichannel.multichannel_wiener_postfilter	Channel geometry/order and timing metadata are correct.	Spatial collapse, combing, or localization bias from misalignment.	Avoid blind spatial processing when channel order/calibration is unknown.
spatial_and_multichannel.phase_aligned_mid_side_field_rotation	Channel geometry/order and timing metadata are correct.	Spatial collapse, combing, or localization bias from misalignment.	Avoid blind spatial processing when channel order/calibration is unknown.
spatial_and_multichannel.phase_consistent_multichannel_denoise	Channel geometry/order and timing metadata are correct. Noise model should be representative of observed noise floor.	Spatial collapse, combing, or localization bias from misalignment. Mismatched noise estimate leaves residue or damages detail.	Avoid blind spatial processing when channel order/calibration is unknown. Avoid static settings on rapidly varying nonstationary noise.
spatial_and_multichannel.pvx_directional_spectral_warp	Channel geometry/order and timing metadata are correct.	Spatial collapse, combing, or localization bias from misalignment.	Avoid blind spatial processing when channel order/calibration is unknown.
spatial_and_multichannel.pvx_interaural_coherence_shaping	Channel geometry/order and timing metadata are correct.	Spatial collapse, combing, or localization bias from misalignment.	Avoid blind spatial processing when channel order/calibration is unknown.
spatial_and_multichannel.pvx_interchannel_phase_locking	Channel geometry/order and timing metadata are correct.	Spatial collapse, combing, or localization bias from misalignment.	Avoid blind spatial processing when channel order/calibration is unknown.
spatial_and_multichannel.pvx_multichannel_time_alignment	Channel geometry/order and timing metadata are correct.	Spatial collapse, combing, or localization bias from misalignment.	Avoid blind spatial processing when channel order/calibration is unknown.
spatial_and_multichannel.pvx_spatial_freeze_and_trajectory	Channel geometry/order and timing metadata are correct.	Spatial collapse, combing, or localization bias from misalignment.	Avoid blind spatial processing when channel order/calibration is unknown.
spatial_and_multichannel.pvx_spatial_transient_preservation	Channel geometry/order and timing metadata are correct.	Spatial collapse, combing, or localization bias from misalignment.	Avoid blind spatial processing when channel order/calibration is unknown.
spatial_and_multichannel.rotating_speaker_doppler_field	Channel geometry/order and timing metadata are correct.	Spatial collapse, combing, or localization bias from misalignment.	Avoid blind spatial processing when channel order/calibration is unknown.
spatial_and_multichannel.spatial_freeze_resynthesis	Channel geometry/order and timing metadata are correct.	Spatial collapse, combing, or localization bias from misalignment.	Avoid blind spatial processing when channel order/calibration is unknown.
spatial_and_multichannel.spectral_spatial_granulator	Channel geometry/order and timing metadata are correct.	Spatial collapse, combing, or localization bias from misalignment.	Avoid blind spatial processing when channel order/calibration is unknown.
spatial_and_multichannel.stereo_width_frequency_dependent_control	Channel geometry/order and timing metadata are correct.	Spatial collapse, combing, or localization bias from misalignment.	Avoid blind spatial processing when channel order/calibration is unknown.
spatial_and_multichannel.stochastic_spatial_diffusion_cloud	Channel geometry/order and timing metadata are correct.	Spatial collapse, combing, or localization bias from misalignment.	Avoid blind spatial processing when channel order/calibration is unknown.
spatial_and_multichannel.transaural_crosstalk_cancellation	Channel geometry/order and timing metadata are correct.	Spatial collapse, combing, or localization bias from misalignment.	Avoid blind spatial processing when channel order/calibration is unknown.
spatial_and_multichannel.vbap_adaptive_panning	Channel geometry/order and timing metadata are correct.	Spatial collapse, combing, or localization bias from misalignment.	Avoid blind spatial processing when channel order/calibration is unknown.

spectral_time_frequency_transforms

Algorithm ID	Assumptions	Failure Modes	When Not To Use
spectral_time_frequency_transforms.chirplet_transform_analysis	Transform parameterization matches target time-frequency structure.	Incorrect parameterization can smear events or over-fragment spectra.	Avoid default settings for highly nonstationary signals without tuning.
spectral_time_frequency_transforms.constant_q_transform_cqt_processing	Transform parameterization matches target time-frequency structure.	Incorrect parameterization can smear events or over-fragment spectra.	Avoid default settings for highly nonstationary signals without tuning.
spectral_time_frequency_transforms.multi_window_stft_fusion	Transform parameterization matches target time-frequency structure.	Incorrect parameterization can smear events or over-fragment spectra.	Avoid default settings for highly nonstationary signals without tuning.
spectral_time_frequency_transforms.nsgt_based_processing	Transform parameterization matches target time-frequency structure.	Incorrect parameterization can smear events or over-fragment spectra.	Avoid default settings for highly nonstationary signals without tuning.
spectral_time_frequency_transforms.reassigned_spectrogram_methods	Transform parameterization matches target time-frequency structure.	Incorrect parameterization can smear events or over-fragment spectra.	Avoid default settings for highly nonstationary signals without tuning.
spectral_time_frequency_transforms.synchrosqueezed_stft	Transform parameterization matches target time-frequency structure.	Incorrect parameterization can smear events or over-fragment spectra.	Avoid default settings for highly nonstationary signals without tuning.
spectral_time_frequency_transforms.variable_q_transform_vqt	Transform parameterization matches target time-frequency structure.	Incorrect parameterization can smear events or over-fragment spectra.	Avoid default settings for highly nonstationary signals without tuning.
spectral_time_frequency_transforms.wavelet_packet_processing	Transform parameterization matches target time-frequency structure.	Incorrect parameterization can smear events or over-fragment spectra.	Avoid default settings for highly nonstationary signals without tuning.

time_scale_and_pitch_core

Algorithm ID	Assumptions	Failure Modes	When Not To Use
time_scale_and_pitch_core.beat_synchronous_time_warping	Frames are locally quasi-stationary and harmonic evolution is reasonably smooth.	High-ratio stretch can introduce phasiness and blurred transients.	Avoid for extreme percussive-only material when attack realism is critical.
time_scale_and_pitch_core.harmonic_percussive_split_tsm	Frames are locally quasi-stationary and harmonic evolution is reasonably smooth.	High-ratio stretch can introduce phasiness and blurred transients.	Avoid for extreme percussive-only material when attack realism is critical.
time_scale_and_pitch_core.lp_psola	Frames are locally quasi-stationary and harmonic evolution is reasonably smooth.	High-ratio stretch can introduce phasiness and blurred transients.	Avoid for extreme percussive-only material when attack realism is critical.
time_scale_and_pitch_core.multi_resolution_phase_vocoder	Frames are locally quasi-stationary and harmonic evolution is reasonably smooth. Phase continuity assumptions hold best for moderate stretch ratios.	High-ratio stretch can introduce phasiness and blurred transients. Extreme settings increase phasiness/transient blur risk.	Avoid for extreme percussive-only material when attack realism is critical. Avoid very large stretch+pitch shifts without transient controls.
time_scale_and_pitch_core.nonlinear_time_maps	Frames are locally quasi-stationary and harmonic evolution is reasonably smooth.	High-ratio stretch can introduce phasiness and blurred transients.	Avoid for extreme percussive-only material when attack realism is critical.
time_scale_and_pitch_core.td_psola	Frames are locally quasi-stationary and harmonic evolution is reasonably smooth.	High-ratio stretch can introduce phasiness and blurred transients.	Avoid for extreme percussive-only material when attack realism is critical.
time_scale_and_pitch_core.wsola_waveform_similarity_overlap_add	Frames are locally quasi-stationary and harmonic evolution is reasonably smooth.	High-ratio stretch can introduce phasiness and blurred transients.	Avoid for extreme percussive-only material when attack realism is critical.

Attribution

See ATTRIBUTION.md.