Wiki Change Log¶

[2026-05-19] ingest | Batch 1 — 10 initial papers¶

New: 10 source summaries, 11 entity pages
Updated: 9 concept pages with source citations, index.md, log.md, home.md
Sources ingested: Demucs (2019), Onsets and Frames (2017), CREPE (2018), GuitarSet (2018), TabCNN (2019), Basic Pitch (2022), HTDemucs (2022), AudioSep (2023), BS-RoFormer (2023), SynthTab (2024)
Entities created: demucs, basic-pitch, tabcnn, audiosef, bs-roformer, guitarset, crepe, spleeter, open-unmix, mt3, onsets-and-frames
Summary: First research pass complete. Source separation landscape: Demucs→HTDemucs→BS-RoFormer progression, with AudioSep offering novel text-conditional approach. Transcription landscape: Basic Pitch best starting point, TabCNN most relevant architecture, GuitarSet the dataset template. Confirmed no existing work on bluegrass/banjo separation or transcription.

New: All topic hubs (6), concept stubs (15), index.md, log.md, home.md
Topics created: transcription-and-pitch-estimation, source-separation, audio-classification-and-representation, classical-techniques, datasets, tools-and-libraries
Concepts created: query-based-source-separation, self-supervised-audio-representation, permutation-invariant-training, spectrogram-unets, musicxml-tab-notation, synthetic-mixing-pipelines, deep-clustering-separation, mixture-invariant-training, auditory-scene-analysis, computational-auditory-scene-analysis, common-fate-principle, sinusoidal-modeling, non-negative-matrix-factorization-audio, shift-invariant-nmf, informed-model-based-separation
Summary: Wiki seeded from Nat's comprehensive audio separation and transcription research taxonomy.