High Resolution Guitar Transcription via Domain Adaptation
Automatic music transcription (AMT) has achieved high accuracy for piano due to the availability of large, high-quality datasets such as MAESTRO and MAPS, but comparable datasets are not yet available for other instruments. In recent work, however, it has been demonstrated that aligning scores to transcription model activations can produce high quality AMT training data for instruments other than piano. Focusing on the guitar, we refine this approach to training on score data using a dataset of commercially available score-audio pairs. We propose the use of a high-resolution piano transcription model to train a new guitar transcription model. The resulting model obtains state-of-the-art transcription results on GuitarSet in a zero-shot context, improving on previously published methods.
Code (0)
등록된 구현이 없습니다.
Tasks
Domain AdaptationMusic TranscriptionSimilar Papers 제목 키워드 기반
Joint Transcription of Acoustic Guitar Strumming Directions and Chords
Automatic transcription of guitar strumming is an underrepresented and challenging task in Music Information Retrieval (MIR), particularly for extracting both strumming directions and chord progressions from audio signal…
Information RetrievalAction DetectionSynthTab: Leveraging Synthesized Data for Guitar Tablature Transcription
Guitar tablature is a form of music notation widely used among guitarists. It captures not only the musical content of a piece, but also its implementation and ornamentation on the instrument. Guitar Tablature Transcript…
SpecificityGOAT: A Large Dataset of Paired Guitar Audio Recordings and Tablatures
In recent years, the guitar has received increased attention from the music information retrieval (MIR) community driven by the challenges posed by its diverse playing techniques and sonic characteristics. Mainly fueled …
Information RetrievalData AugmentationA Data-Driven Methodology for Considering Feasibility and Pairwise Likelihood in Deep Learning Based Guitar Tablature Transcription Systems
Guitar tablature transcription is an important but understudied problem within the field of music information retrieval. Traditional signal processing approaches offer only limited performance on the task, and there is l…
Information RetrievalMusic Information RetrievalRetrievalMulti-task learning-based temporal pattern matching network for guitar tablature transcription
Guitar tablature transcription poses unique challenges in automatic music transcription, as it requires capturing both pitch and string usage on a multi-string instrument with various expressive techniques. While guitar …
Multi-Task LearningMusic Transcription