English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Constructing an embedding space for musical instrument sounds that canmeaningfully represent new and unseen instruments is important for downstreammusic generation tasks such as multi-instrument synthesis and timbre transfer.The framework of Automatic Speaker Verification (ASV) provides us witharchitectures and evaluation methodologies for verifying the identities ofunseen speakers, and these can be repurposed for the task of learning andevaluating a musical instrument sound embedding space that can support unseeninstruments. Borrowing from state-of-the-art ASV techniques, we construct amusical instrument recognition model that uses a SincNet front-end, a ResNetarchitecture, and an angular softmax objective function. Experiments on theNSynth and RWC datasets show our model's effectiveness in terms of equal errorrate (EER) for unseen instruments, and ablation studies show the importance ofdata augmentation and the angular softmax objective. Experiments also show thebenefit of using a CQT-based filterbank for initializing SincNet over a Melfilterbank initialization. Further complementary analysis of the learnedembedding space is conducted with t-SNE visualizations and probingclassification tasks, which show that including instrument family labels as amulti-task learning target can help to regularize the embedding space andincorporate useful structure, and that meaningful information such as playingstyle, which was not included during training, is contained in the embeddingsof unseen instruments.

Use of speaker recognition approaches for learning and evaluating embedding representations of musical instrument sounds