PulseExploreJournal ClubResearchersJournals
Instagram
HomeJournal ClubExplore
Synapse
⌘+K
Synapse
March 3, 2026Open Access

Audio Foundation-Modeller för Generering av Ljudeffekter : Utvärdering, Kontrollerbarhet och Redigerbarhet av de Senaste Modellerna

View Full Paper
Ask AI
Bookmark
Share

Authors

THTheresa Anna Hösl

Discussion

Loading...

Member takes

Overview

Observational analysis evaluates sound effects generated by textto-audio models, indicating editing potential exists.

Key Points

  • Sound effect generation often lacks accuracy under specific prompts, limiting use cases.
  • The model Stable Audio Open was chosen for further investigation due to its editing capabilities.
  • Evaluation included user studies with methods like semantic guidance and cross-attention map blending.
  • The outcomes suggest a need for improved evaluation metrics to align subjective ratings with objective assessments.

Cite This Study

Theresa Anna Hösl (2025) studied this question.

synapsesocial.com/papers/69a75d5bc6e9836116a274a9https://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-375841
View Full Paper
Ask AI
Bookmark
Share