ADIFF: Explaining audio difference using natural language
Year: 2025
Type: conference-paper
Abstract: ADIFF is an audio prefix tuning-based language model with a cross-projection module and undergoes a three-step training process. ADIFF takes two audios and text prompt as input and produces different tiers of difference explanations as output. This involves identifying and describing audio events, acoustic scenes, signal characteristics, and their emotional impact on listeners. The code repository is: soham97/ADIFF
Authors: Deshmukh, Soham, Han, Shuo, Singh, Rita, Raj, Bhiksha
Institutions: Microsoft (United States), Carnegie Mellon University
Language: English
Topic: Music and Audio Processing
Subfield: Signal Processing
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W6949384336
FWCI:
Cites:
Cited by:
Related to: