Fusion of Modulation Spectrogram and SSL with Multi-head Attention for Fake Speech Detection

Rishith Sadashiv T N 1 Abhishek Bedge 1 Saisha Suresh Bore 2 Jagabandhu Mishra 3 Mrinmoy Bhattacharjee 4 S R Mahadeva Prasanna 1,2 1 IIT Dharwad 2 IIIT Dharwad 3 University of Eastern Finland 4 IIT Jammu Email: [email protected] · 2025

Read Fusion of Modulation Spectrogram and SSL with Multi-head Attention for Fake Speech Detection by Rishith Sadashiv T N 1 Abhishek Bedge 1 Saisha Suresh Bore 2 Jagabandhu Mishra 3 Mrinmoy Bhattacharjee 4 S R Mahadeva Prasanna 1,2 1 IIT Dharwad 2 IIIT Dharwad 3 University of Eastern Finland 4 IIT Jammu Email: [email protected] (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-01034.txt.

Description

Rishith Sadashiv and co-authors fuse two views of speech in a 2025 paper to catch fake voices. Fake speech tools ace their home sets and fail on new ones, since training speech is too narrow. A richer front view of the voice may travel better.

Their view joins deep self-taught voice codes with a chart of how voice tones wobble over time. A focus head blends the two before a known back-end judge scores the clip. Tests span one-tongue and many-tongue fake sets. Guard teams gain a front end built to hold up when the fake source shifts.

Catalog record

CREATORRishith Sadashiv T N 1 Abhishek Bedge 1 Saisha Suresh Bore 2 Jagabandhu Mishra 3 Mrinmoy Bhattacharjee 4 S R Mahadeva Prasanna 1,2 1 IIT Dharwad 2 IIIT Dharwad 3 University of Eastern Finland 4 IIT Jammu Email: [email protected]
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE25 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2508-01034

Browse the full library · About the Tower of Babel