@inproceedings{chetia-phukan-etal-2025-investigating,
title = "Investigating Prosodic Signatures via Speech Pre-Trained Models for Audio Deepfake Source Attribution",
author = "Chetia Phukan, Orchid and
Singh, Drishti and
Behera, Swarup Ranjan and
Buduru, Arun Balaji and
Sharma, Rajesh",
editor = "Che, Wanxiang and
Nabende, Joyce and
Shutova, Ekaterina and
Pilehvar, Mohammad Taher",
booktitle = "Findings of the Association for Computational Linguistics: ACL 2025",
month = jul,
year = "2025",
address = "Vienna, Austria",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.findings-acl.218/",
doi = "10.18653/v1/2025.findings-acl.218",
pages = "4206--4214",
ISBN = "979-8-89176-256-5",
abstract = "In this work, we investigate various state-of-the-art (SOTA) speech pre-trained models (PTMs) for their capability to capture prosodic sig-natures of the generative sources for audio deepfake source attribution (ADSD). These prosodic characteristics can be considered oneof major signatures for ADSD, which is unique to each source. So better is the PTM at capturing prosodic signs better the ADSD per-formance. We consider various SOTA PTMs that have shown top performance in different prosodic tasks for our experiments on benchmark datasets, ASVSpoof 2019 and CFAD. x-vector (speaker recognition PTM) attains the highest performance in comparison to allthe PTMs considered despite consisting lowest model parameters. This higher performance can be due to its speaker recognition pre-training that enables it for capturing unique prosodic characteristics of the sources in a better way. Further, motivated from tasks suchas audio deepfake detection and speech recognition, where fusion of PTMs representations lead to improved performance, we explorethe same and propose FINDER for effective fusion of such representations. With fusion of Whisper and x-vector representations through FINDER, we achieved the topmost performance in comparison to all the individual PTMs as well as baseline fusion techniques and attaining SOTA performance."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="chetia-phukan-etal-2025-investigating">
<titleInfo>
<title>Investigating Prosodic Signatures via Speech Pre-Trained Models for Audio Deepfake Source Attribution</title>
</titleInfo>
<name type="personal">
<namePart type="given">Orchid</namePart>
<namePart type="family">Chetia Phukan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Drishti</namePart>
<namePart type="family">Singh</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Swarup</namePart>
<namePart type="given">Ranjan</namePart>
<namePart type="family">Behera</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Arun</namePart>
<namePart type="given">Balaji</namePart>
<namePart type="family">Buduru</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Rajesh</namePart>
<namePart type="family">Sharma</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: ACL 2025</title>
</titleInfo>
<name type="personal">
<namePart type="given">Wanxiang</namePart>
<namePart type="family">Che</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Joyce</namePart>
<namePart type="family">Nabende</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ekaterina</namePart>
<namePart type="family">Shutova</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mohammad</namePart>
<namePart type="given">Taher</namePart>
<namePart type="family">Pilehvar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Vienna, Austria</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-256-5</identifier>
</relatedItem>
<abstract>In this work, we investigate various state-of-the-art (SOTA) speech pre-trained models (PTMs) for their capability to capture prosodic sig-natures of the generative sources for audio deepfake source attribution (ADSD). These prosodic characteristics can be considered oneof major signatures for ADSD, which is unique to each source. So better is the PTM at capturing prosodic signs better the ADSD per-formance. We consider various SOTA PTMs that have shown top performance in different prosodic tasks for our experiments on benchmark datasets, ASVSpoof 2019 and CFAD. x-vector (speaker recognition PTM) attains the highest performance in comparison to allthe PTMs considered despite consisting lowest model parameters. This higher performance can be due to its speaker recognition pre-training that enables it for capturing unique prosodic characteristics of the sources in a better way. Further, motivated from tasks suchas audio deepfake detection and speech recognition, where fusion of PTMs representations lead to improved performance, we explorethe same and propose FINDER for effective fusion of such representations. With fusion of Whisper and x-vector representations through FINDER, we achieved the topmost performance in comparison to all the individual PTMs as well as baseline fusion techniques and attaining SOTA performance.</abstract>
<identifier type="citekey">chetia-phukan-etal-2025-investigating</identifier>
<identifier type="doi">10.18653/v1/2025.findings-acl.218</identifier>
<location>
<url>https://aclanthology.org/2025.findings-acl.218/</url>
</location>
<part>
<date>2025-07</date>
<extent unit="page">
<start>4206</start>
<end>4214</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Investigating Prosodic Signatures via Speech Pre-Trained Models for Audio Deepfake Source Attribution
%A Chetia Phukan, Orchid
%A Singh, Drishti
%A Behera, Swarup Ranjan
%A Buduru, Arun Balaji
%A Sharma, Rajesh
%Y Che, Wanxiang
%Y Nabende, Joyce
%Y Shutova, Ekaterina
%Y Pilehvar, Mohammad Taher
%S Findings of the Association for Computational Linguistics: ACL 2025
%D 2025
%8 July
%I Association for Computational Linguistics
%C Vienna, Austria
%@ 979-8-89176-256-5
%F chetia-phukan-etal-2025-investigating
%X In this work, we investigate various state-of-the-art (SOTA) speech pre-trained models (PTMs) for their capability to capture prosodic sig-natures of the generative sources for audio deepfake source attribution (ADSD). These prosodic characteristics can be considered oneof major signatures for ADSD, which is unique to each source. So better is the PTM at capturing prosodic signs better the ADSD per-formance. We consider various SOTA PTMs that have shown top performance in different prosodic tasks for our experiments on benchmark datasets, ASVSpoof 2019 and CFAD. x-vector (speaker recognition PTM) attains the highest performance in comparison to allthe PTMs considered despite consisting lowest model parameters. This higher performance can be due to its speaker recognition pre-training that enables it for capturing unique prosodic characteristics of the sources in a better way. Further, motivated from tasks suchas audio deepfake detection and speech recognition, where fusion of PTMs representations lead to improved performance, we explorethe same and propose FINDER for effective fusion of such representations. With fusion of Whisper and x-vector representations through FINDER, we achieved the topmost performance in comparison to all the individual PTMs as well as baseline fusion techniques and attaining SOTA performance.
%R 10.18653/v1/2025.findings-acl.218
%U https://aclanthology.org/2025.findings-acl.218/
%U https://doi.org/10.18653/v1/2025.findings-acl.218
%P 4206-4214
Markdown (Informal)
[Investigating Prosodic Signatures via Speech Pre-Trained Models for Audio Deepfake Source Attribution](https://aclanthology.org/2025.findings-acl.218/) (Chetia Phukan et al., Findings 2025)
ACL