Saved in:
Bibliographic Details
Main Authors: Shi, Haohan, Shi, Xiyu, Dogan, Safak, Huang, Tianjin, Zhang, Yunxiao
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2601.19573
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916000557957120
author Shi, Haohan
Shi, Xiyu
Dogan, Safak
Huang, Tianjin
Zhang, Yunxiao
author_facet Shi, Haohan
Shi, Xiyu
Dogan, Safak
Huang, Tianjin
Zhang, Yunxiao
contents This paper focuses on audio deepfake detection under real-world communication degradations, with an emphasis on ultra-short inputs (0.5-2.0s), targeting the capability to detect synthetic speech at a conversation opening, e.g., when a scammer says "Hi." We propose Short-MGAA (S-MGAA), a novel lightweight extension of Multi-Granularity Adaptive Time-Frequency Attention, designed to enhance discriminative representation learning for short, degraded inputs subjected to communication processing and perturbations. The S-MGAA integrates two tailored modules: a Pixel-Channel Enhanced Module (PCEM) that amplifies fine-grained time-frequency saliency, and a Frequency Compensation Enhanced Module (FCEM) to supplement limited temporal evidence via multi-scale frequency modeling and adaptive frequency-temporal interaction. Extensive experiments demonstrate that S-MGAA consistently surpasses nine state-of-the-art baselines while achieving strong robustness to degradations and favorable efficiency-accuracy trade-offs, including low RTF, competitive GFLOPs, compact parameters, and reduced training cost, highlighting its strong potential for real-time deployment in communication systems and edge devices.
format Preprint
id arxiv_https___arxiv_org_abs_2601_19573
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Audio Deepfake Detection at the First Greeting: "Hi!"
Shi, Haohan
Shi, Xiyu
Dogan, Safak
Huang, Tianjin
Zhang, Yunxiao
Audio and Speech Processing
This paper focuses on audio deepfake detection under real-world communication degradations, with an emphasis on ultra-short inputs (0.5-2.0s), targeting the capability to detect synthetic speech at a conversation opening, e.g., when a scammer says "Hi." We propose Short-MGAA (S-MGAA), a novel lightweight extension of Multi-Granularity Adaptive Time-Frequency Attention, designed to enhance discriminative representation learning for short, degraded inputs subjected to communication processing and perturbations. The S-MGAA integrates two tailored modules: a Pixel-Channel Enhanced Module (PCEM) that amplifies fine-grained time-frequency saliency, and a Frequency Compensation Enhanced Module (FCEM) to supplement limited temporal evidence via multi-scale frequency modeling and adaptive frequency-temporal interaction. Extensive experiments demonstrate that S-MGAA consistently surpasses nine state-of-the-art baselines while achieving strong robustness to degradations and favorable efficiency-accuracy trade-offs, including low RTF, competitive GFLOPs, compact parameters, and reduced training cost, highlighting its strong potential for real-time deployment in communication systems and edge devices.
title Audio Deepfake Detection at the First Greeting: "Hi!"
topic Audio and Speech Processing
url https://arxiv.org/abs/2601.19573