BanglaFake: Constructing and Evaluating a Specialized Bengali Deepfake Audio Dataset

TL;DR

This paper introduces BanglaFake, a large Bengali deepfake audio dataset with high-quality synthetic speech, aiming to facilitate research in deepfake detection for low-resource languages.

Contribution

The creation and evaluation of BanglaFake, a novel Bengali deepfake audio dataset with over 25,000 utterances generated using state-of-the-art TTS models.

Findings

High naturalness and intelligibility scores for synthetic speech

t-SNE visualization reveals challenges in differentiating real and fake audio

Dataset addresses resource limitations in Bengali deepfake detection

Abstract

Deepfake audio detection is challenging for low-resource languages like Bengali due to limited datasets and subtle acoustic features. To address this, we introduce BangalFake, a Bengali Deepfake Audio Dataset with 12,260 real and 13,260 deepfake utterances. Synthetic speech is generated using SOTA Text-to-Speech (TTS) models, ensuring high naturalness and quality. We evaluate the dataset through both qualitative and quantitative analyses. Mean Opinion Score (MOS) from 30 native speakers shows Robust-MOS of 3.40 (naturalness) and 4.01 (intelligibility). t-SNE visualization of MFCCs highlights real vs. fake differentiation challenges. This dataset serves as a crucial resource for advancing deepfake detection in Bengali, addressing the limitations of low-resource language research.