Voxenes-2026: Bilingual Benchmark for Audio Deepfake Detection

A bilingual benchmark dataset for evaluating audio deepfake detectors against text-to-speech and voice conversion systems. The dataset's author, organization, and specific scale are not provided in the metadata. It was sourced from the Kaggle platform, but the last update date is unknown.

Use Cases

• Benchmarking audio deepfake detection models based on the described TTS and VC audio samples.
• Training classifiers to distinguish between real and synthetic speech based on the bilingual audio content.
• Evaluating the robustness of speech synthesis systems against detection algorithms based on the benchmark's structure.

Strengths

• Dataset is designed as a benchmark for a specific, high-relevance task in audio security.
• Includes audio samples from both text-to-speech and voice conversion systems, covering two common deepfake methods.
• Benchmark is bilingual, which may allow for cross-linguistic evaluation of detection models.

Limitations

• Row count and total audio duration are unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Last update date is unknown; freshness unverified.

Provenance

Source

Kaggle