ClonEval: An Open Voice Cloning Benchmark
By: Iwona Christop, Tomasz Kuczyński, Marek Kubis
Potential Business Impact:
Tests how well computers copy voices.
We present a novel benchmark for voice cloning text-to-speech models. The benchmark consists of an evaluation protocol, an open-source library for assessing the performance of voice cloning models, and an accompanying leaderboard. The paper discusses design considerations and presents a detailed description of the evaluation procedure. The usage of the software library is explained, along with the organization of results on the leaderboard.
Similar Papers
Voice Cloning: Comprehensive Survey
Sound
Makes computers sound like anyone with a few words.
VoiceAssistant-Eval: Benchmarking AI Assistants across Listening, Speaking, and Viewing
Computation and Language
Tests AI assistants on hearing, talking, and seeing.
NorEval: A Norwegian Language Understanding and Generation Evaluation Benchmark
Computation and Language
Tests how well computers understand Norwegian.