Explore and analyze audio data with AudioBench Leaderboard
Generate realistic voices from text
Generate speech from text with custom voice
Generate speech from text
ExpressivText-to-Speech
Generate audio from text
Kokoro is an open-weight TTS model with 82 million parameters.
Transcribe spoken Russian into text
Convert text into speech in Japanese
Generate speech from text with adjustable rate and pitch
Generate speech from text with customizable voices
Generate audio from text or modify voice pitch
Sound effect from description
Leaderboard / AudioBench is a platform designed for benchmarking, analyzing, and comparing audio data in the context of speech synthesis. It serves as a tool to evaluate and explore the performance of different models and systems, providing insights into their capabilities and limitations.
What is the purpose of Leaderboard / AudioBench?
The purpose of Leaderboard / AudioBench is to provide a comprehensive platform for evaluating and comparing audio data, particularly in speech synthesis, to help users identify top-performing models and systems.
Can I use Leaderboard / AudioBench for custom datasets?
Yes, Leaderboard / AudioBench supports custom datasets, allowing users to upload their own audio data for benchmarking and analysis.
What makes Leaderboard / AudioBench unique?
Its ability to provide detailed, customizable analysis and comparison of speech synthesis models, along with advanced visualization tools, makes it a standout platform in the field of audio data evaluation.