Drama Studio: local realistic text-to-speech synthesis engine for Apple Silicon (github.com)

Fish Audio has launched Drama Studio, a local text-to-speech synthesis engine designed specifically for Apple Silicon Macs. This native macOS application leverages the Drama 3 model, built on a 35 billion parameter mixture-of-experts architecture, utilizing custom Metal compute kernels for efficient execution. The entire voice processing is handled locally, with real-time audio generation through Core ML, eliminating the need for heavier frameworks like PyTorch during runtime. Users will need at least 49 GB of free disk space and 64 GB of unified memory for optimal performance, as the application is optimized for M3 Max chips and may encounter issues on lower-spec machines. The release is significant for the AI/ML community as it offers a suite of features tailored for spoken dialogue application, including a screenplay mode for generating multiple character voices, support for various script formats, and built-in word recognition checks to ensure accuracy in voice output. Voice cloning capabilities allow users to create custom voices from reference clips, fostering creativity and personalization in content generation. With Drama Studio's emphasis on local processing, users can maintain privacy and control over their generated speech, setting a new standard for accessibility and performance in text-to-speech technologies.