OpenAI Startup Fund’s Portfolio Company Improves RVQGAN: 90x Compression of 44.1 KHz Audio at 8kbps Bandwidth

In a new paper High-Fidelity Audio Compression with Improved RVQGAN, a Descript research team presents Improved RVQGAN, a high fidelity universal audio compression model that combines advances in high-fidelity audio generation and improved adversarial and reconstruction losses to achieve 90x compression of 44.1 KHz audio at only 8kbps bandwidth.

ai artificial intelligence audio compression deep-neural-networks descript fidelity generative adversarial network losses machine learning machine learning & data science ml neural audio compression openai openai startup fund paper portfolio research research team startup team technology