Nvidia and NDI Team Up to Cut Multilingual Video Bandwidth Eightfold with AI
NDI and Nvidia have collaborated to develop AI-powered broadcast workflows that enable the creation of multiple language versions from a single media stream, resulting in significantly reduced video bandwidth. In internal tests, the companies achieved remarkable results by cutting the volume of video carried to around one-eighth for a 10-language workflow.
The new method supports real-time translation, lip-synced dubbing, and regional language adaptation, allowing organisations to adapt their content for different audiences without replicating full production operations. This innovative solution combines NDI-native media connectivity with Nvidia AI technologies, including NIM microservices.
NDI head Miguel Coutinho stated that this collaboration has enabled a 'fundamentally different model' where a single media workflow can support multiple language experiences. Lewis Smithingham, Nvidia's principal product manager for media and entertainment, added that AI is transforming the content creation process by providing organisations with the ability to increase viewer engagement through multilingual content delivery.
The solution is currently being tested with partners in live production and enterprise communication, holding promise for significantly reducing bandwidth and distribution costs while expanding access to multilingual content for larger audiences.