In this video, I demonstrate how to drive MetaHuman lip sync using local Text-to-Speech (TTS) in Unreal Engine 5.6, completely offline and fully supported in packaged projects.
This setup works for MetaHumans created with the new MetaHuman Creator system introduced in Unreal Engine 5.6 (as well as any custom MetaHuman characters), using three plugins working together: Runtime MetaHuman Lip Sync – for generating facial animation from audio Runtime Text To Speech – for generating speech from text using local voices Runtime Audio Importer – for processing audio at runtime
The local TTS system supports 40+ languages, 900+ voices, and 160+ voice styles, all without requiring an internet connection.
Hi, this tutorial is basically suitable for UE versions 5.0 and later (including the latest 5.6). For engine versions earlier than 5.6, the process of creating MetaHuman characters is different, but the lip sync + text-to-speech integration is the same for all engine versions from 5.0 to 5.6.
MetaHuman SDK is another plugin which requires a subscription to an external server and internet connection, and this plugin is quite outdated and may not even work nowadays (their website is not working, their last update was in October 2024, and the latest UE version supported is 5.4). This plugin is not used in the tutorial intentionally, but instead, it uses a newer and more reliable plugin called Runtime MetaHuman Lip Sync (and Runtime Text To Speech + Runtime Audio Importer) to perform the lip sync fully locally and natively on your device, which doesn’t require an internet connection or subscriptions, since everything works locally.
Runtime MetaHuman Lip Sync – for generating facial animation from audio