Update: devAIce® SDK 4.0 and Web API 4.8
With devAIce® SDK 4.0, Web API 4.8, and Microservice Support 1.0, audEERING is releasing an update that significantly reduces the workload for developers and companies. The new version improves the developer experience for anyone using advanced audio AI. It also simplifies key modules and opens up new ways to scale devAIce® in modern environments.
Modern developer APIs for C++ and Python
With the SDK version 4.0, audEERING is introducing a completely new C++ API. It is based on modern C++ standards and makes integration much easier. Developers work with clear structures, consistent naming conventions, and well-organized interfaces.
The Python API also benefits from this: The new Python wrapper is now based directly on the new C++ API. The result is a more intuitive, “pythonic” API that can be put into productive use more quickly. As a part of this realignment, the previous C API is officially deprecated. Android and iOS wrappers have been removed, while audEERING is working in parallel on improved mobile solutions for the future.
Unified Expression module and simple configuration
A key architectural innovation is the unification of the expression modules. Instead of separate variants, there is now a single expression module that can be flexibly configured. Users can choose between the large, small, and new tiny models depending on the application. The tiny model in particular saves considerable resources and is suitable for lightweight applications while maintaining high accuracy. The Speaker Attributes module is now configured via a single model flag, which simplifies the overall setup. Audio quality analysis is more precise with the new option to process only voice segments using VAD. The Multi-Modal Expression module has been discontinued.
Introduction of microservice support
With Microservice Support 1.0, audEERING is introducing a solution based on devAIce® SDK 4.0 that integrates seamlessly into cloud-native and containerized environments. Companies can flexibly scale their audio AI and integrate it into distributed systems. At the same time, data processing remains local and data protection is ensured.
Additional improvements
The documentation has been updated and expanded, and all C++ and Python sample code has been updated for the new APIs.
In Web API 4.8, the ASR module no longer supports the small model option due to high resource consumption.
The most important features at a glance
– Modernized C++ and Python APIs with clear documentation
– Unified Expression module with large, small and tiny models
– Microservice Support 1.0 for cloud-native and containerized environments
– Simplified configuration of Speaker Attributes and more precise analysis of the Audio Quality using VAD
Upgrade to devAIce® SDK 4.0, Web API 4.8, and Microservice Support 1.0 now and benefit from greater flexibility, improved scalability, and a significantly enhanced developer experience. If you have any questions or require assistance, please contact sales@audeering.com.
