Feature: Integrate ONNX Runtime for Existing AudioProcessor

[omeryusufyagci]

We need to integrate the existing machine learning model into the project using ONNX Runtime, ensuring the model functions correctly when processing small audio chunks. The goal is to achieve real-time audio processing while maintaining the internal state across potentially different threads.

An initial implementation of the respective C API has been made, which should provide a starting point.

Please see the header for details.