Add support for playing audio generated at runtime

Currently, the plugin only supports playing audio data which is loaded as wav files. It would be favorable to load audio data at runtime in the form of raw PCM / Linear16 data, e.g. as TArray<uint8> or TArray<uint16>. This would be useful for example, when text-to-speech data is generated at runtime, because currently one would have to 1) convert the data to a complete wav file by creating the WAV header and appending the data 2) saving the file to disk 3) loading the file again 4) removing the temporary file after it was played, which results in unnecessary I/O operations.