Hi, my organization would like to use Azure Speech Service for a couple of use cases that involve client data. Here (https://azure.microsoft.com/en-us/services/cognitive-services/speech-to-text/#security) it says that "Your data remains yours. Your audio input and transcription data aren't logged during audio to text processing." but would you have more technical documentation on this? I.e. is audio input processed as speech stream and no audio nor text are stored outside of the buffer? Could somebody explain this in more technical terms, please?