@Gabriele Serra I think there is a bit of confusion with the documentation for enable audio logging. I have tried this scenario and the logs are enable on the service. This is detailed in the speech to text FAQ document here.
The logs can be retrieved using the Get Base Model Endpoint API
For a configuration that I used with the python SDK.
speech_config.enable_audio_logging()
The logs location is available in the response of this API.
The format of the file should contain all the details of the requests you pass to the speech API.