ESP/Speaker: Remove std::string

`std::string` uses dynamic allocation under the hood, and should be avoided to try and avoid long term fragmentation.

consider using `std::string_view` where applicable.

Start by locating the ESP/Speaker code and every use of std::string, then determine which uses can safely be replaced with std::string_view. Done means applicable dynamic string usage is removed without changing the speaker behavior, but the issue does not name files or tests to guide verification.

Written by the indexing model from the issue text.