The former method lost any XML markup (such as say-as tags, etc) embedded within the voice tag. In the future if support for setting voice speed is added, it can be inserted here easily as well.
* WIP: initial commit for observatory herald
* Plugin error handling refactor
* make error window non-modal
* tidy up plugin error handling
* first pass for basic herald functionality
* corrections for linux env
* Use FNV hash directly instead of managing through dictionary/index file
* resolve audio queuing issue, switch to personal NetCoreAudio fork
* merge cleanup
* add enable setting, populate defaults
* framework xml doc update
* Adjust settings, add style selection, replace locale with demonym in dropdown list.
* Test is position is on screen before saving/loading.
* use a default that's actually in the list