The problem in question seems to have been going for a while.

One way to try and go around this is to use hard-coded voice objects as suggested here. However, it is unclear to me how to use an object like: { name: "Samantha (Enhanced)", voiceURI: "com.apple.ttsbundle.Samantha-premium", lang: "en-US", localService: true, "default": true }

to instantiate a SpeechSynthesisVoice object.

Has anyone come across a way to instantiate a SpeechSynthesisVoice object as a workaround for an empty array of available voices? Thanks!