The SSML 1.0 spec shows various options to support the voice tag:
<voice gender="female">Female voice default</voice>
<voice gender="female" variant="2">Female voice number 2</voice>
<voice name="Mike" xml:lang="en-GB">Mike voice</voice>
SSML 1.1 added a <lang> tag with a single xml:lang attribute.
Bixby has opted for <lang> tag with a non-standard voice attribute.
Please fix to follow SSML with a preference toward the name attribute:
<voice name="Louise">Louise voice</voice>
<voice name="Chris"><lang xml:lang="en-GB">Chris voice</lang></voice>
This will help developers and designers who are already familiar with SSML to implement speech formatting in Bixby. It will also allow better support in open source frameworks such as Speech Markdown.
Please sign in to leave a comment.