This issue occurs in NVDA 2019.3.1 and above.
During the key navigation in Windows directories three separate speak calls are generated from the sapi5.py synthDriver. For example, the calls may contain following texts:
1. <pitch absmiddle="0"><spell>s<Bookmark Mark="166" /></spell></pitch>
2. <Bookmark Mark="167" />
3. <pitch absmiddle="0">Spotify 4 z 16 <Bookmark Mark="168" /></pitch>
Unfortunately, IVONA SAPI 5 voices don't generate the bookmark callback for a speak command that contains text with only bookmark tag. In effect the speak command with the last text is not called.
IVONA voice authors reported they can't fix that on their side.
This issue occurs in NVDA 2019.3.1 and above.
During the key navigation in Windows directories three separate speak calls are generated from the sapi5.py synthDriver. For example, the calls may contain following texts:
Unfortunately, IVONA SAPI 5 voices don't generate the bookmark callback for a speak command that contains text with only bookmark tag. In effect the speak command with the last text is not called.
IVONA voice authors reported they can't fix that on their side.