I’ve noticed a consistent issue with voice selection when using project-based chats.
Issue Summary:
I have selected a specific voice in my settings (Cove, a male voice).
In normal (non-project) conversations, my chosen voice works correctly in both active conversation and read-aloud mode.
In project-based conversations, the UI still shows my chosen voice, but the actual voice speaking in Active Voice is a default female voice instead.
However, the read-aloud function in projects still uses the correct selected voice (Cove).
This means that my chosen voice is being recognized in some functions (read-aloud) but not in active conversations inside projects.
What I’ve Tried:
I confirmed this only happens inside project-based chats.
I checked voice settings, but the issue persists.
Other users have reported similar issues:
One thread (mislabeled under “Advanced Voice Mode”) describes a similar issue but isn’t categorized correctly.
Another thread about Tasks reports the same problem, but in that feature instead of Projects.
Request for Fix or Clarification:
Since the system incorrectly displays the selected voice in settings but doesn’t apply it to active project conversations, instead defaulting to a female voice, this is either:
-
A bug that needs fixing.
-
An undocumented limitation in projects, tasks, and potentially others that should be clarified.
Given that this issue has now been observed in both Projects and Tasks, it may be a broader system issue affecting voice selection in multiple features.
Can anyone from @OpenAI confirm if this is an intentional system limitation or an oversight? If it’s a bug, is there a fix in progress?
Thanks for looking into this!