Same here and that voice really was the only good one. I don't know why they don't bring the voices from their API over, which are all much better, like Nova or Shimmer (https://platform.openai.com/docs/guides/text-to-speech)