The discussion regarding "MT" vs "Mt" merits some explanation.
The
Big TTS Test List initially was derived from the abbreviation lists of several official sources. Most of them used only capital letters for the abbreviations. They were imported as is into the first worksheet.
Therefore the initial round of testing was NOT necessarily case aware. Jencard and I may have submitted our tests in upper, lower or mixed case.
Subsequent testing using specific capitalization has the abbreviation in
BOLD letters.
This is documented somewhere in the bottom notes in the list.
The sand box areas for those tests were deleted by others a long time ago, so no further record exists on how they were tested.
Since the list was already so large, and since we had no progress for years with Waze understanding how it worked, testing was never resumed let alone started over from scratch. Further complications due to the dual file issues also thwarted further testing.
To be fair to Waze, they originally implied that they did not have a clear understanding of how it all worked either. We were told then that they would have to submit new rules to Nuance, which I do not think is the case any longer.
Please note that the Nuance TTS demo sites are NOT a good place to test anything for Waze. From what I can determine, that test site uses only the base data for the Samantha and Tom voices for the "American English". We believe that Waze actually uses the Nuance enhanced "Mobile" version built onto those base voice systems. Changes we submit are then exceptions introduced by Waze as an additional layer into this process.
Nuance Vocalizer 5 Interactive Demo (Includes Samatha and Tom)Nuance Interactive Loquendo TTS Demo (Samantha voice is not available).
