MBTA switches to text-to-speech voice for all Green Line announcements

Within a few years, Frank Oglesby, Jr.'s voice will be fully phased out. The Green Line is the first to see the new text-to-speech software, previously seen on the Type 9s, replace all prerecorded announcements. Other lines will soon follow.

https://huntnewsnu.com/69983/city-pulse ... bys-voice/

Brief, fair-use quote:

With the increasing need to convey “real-time information through many more channels,” the MBTA is switching to automated voice announcements, said Lisa Battiston, the deputy press secretary.

“Unfortunately, working with Frank every time a new audio file was needed was not a sustainable process — it added time and cost to an already laborious process,” Battiston said in a statement to The News. “So, for the Green Line, the decision was made to use a text-to-speech service. Over time, the MBTA intends to bring all fleets into alignment with this.”

We get less human over time sigh

Interesting how New York doesn't seem to have this problem.

NYCT has an entire department that makes real time service disruption announcements - and some of the people in that department record the AAS announcements as well.

“Unfortunately, working with Frank every time a new audio file was needed was not a sustainable process — it added time and cost to an already laborious process,”

In short, it was extra work for people. I doubt paying Frank was the dealbreaker in the T's $2.5+ billion dollar budget.

Regardless, I hope we get more announcements like the TTS ones at North Station that are a) hard to understand and b) mispronounce many place names and other words.

If they wanted to be clever, deepfake technology could be used to simulate his voice.

The advances and capabilities of text to speech have been amazing, so why not use a real-time service vs. trying to match up piecemeal recordings?

It's never been about the cost. It's the portability and flexibility. You can't always predict what's needed on a given day.

Sent from my SM-G981U using Tapatalk

I wonder if it'll be like North Station, where the voice says Con-corde and West Con-corde.

It's not hard to get it correctly - you just define a phonetic replacement...

Harvard = Haaavuhd
Concord = Conkerd or Concawd

I've done some convincing NYC accented profanity using phonetics...

Sent from my SM-G981U using Tapatalk

I miss the live announcements that you could not understand, lol

I've been on plenty of commuter rail trains lately where the conductor opens a mic in a vestibule with the door open and all you hear is wooOOOOOOOOOO! as feedback from the speakers drowns out whatever the conductor has to say.

The new PTIS on the Commuter Rail can't pronounce "is". I don't understand what's so difficult about recording announcements.

Or follow NY MTA and have local personalities and celebrities do voiceovers.