Speech-to-text changed, is not as good now

(EDIT: Originally I said "text-to-speech", but I meant "speech-to-text".)

I have a Samsung S9, with Verizon, on Android 10. I use the Google speech-to-text engine.

A couple of months ago something changed - there was some software update that really made the speech-to-text function a lot worse. I'm not sure if it's a Google thing, a Verizon thing, a Samsung thing, an Android thing, or what. Pretty sure it's the latter - Android, as it affects all the apps.

Specifically - here's the new problematic behavior:

• Very often when I speak a sentence, there's a bug that causes the sentence to be duplicated. E.g. I'll say "I'll be home in 15 minutes", but the text that shows up is "I'll be home in 15 minutes; I'll be home in 15 minutes"
• There's also a new bug where it doesn't translate punctuation correctly. E.g. I'll say "period" at the end of the sentence - sometimes it changes this to '.', but sometimes it just leaves it as "period".
• In general it's just not as good at detecting what I'm saying, as it used to be - getting words wrong more often.
• Often times it just doesn't detect anything at all - just sit there dumbly and time out as if I didn't say anything, and time out, even though I have good internet connection.
• Or often it'll just fail with a double-beep for some reason, even though I have good internet connectivity.
• The functionality to be able to switch back and forth between manual typing and text-to-speech is now gone. Now once I start typing, the microphone icon no longer shows up, and thus I can't switch to speech. (This is a desirable thing to have - e.g. to speak a sentence, go back and correct a word via keyboard, then continue speaking the rest; this is no longer do-able)

Anyone know what update may have broken this, and if there's anything that can be done about it?

Thanks.

 

Thanks for the tips, though I generally have found speech-to-text a pretty useful function - certainly more efficient than typing when doing a lot of text; even with the recent changes that made it more crappy. So I plan to keep using it either way; just hoping that someone can enlighten as to what's changed to make it worse, and perhaps how it can be improved.

 

I'll try that, though right off the bat I can tell it won't help most of the problems.. I tried doing a test, and it double-beeped at me as failed on the second try (I'm on good Wi-Fi). Also the problem where the microphone icon isn't there when using the keyboard still exists. (it can be brought back, but only when exiting out of the keyboard first; something new and an extra step added)

Most or all of the problems aren't actually the text-to-speech engine, which is something on the back end I believe, but more an application thing - e.g. the application fails to even invoke the engine (the beep failure), or it accidentally invokes it twice (duplicate text).

 

Oh shoot - I did mean speech-to-text. Thanks. Fixed.