Follow-up to this thread - this is way more specific, so hopefully worthy of its own thread. I think wildcards are the best option for my skill level (basically none), and have gotten a good chunk of what I wanted to accomplish done with those.

An issue I’ve run into and can’t seem to google my way out is making TTS pronounce acronyms in a sensible way. For example “PACU” (post-anesthesia care unit) is usually vocalized as “pack-you” but my TTS software likes to say things like “pace-uh”. Or “PO” (latin abbreviation for ‘by mouth’) is vocalized by just saying the letters, but TTS says “Poe”. Stuff like that.

When the TTS comes across a capitol letter with a space on either side, it just pronounces the letter, so I’d still lose things like “pack-you” but at least hearing it spell out “pee ay see you” would make sense, vs “pace-uh” which is gibberish and confusing at high playback speeds.

Best I’ve come up with so far is <([A-Z]{2})> on the Find side, but that’s only spotting the two character terms like PO, and ignoring the longer ones… I’d hoped it would see PACU and detect PA, AC, and CU as three distinct sets of two that could cobbled into “P A C U”.

Nothing I’ve done on the Replace side comes close to working. It either does nothing at all, or it’ll do something like turn “PO” into <([A- Z]{2})>. Not sure if preserving the original characters is something A-Z is actually capable of - seems not, but I’m kind of an idiot with stuff like this, so any tips would be appreciated!

Thank you!

  • Sterile_Technique@lemmy.worldOP
    link
    fedilink
    English
    arrow-up
    2
    ·
    1 month ago

    The more I play with this the more I love it. Just modified your code to produce:

    Find ([0-9])/([0-9])

    Replace \1 over \2

    which changes things like “blood pressure 129/81” to “blood pressure 129 over 81” to prevent TTS from saying a fraction like “blood pressure one-hundred and twenty nine eighty firsts”

    …will have to check for actual fractions first, but in the context, two numbers separated by a slash is pretty much always blood pressure.

    Hell yeah!