News
box
box
Events
box
box
Community
box
box

MacSpeech Forums and Discussion » MacSpeech Scribe » Using Scribe

deleting problem words?

(6 posts)
  • Started 1 month ago by Chris Holdsworth
  • Latest reply from fiverlocker
  1. Chris Holdsworth
    Member

    Is there a way to delete problem words from the built in dictionary like you can in dragon naturally speaking?

    I use Dragon at the moment for work. I dictate inventories on properties, and it is pretty accurate. But it's windows only, hence why I am trying to move to Scribe.

    The problem is that it insists on transcribing wrong words even after numerous attempts to train it, for example:

    Ceiling comes out as seeling
    Door to... comes out as daughter

    these are words Iam never going to use and in Dragon you just delete them and you don't have a problem again.

    Any thoughts??

    cheers

    Posted 1 month ago #
  2. There is a Vocabulary Editor which you should take a look at. However, it only allows you to add and/or train words (and very short phrases). There is not a way to delete built-in words. You can train "problem" words so Dictate/Scribe know exactly how you pronounce those words.

    Posted 1 month ago #
  3. Chris Holdsworth
    Member

    many thanks. I'll take a look.

    Posted 1 month ago #
  4. CSLong
    Member

    I'm afraid jayg suggestion will not satisfy you. There are other posts concerning editing the inbuilt dictionary (i.e. it is not possible).

    My wife's name is 'bruna', which is contained in the in-built dictionary. Despite extensive training I still get 'bruner' another in built word. I have even tried to train the word 'bruner' as 'rubbish', but it still appears when I say 'bruna'.

    Basically you're stumped .......

    Posted 4 weeks ago #
  5. My wife's name is 'bruna', which is contained in the in-built dictionary. Despite extensive training I still get 'bruner' another in built word. I have even tried to train the word 'bruner' as 'rubbish', but it still appears when I say 'bruna'.

    If the training isn't close to the word, Dictate will ignore the training. In this case it would ignore "rubbish".

    Try training "bruner" as "Broon Urr". You can exaggerate the enunciation, but it still needs to be pronounced within reason as close to the actual pronunciation. I would also add the whole name: first name and last name, and then train that as a phrase.

    I would also create a text macro with a command name of "Bruner" which will print "Bruna" when recognized.

    Posted 4 weeks ago #
  6. fiverlocker
    Member

    I would second the request to be able to delete words from the custom dictionary. I can't see what the downside is? If you accidentally delete single words, you can always re-train them, and if you accidentally delete a huge amount, you can always re-install the data from disc.

    This insistence that I create text macros for names is ludicrous, recognising text and learning my pronunciation is what these programs are all about.

    I find both Dictate and Scribe terrible with names of people and places. It also seems to be incapable of learning my preferences, i.e. my husband's name is Stuart, not Steward! Arg!

    Fiver

    Posted 2 weeks ago #

RSS feed for this topic

Reply

You must log in to post.