Start a new topic

Adding tags in a second run when using speech recognition

Michael asked:


>re ‘post-tagging’: I suppose it would actually be very easy to try this in CafeTran, as we could just switch on Translation > Tags > Automatic transfer of remaining tags, so all tags would just get inserted at the end of each segment as we went along. We could then fix them at the end.


That's indeed how I'd do it too.


However, there's room for improvement here too (where isn't?):

  • A new "Dictation" mode that doesn't insert a single tag (so that you don't have to remove a single one in the second and tagging run)
  • Use of the glossaries, regular expressions and non-translatables to automatically wrap paired tags around matching target terms/expressions (if the source terms/expressions are tagged). Of course, the tag numbers have to be correct (e.g. CafeTran automatically inserts tags 1,2 and 5 and 6, translator inserts 3 and 4 manually).

Good ideas. 


Another very good one would be a shortcut for wrapping a selection with the relevant tags: basically what dragging over a selection with Ctrl held down does now, but via a KBS.

The user selects a fragment with a keyboard (not with a mouse) and presses a keyboard shortcut, the selected fragment is tagged. Correct?


Igor 

I think that that one is essential indeed.


I guess the workflow with speech recognition would be:


  • Translator speeks: Select speech recognition.
  • Speech software selects 'speech recognition'.
  • Translator speeks: Tag this

A voice command 'Tag this' has to be mapped to te keyboard shortcut assigned to wrapping any selection with tags.

@Igor: exactly.

"The user selects a fragment with a keyboard (not with a mouse) and presses a keyboard shortcut, the selected fragment is tagged. Correct? "

But the user could also select it with the mouse.


You could then also add a voice command to trigger this keyboard shortcut, to tag selections.

@Hans: Yes, ideally. However, CafeTran currently doesn't allow Dragon to ‘get at’ the text in the target pane, so it is not yet possible to select words/phrases by voice. This is called ‘Full Text Control’, or ‘Select-and-Say’ in Dragon lingo.


I'm not sure if other speech recognition software can select words/phrases in the CafeTran target pane by voice.

>But the user could also select it with the mouse.


That's already implemented! Press Command, click, drag, release.

@Juniper's dad:


@Hans: Yes, ideally. However, CafeTran currently doesn't allow Dragon to ‘get at’ the text in the target pane, so it is not yet possible to select words/phrases by voice. This is called ‘Full Text Control’, or ‘Select-and-Say’ in Dragon lingo.


I'm not sure if other speech recognition software can select words/phrases in the CafeTran target pane by voice.


==


As demonstrated in one of those awfully exciting screencasts of mine, this is no problem at all with Dictation in OS X. Nor is defining new voice commands to emulate key presses, like 'Tag this'.


I'm starting to fear that you didn't watch my movies. For whom else but for you did moi record them?!

Yes, I know that. All I meant was: to use the proposed "tag this" / "tag selection" shortcut, you need to select something. Selecting can be done via mouse or keyboard. If the Ctrl/Command key was pressed down before you selected it with the mouse, when you let it go, it would get tagged (using the already present function). However, if the Ctrl/Command key was not pressed down before you selected it with the mouse, you could use the new new KBS.


 

This one: http://cafetran.wikidot.com/dictation


OK, just watched it. Lucky bastard: select-and-say is indeed built into the Mac's Dictation app, and works inside CT. 

>Lucky bastard?


I'd rather say that I've carefully chosen my OS and was not afraid to make the jump from Windows. Never regretted it. Au contraire, mon frère.


And of course I have full access to all my Windows apps (that I hardly use, but that's because I have Mac alternatives).


End of commercial

Login to post a comment