Train Single Word Multiple Times

[Mikestriken]

I know about the speech dictionary, and how it's possible to "add a new word" and "Record a pronunciation upon Finish".

But is it possible to do the exact same thing but train the word more than once? Can I say the same word multiple times and in different ways to improve the confidence level of the software, and get it to better recognize what I am saying?

Can I make my own "training" so that I can say different things during the training to perhaps get it to learn different words?

If I have gone through training 8 times (which I have), have I reached the limit of what the training can offer?

Does the software always continue to learn (even if you aren't training or adding a new word)? Will the software get better over time while using it normally (not opening the Windows Speech Recognition window)? What about when talking into the Windows Speech Recognition Window normally and dictating onto a notepad, will it learn then?

For context, I am using voice attack in DCS using the Vaicom app.

[Gary]

Hi, Mikestriken.

If you've trained the speech engine multiple times (including training individual words) and things still are not working satisfactorily, it's possible that your speech profile could just be borked up.  The speech engine is supposed to 'learn' over time (this would also include using the Windows speech app, as it shares the same engine/profile as VA), and sometimes ambient/line noise will play a part in affecting the overall quality of what it stores.  Here is something to try (this has worked for a lot of folks over the years): https://forum.voiceattack.com/smf/index.php?topic=64.0    If that doesn't help, you'll want to take a look at this thread: https://forum.voiceattack.com/smf/index.php?topic=1635.0
Look for the section labeled, 'VoiceAttack is listening, but what it hears is complete gibberish compared to what I said'.  Chances are a solution *could* be in there.

Hope that helps!

[Mikestriken]

Hi, Mikestriken.

If you've trained the speech engine multiple times (including training individual words) and things still are not working satisfactorily, it's possible that your speech profile could just be borked up.  The speech engine is supposed to 'learn' over time (this would also include using the Windows speech app, as it shares the same engine/profile as VA), and sometimes ambient/line noise will play a part in affecting the overall quality of what it stores.  Here is something to try (this has worked for a lot of folks over the years): https://forum.voiceattack.com/smf/index.php?topic=64.0    If that doesn't help, you'll want to take a look at this thread: https://forum.voiceattack.com/smf/index.php?topic=1635.0
Look for the section labeled, 'VoiceAttack is listening, but what it hears is complete gibberish compared to what I said'.  Chances are a solution *could* be in there.

Hope that helps!

1. So I take it there is no way to pronounce the same word in different ways in the speech dictionary? Would you know why this is the case out of curiosity? (Seems to be a great way to potentially train the engine)

2. I will try creating a new profile as it looks like I won't lose any progress on my old one if it doesn't solve the issue. (Don't want to waste all that time I spent teaching the first profile).

3. I did read the Frequently Addressed Topics and tried what I could (I glanced over the point regarding profiles as it mentioned a clean install of windows (which I don't have). (Point is I already tried what tips in there seemed applicable to me)

Examples of my problems (in case it might help the situation further somehow):
A lot of my problems are words like "air" being confused with "their", "there" and "they're" despite me going out of my way to use the word dictionary to pronounce "their", "There", and "They're" in a noticeably different way (I am still saying the word, just emphasizing certain parts of it especially to distinguish it from "air", I also did "air" ofc). In the end, I kinda gave up and just told the program to recognize me saying "their" instead of "air".

Then after that, I started having the issue of it getting the phrase correct but not being confident enough to activate the command. I am aware I could lower its confidence level but it's at the minimum recommended confidence level according to the Vaicom manual.

In general, it doesn't feel smooth, and it feels like I have to strain my voice to get it to be recognized by the computer easier (I was hoping to be able to communicate a bit more naturally, like when I use the built-in speech engine on my phone to write messages).