1. Speech recognition
Whilst you're visiting the forum, Chris, I'd like to thank you for
posting the speech.dll - Any hints on how I can create my own
(assuming this can be done by logical comparison without a steep
C learning curve)?
I built an editor around it which incorporates both TTS and SR, and
toggles between the two using VoiceShutdown and VoiceInit switching
one off before the other is switched on. I take it that this is the
usual practice and that speech.dll contains a sufficiently full
repertoire of available commands.
I also found a bug in the demo SpeechToText.exw which causes a crash
upon uttering percent sign, pound sign. Perhaps Voice_Check should
read as follows;
procedure Voice_Check(integer self, integer event, sequence params)
object voice
object block
if Voice_Result()=1
then
block = Voice_Block()
voice=getRichText(Text,-1)
setText(Text, voice & " " & block)
setFocus(Text)
setIndex(Text, -1 )
end if
end procedure
setHandler(MainWin, w32HTimer, routine_id("Voice_Check"))
Regards,
Bob
2. Re: Speech recognition
- Posted by C Bouzy <eucoder at hotmail.com>
Jan 23, 2006
-
Last edited Jan 24, 2006
Bob Thompson wrote:
>
>
> Whilst you're visiting the forum, Chris, I'd like to thank you for
> posting the speech.dll
It is my pleasure, and I am glad you found use for it :)
> Any hints on how I can create my own
> (assuming this can be done by logical comparison without a steep
> C learning curve)?
>
You can try the Microsoft Speech SDK, that is where I started.
> I built an editor around it which incorporates both TTS and SR, and
> toggles between the two using VoiceShutdown and VoiceInit switching
> one off before the other is switched on. I take it that this is the
> usual practice and that speech.dll contains a sufficiently full
> repertoire of available commands.
>
Cool....you should add it to the archive.
> I also found a bug in the demo SpeechToText.exw which causes a crash
> upon uttering percent sign, pound sign.
Thanks for catching this! If possible please send the updated file to Robert
so he can update the Speech package on the site.
----If you continue to do what you have always done,
you will get what you have always gotten.----
3. Re: Speech recognition
- Posted by ags <eu at 531pi.co.nz>
Jan 23, 2006
-
Last edited Jan 24, 2006
Bob Thompson wrote:
> I also found a bug in the demo SpeechToText.exw which causes a crash
> upon uttering percent sign, pound sign. Perhaps Voice_Check should
> read as follows;
Are you sure you weren't saying something like "$#%@ing computer!"?
I've found in my line of work that people who are angry with computers seem to
have ones that crash most often (while the same computers don't crash on other
people)...
Gary
(apologies for levity)
4. Re: Speech recognition
ags wrote:
>
> Bob Thompson wrote:
> > I also found a bug in the demo SpeechToText.exw which causes a crash
> > upon uttering percent sign, pound sign. Perhaps Voice_Check should
> > read as follows;
>
> Are you sure you weren't saying something like "$#%@ing computer!"?
>
> I've found in my line of work that people who are angry with computers seem
> to
> have ones that crash most often (while the same computers don't crash on other
> people)...
>
> Gary
> (apologies for levity)
Lol. SR doesn't understand my accent so it might come out with something
like that one day.
To be accurate, though, sprint doesn't like % as the third argument for
obvious reasons.
Regards,
Bob
5. Re: Speech recognition
C Bouzy wrote:
> Cool....you should add it to the archive.
So how does one zip a bunch of files and upload them? I could do this
as a working demo for newbies to Windows API once I have refined it a bit.
> Thanks for catching this! If possible please send the updated file to Robert
> so he can update the Speech package on the site.
Comments as above so I hope Rob can save the following to SpeechToText.exw
By the way, I've also changed voice and block to sequences.
include Speech.ew
include Win32Lib.ew
without warning
---------------------------------------------------------------------------------------------
constant
MainWin = create(Window, "Speech To Text", 0, Default, Default, 320, 180,
{WS_MINIMIZEBOX,WS_SYSMENU}),
Text = create(RichEdit, "", MainWin, 5, 8, 304, 100, 0),
Instructions = create(CText, "Talk into your microphone to see the
text.",MainWin, Center, 120, 304, 20,0 )
---------------------------------------------------------------------------------------------
procedure Open_MainWin(integer self, integer event, sequence params)
Voice_Init()
setTimer(MainWin, 0, 100)
end procedure
setHandler(MainWin, w32HOpen, routine_id("Open_MainWin"))
---------------------------------------------------------------------------------------------
procedure Voice_Check(integer self, integer event, sequence params)
sequence voice, block
if Voice_Result()=1
then
block = Voice_Block()
voice=getRichText(Text,-1)
setText(Text, voice & " " & block)
setFocus(Text)
setIndex(Text, -1 )
end if
end procedure
setHandler(MainWin, w32HTimer, routine_id("Voice_Check"))
---------------------------------------------------------------------------------------------
procedure Close_MainWin(integer self, integer event, sequence params)
Voice_Shutdown()
end procedure
setHandler(MainWin, w32HClose, routine_id("Close_MainWin"))
---------------------------------------------------------------------------------------------
WinMain(MainWin, Normal )
Regards,
Bob
6. Re: Speech recognition
Bob Thompson wrote:
> C Bouzy wrote:
>
> > Cool....you should add it to the archive.
>
> So how does one zip a bunch of files and upload them? I could do this
> as a working demo for newbies to Windows API once I have refined it a bit.
>
> > Thanks for catching this! If possible please send the updated file to Robert
> > so he can update the Speech package on the site.
>
> Comments as above so I hope Rob can save the following to SpeechToText.exw
Please don't give me individual files, or pieces of files to
patch in. Re-zip the entire original package. Normally the original author
would do this, but he seems to be ok with your doing it.
Click the up-arrow at the right side of Bouzy's
speech program, and upload a new .zip. You just have to fill in
a sentence or two stating what the change was. It will warn you
that you aren't Bouzy. That's OK, I know who you are.
Regards,
Rob Craig
Rapid Deployment Software
http://www.RapidEuphoria.com
7. Re: Speech recognition
Bob Thompson wrote:
> To be accurate, though, sprint doesn't like % as the third argument for
> obvious reasons.
Whoops, that wasn't very accurate, sprintf only takes 2, but you get the
gist, I'm sure.
Regards,
Bob
