Working Sample of Speech Recognition from a file?

Please direct me to a working sample of Speech Recognition from a file 
preferably
with (as many as possible of) the following traits:

   Runable on  XP/2003 (not just Vista)
        installation of some specific SDK/Net version is ok here as long as 
that works
        on the older OSs

  .Net (recent is better but previous item takes precedence)

  C# (or C++ then VB etc)

  Using MP3 file (but if this isn't possible please tell me so that I can 
plan to convert
        to WAV (etc)

My initial searches found either partial or non-working samples.
(I.e., wouldn't compile in VS2008)

I intend to train the engine with a fixed set of recorded responsed (less 
than
50), and then expect it to recognize these distinctive short phrases 
combined
in continous speech of 5-10 of these elements.

Thanks in advance.

--
Herb Martin

I don't know if this helps. It's C# and does work. Although I haven't trained 
my engine properly so result is not very good!

//Class members
SpInProcRecoContext sprec = null;
ISpeechRecoGrammar spg = null;
//Initialiozation
sprec = new SpInProcRecoContext();
sprec.Recognition += new 
_ISpeechRecoContextEvents_RecognitionEventHandler(SPRec_Recognition);
sprec.StartStream += new 
_ISpeechRecoContextEvents_StartStreamEventHandler(SPRec_StartStream);
sprec.EndStream += new 
_ISpeechRecoContextEvents_EndStreamEventHandler(SPRec_EndStream);
spg = sprec.CreateGrammar(null);
spg.DictationLoad(null, SpeechLoadOption.SLOStatic);
//Event handlers
void SPRec_EndStream(int StreamNumber, object StreamPosition, bool 
StreamReleased)
{
                //whatever you want to do
}
void SPRec_StartStream(int StreamNumber, object StreamPosition)
{
                //
}
void SPRec_Recognition(int StreamNumber, object StreamPosition, 
                SpeechRecognitionType RecognitionType, ISpeechRecoResult 
Result)
{
    spg.DictationSetState(SpeechRuleState.SGDSInactive);
    string s = "";
    for (int i = 0; i < Result.PhraseInfo.Elements.Count; i++)
    {
         ISpeechPhraseElement e = Result.PhraseInfo.Elements.Item(i);
         s += e.DisplayText + " ";
    }
    MessageBox.Show(s);
}
 //Open file
SpFileStream fs = new SpFileStream();
fs.Open(filename, SpeechStreamFileMode.SSFMOpenForRead,false);
sprec.Recognizer.AudioInputStream = fs;
spg.DictationSetState(SpeechRuleState.SGDSActive);


"Herb Martin" wrote:

> Please direct me to a working sample of Speech Recognition from a file 
> preferably
> with (as many as possible of) the following traits:
> 
>    Runable on  XP/2003 (not just Vista)
>         installation of some specific SDK/Net version is ok here as long as

> that works
>         on the older OSs
> 
>   .Net (recent is better but previous item takes precedence)
> 
>   C# (or C++ then VB etc)
> 
>   Using MP3 file (but if this isn't possible please tell me so that I can 
> plan to convert
>         to WAV (etc)
> 
> My initial searches found either partial or non-working samples.
> (I.e., wouldn't compile in VS2008)
> 
> I intend to train the engine with a fixed set of recorded responsed (less 
> than
> 50), and then expect it to recognize these distinctive short phrases 
> combined
> in continous speech of 5-10 of these elements.
> 
> Thanks in advance.
> 
> --
> Herb Martin
> 
> 

"raminarya" <raminarya@discussions.microsoft.com> wrote in
message 
news:6F9A0AE2-38CB-497A-A123-CF07968F07E4@microsoft.com...
>I don't know if this helps. It's C# and does work. Although I haven't 
>trained
> my engine properly so result is not very good!

I didn't even come close to getting a compile -- even by trying to look
up the missing class definitions and such...

> //Class members
> SpInProcRecoContext sprec = null;
> ISpeechRecoGrammar spg = null;
> //Initialiozation
> sprec = new SpInProcRecoContext();
> sprec.Recognition += new
> _ISpeechRecoContextEvents_RecognitionEventHandler(SPRec_Recognition);
> sprec.StartStream += new
> _ISpeechRecoContextEvents_StartStreamEventHandler(SPRec_StartStream);
> sprec.EndStream += new
> _ISpeechRecoContextEvents_EndStreamEventHandler(SPRec_EndStream);
> spg = sprec.CreateGrammar(null);
> spg.DictationLoad(null, SpeechLoadOption.SLOStatic);
> //Event handlers
> void SPRec_EndStream(int StreamNumber, object StreamPosition, bool
> StreamReleased)
> {
>                //whatever you want to do
> }
> void SPRec_StartStream(int StreamNumber, object StreamPosition)
> {
>                //
> }
> void SPRec_Recognition(int StreamNumber, object StreamPosition,
>                SpeechRecognitionType RecognitionType, ISpeechRecoResult
> Result)
> {
>    spg.DictationSetState(SpeechRuleState.SGDSInactive);
>    string s = "";
>    for (int i = 0; i < Result.PhraseInfo.Elements.Count; i++)
>    {
>         ISpeechPhraseElement e = Result.PhraseInfo.Elements.Item(i);
>         s += e.DisplayText + " ";
>    }
>    MessageBox.Show(s);
> }
> //Open file
> SpFileStream fs = new SpFileStream();
> fs.Open(filename, SpeechStreamFileMode.SSFMOpenForRead,false);
> sprec.Recognizer.AudioInputStream = fs;
> spg.DictationSetState(SpeechRuleState.SGDSActive);
>
>
> "Herb Martin" wrote:
>
>> Please direct me to a working sample of Speech Recognition from a file
>> preferably
>> with (as many as possible of) the following traits:
>>
>>    Runable on  XP/2003 (not just Vista)
>>         installation of some specific SDK/Net version is ok here as
long 
>> as
>> that works
>>         on the older OSs
>>
>>   .Net (recent is better but previous item takes precedence)
>>
>>   C# (or C++ then VB etc)
>>
>>   Using MP3 file (but if this isn't possible please tell me so that I
can
>> plan to convert
>>         to WAV (etc)
>>
>> My initial searches found either partial or non-working samples.
>> (I.e., wouldn't compile in VS2008)
>>
>> I intend to train the engine with a fixed set of recorded responsed
(less
>> than
>> 50), and then expect it to recognize these distinctive short phrases
>> combined
>> in continous speech of 5-10 of these elements.
>>
>> Thanks in advance.
>>
>> --
>> Herb Martin
>>
>>
>> 

You should add SAPI (MS Speech in COM objects) to your references and then 
"using SpeechLib;" to your source code. Sorry I should have mentioned.
You 
can download a sample (MySR.zip) here:
http://www.csit.carleton.ca/~arya/tmp/

If you are using .NET 3.5 then there are speech classes under System.Speech 
but I'm using .NET 2.

"Herb Martin" wrote:

> 
> 
> I didn't even come close to getting a compile -- even by trying to look
> up the missing class definitions and such...
> 

Claudius Coenen posted this recently: 
http://msdn2.microsoft.com/en-us/library/system.speech.recognition.speechrecogni
tionengine.setinputtowavefile.aspx

Just use SetInputToWaveFile instead of SetInputToDefaultAudioDevice. You'll 
need to figure out how to convert to WAV though.

Thanks,
Phil.

-- 
This post is provided "AS IS" with no warranties, and confers no
rights.

"Herb Martin" <news@learnquick.com> wrote in message 
news:uF7VgUXaIHA.1168@TK2MSFTNGP02.phx.gbl...
> Please direct me to a working sample of Speech Recognition from a file 
> preferably
> with (as many as possible of) the following traits:
>
>   Runable on  XP/2003 (not just Vista)
>        installation of some specific SDK/Net version is ok here as long as

> that works
>        on the older OSs
>
>  .Net (recent is better but previous item takes precedence)
>
>  C# (or C++ then VB etc)
>
>  Using MP3 file (but if this isn't possible please tell me so that I can 
> plan to convert
>        to WAV (etc)
>
> My initial searches found either partial or non-working samples.
> (I.e., wouldn't compile in VS2008)
>
> I intend to train the engine with a fixed set of recorded responsed (less 
> than
> 50), and then expect it to recognize these distinctive short phrases 
> combined
> in continous speech of 5-10 of these elements.
>
> Thanks in advance.
>
> --
> Herb Martin
>