Direct Speech Recognition

[asabi]

I have the following code:

'this in the form_load
DirectSR1.GrammarFromFile App.Path & "g:\stuff\vb\msagents\grammer.txt"
DirectSR1.Activate

' this in the control
Private Sub DirectSR1_PhraseFinish(ByVal flags As Long, ByVal beginhi As Long, ByVal beginlo As Long, ByVal endhi As Long, ByVal endlo As Long, ByVal Phrase As String, ByVal parsed As String, ByVal results As Long)
MsgBox "hh"
If Phrase = "HI" Then
'Do Something
ElseIf Phrase = "Bye" Then
End
Else
MsgBox "Here"
End If
'DirectSR1.Speak "Hello"

End Sub


and I have the text file (grammer.txt):
[Grammar]
type=cfg
[<start>]
<start>=HI
<start>=Bye

It doesn't work !!

How do I make it listen to me ???

thank you

Is this what you meant to have wrote?:

DirectSR1.GrammarFromFile App.Path & "g:\stuff\vb\msagents\grammer.txt

Shouldn't it have been either:

DirectSR1.GrammerFromFile App.Path & "\grammer.txt"

or

DirectSR1.GrammerFromFile "g:\stuff\vb\msagents\grammer.txt"

??

But I don't see any problems with your actual PhraseFinish event. If it still doesn't work, try a simpler Select Case to choose the Phrase and try simpler words first. Like just "1" and "2"


Sunny

[asabi]

Are you suppose to press anything when you speak to the microphone ? or is the "channel" open all the time, and you just need to speak ?

You are right about the syntax .. :-) thanks, it was late at night ..
thanks

[asabi]

o.k, forget about the last one ..

It is working now..

I would like to know what "tags" I can use other then the "<start>" in my grammer.txt file (if there are any) or if you know where I can find good documentation about it ..

THANK YOU ! :-)

[HarryW]

I know it's petty, but 'grammar' doesn't have an 'e' in it

[asabi]

well, in my GRAMMER there is :-) !!!
You can have yours and call it what ever you like

by the way, how do you do the "faces"/ bold writing in the body of the message ?

[Doc Scheinder]

There is a possibility that your coding does work. Unfortunately Speech Recognition has several draw backs, with the major one being pronounciation. The progam may be hearing you OK but doesn't recognize what it thinks it heard. There is also a possibility of a faulty sound card. Run your speech engine's test feature to make sure evrything is turned on and working properly. I have a speech engine that I downloaded from Microsoft's Developer Web Site. It's called Microsoft Speech SDK / Microsoft Voice, this program has so much support that the impossible will become possible, and it readily connects to VB. Good luck. Any clue as to what you are working on? I've been working with voice activation for over 3 years, maybe I can help.

[asabi]

Well, for now I am making something for my own use.

It is a MSagent that is like a "desktop friend", it reads me the things I put in the clipboard, from time to time it says stuff and moves (randomly).


The speech recognition is working now, and to be quite honest it even understands me :-)

things like "date", "time" ,"Joke" ...



you say that this SDK is a better engine ?
I will try it out.
do you have an exact link that I can reffer to ?

Thank you

[asabi]

Well, I downloaded the SDK engine, installed it and now I want to use the dictation control.

I put a control on my form (vdict1) and triggered the activate method ...
and put some code like:

Private Sub Vdict1_PhraseFinish(ByVal flags As Long, ByVal phrase As String)
Text1.Text = Text1.Text & phrase
End Sub

but nothing happened when I spoke ... so I went to the help files, and they say that the "mode" should be set to:

VSRMODE_DCTONLY

but what the hell is this constant ???

I will appriciate it if someone can send me a demo for it, or post something here ..

Thank you

I have sent you a demo...say 1, 2 or 3

With your last one, you've activated it, by did you load a grammar file?

Sunny

Hi there.

Whre can i get direct speech control, so i can make my self "Talking computer" , thank´s

numibesi

You can download it from Microsoft Agent's website:

http://msdn.microsoft.com/workshop/i...nt/default.asp


Sunny

When using GrammarFromFile you don't need all that <start> stuff. You would only need that when using GrammarFromString.

[Doc Scheinder]

You seem to be having a rough go at it. When I started doing Speech Recognition. I had similar errors that seemed to come from no where. There's are alot of simple coding that comes with Microsoft SAPI 4.0. I didn't bring any voice activation coding to work with me. But i can send you some tomorrow. But here is some of the coding that comes with the program for VB.

To test if your program is listening and hearing correctly use this code:

Dim gMymenu As Long

Private Sub UpdateList()
Dim i As Long
Dim command As String
Dim description As String
Dim category As String
Dim flags As Long
Dim action As String

List1.Clear
tcount = Vcommand1.CountCommands(gMymenu)
For i = 1 To tcount
Vcommand1.GetCommand gMymenu, i, command, description, category, flags, action
List1.AddItem command
Next i
End Sub

Private Sub Command1_Click()
If (Text2.Text <> "") Then
Vcommand1.AddCommand gMymenu, 1, Text2.Text, "when you say " + Text2.Text, "listen list", 0, ""
End If
UpdateList

End Sub

Private Sub Command2_Click()
If (List1.ListIndex >= 0) Then
Vcommand1.Remove gMymenu, List1.ListIndex + 1
End If
UpdateList
End Sub

Private Sub Command3_Click()
If Command3.Caption = "Listen" Then
Vcommand1.Activate gMymenu
Command3.Caption = "Stop Listening"
List1.Enabled = True
Else
Vcommand1.Deactivate gMymenu
Command3.Caption = "Listen"
List1.Enabled = False
End If

End Sub

Private Sub Form_Load()
Vcommand1.Initialized = 1
gMymenu = Vcommand1.MenuCreate(App.EXEName, "state1", 4)
Vcommand1.Enabled = 1
Vcommand1.AddCommand gMymenu, 1, "red", "when you say" + "red", "listen list", 0, ""
Vcommand1.AddCommand gMymenu, 1, "blue", "when you say" + "blue", "listen list", 0, ""
Vcommand1.AddCommand gMymenu, 1, "green", "when you say" + "green", "listen list", 0, ""
UpdateList
Vcommand1.Activate gMymenu
End Sub


Private Sub Form_Unload(Cancel As Integer)
Vcommand1.ReleaseMenu gMymenu
End Sub

Private Sub Vcommand1_CommandRecognize(ByVal ID As Long, ByVal CmdName As String, ByVal flags As Long, ByVal action As String, ByVal NumLists As Long, ByVal ListValues As String, ByVal command As String)
Label3.Caption = command
End Sub

You can also modify this coding to perform text to speech. In my long journy to the land of voice activation this coding opened the doors that had been closed to me for a long time. Let me know what happens.

When I was playing with MSAgent I fiddled for hours with no luck from the voice recognition, the key was to press scroll lock to make the agent listen.

I find it hard to believe that the Scroll lock key needs to be pressed, mainly because I've used voice recognition without any problems, and also, if a key was needed to be pressed before it listens, it wouldn't be Scroll lock - its a very unconventional place to put a key for such a purpose as it is usually not within fingers' reach of the typist like Control or shift.

Sunny

Sunny, the following is a cut and paste from troubleshooting Agents from MSDN, you will find it makes for dull bed time reading but is nevertheless a good source of info!


Verify that the Agent-enabled application or Web page supports speech input. Not all pages or applications support speech input. Press and hold the Listening key. Typically this will be the Scroll Lock key unless you changed it. A pop-up window should appear under the character. The text in the tip will tell you the listening state of the character. If no tip appears, either the application or Web page does not support speech input, or you don't have a compatible speech engine installed. If the tip does appear and indicates the character is listening, speak one of the character's voice commands. If you do not know what voice commands are available, release the Listening key and right-click the character.


SDK beta 5 download at http://www.microsoft.com/speech/

[Edited by MadWorm on 09-21-2000 at 06:15 AM]

[dekelc]

What is this DirectSP control and where can I get it?

Maybe someone can clear this up, there seems to be two types of TTS and STT here... one seems to be Microsoft Agent (involving the characters) which MadWorm mentioned, and there seems to be another one which I think is the one asabi is having problems with (he is trying to use Voice Dictation).

MS Agent:
http://msdn.microsoft.com/workshop/i...nt/default.asp

'the other one'
http://www.microsoft.com/IIT/download/speechengines.htm


Sunny

[Edited by sunnyl on 09-21-2000 at 07:10 AM]

[asabi]

Well said sunnyl !

You are right, and I still can't make the dictation to work (the commands and the voice recognition are working perfectly..) :-)

Thank you for helping

I've downloaded the SAPI SDK and help files. Although it says it has sample code, the 'sample code' for voice dictation is a full blown application, it wasn't easy to understand (and I still don't), but this code seems to work when dictating:

Option Explicit

Private Sub Form_Load()
    Vdict1.Speaker = "Sunny"     'uses my profile?
    Vdict1.Mode = 32       'dictation only mode (?)
    Vdict1.Activate
End Sub

Private Sub Vdict1_PhraseFinish(ByVal flags As Long, ByVal phrase As String)
    Text1.Text = Text1.Text & " " & phrase
End Sub

When I ran it and said "testing testing 1 2 3" I got a result of "testing testing one to three" several times.

Sunny

[asabi]

GREAT !!!

Thank you very much !

It is working at last ... but it is really funny ...

I say "go" it writes "no"

I say "wolf" it writes "would"

But I guess that's as good as it can get :-)

Thank you very much for all !!!

Cheers

You should train it more so that it becomes more accurate. I tried running the program with and without the lines Vdict1.Speaker and there seemed to be a bit of difference. I also find that I actually have to set the Accuracy vs Speed setting right to the speed end to get a reasonable response time. I have a PII 350.


Sunny

It is annoying.

Last time I've make the same thing for word (Auto-type) but Guess what? I gave the command "Type R V S E T Y D E O R R" and it typed "ARVEASEEIEYEEORR".

Expect it, I trained it 3 or 4 times...

[asabi]

I also have a PII 350, it has 256MB ram, and it still takes it a sec or so before it reacts .., and when it does ... :-) it almost seems as if it has its own imagination ..

[dekelc]

Where can I get one of these control?
It seems realy cool to add them to my app.

[asabi]

MSAgents

http://msdn.microsoft.com/workshop/c...nt/default.asp


Speech

http://www.microsoft.com/IIT/download/speechengines.htm

[nitrolic2]

I'm downloading: Microsoft Speech Recognition Engine for PC command &
control applications

any help on how i can use that?

[sparkle52988]

Whenever I try to use the speech dictation control from Microsoft, as well as some other MS voice controls, VB says that "another program is requesting the same control", or something like that, and I can never put down one of those controls. What should I do, because I REALLY want to use speech dictation in my software? Please e-mail me back at [email protected]

[dekelc]

The activeX from MS is an exe, hoe can I use it?

dekelc, MS has many different packages for speech recognition and text to speech, which one have you downloaded?

Sunny

[dekelc]

I've downloaded the Microsoft Speech Recognition Engine for PC command & control applications control.

[dekelc]

I've activated the exe for the
Microsoft Speech Recognition Engine for PC command & control applications control,
it show me a few files, non of them is ocx.
What am I doing wrong?

[Doc Scheinder]

I've tried that tag bit, but when your command list get's long you will run into a line continuation error. I had to resort to storing the commands in a list box. This works, the only drawback is, if the computer is turned off and back on, the SDK doesn't automatically load the commands from the list box, however the list box still holds the commands. This suggests that the command list is not permantly saved. Any of you CDM/W's know how to solve this problem and permanently save the commands storeed in the list box?

dekelc, start a new VB project, goto Project => Components and in the list look for Microsoft Direct Speech Recognition. When you find it, check its box and click ok then you should have a new control on your toolbox.

Doc, Can you do it manually by saving the contents of the list to a file when your program closes and reloads the file when it starts up?

Sunny

[dekelc]

There is no control like that in the components box!

Don't panic! It just means the package you downloaded didn't have DirectSR (Speech Recognition). I suggest you download the largest package, which has everything... http://www.microsoft.com/iit

For the time being, look for these other ones, at least you'll know you've installed something.

Microsoft Direct Speech Recognition
Microsoft Direct Text-to-Speech
Microsoft Voice Dictation (theres an example i wrote earlier in the thread for this)
Microsoft Voice Command
Microsoft Voice Text


Sunny

[Doc Scheinder]

Originally posted by sunnyl
dekelc, start a new VB project, goto Project => Components and in the list look for Microsoft Direct Speech Recognition. When you find it, check its box and click ok then you should have a new control on your toolbox.

Doc, Can you do it manually by saving the contents of the list to a file when your program closes and reloads the file when it starts up?

Sunny

[Doc Scheinder]

Sunny,

I have an idea of what you are saying, but still a little confused as to how it will load. Do you have any sample code that displays the process?

I may be wrong here, but you say that if there are too many commands in the file, it will go into a line continuation error. And you loaded the commands into a listbox instead. but the next time around, the commands are in the listbox but they are not recognised, right?

Then you don't have to use GrammarFromFile, you can use GrammarFromString. You can load each of the commands in the listbox with GrammarFromString if the file is too long and causes an error.


Sunny