Speech Dictation
-
Gud Day all I've been trying to solve this problem for quite some time now...I hope someone might have a clue or at least a place where I can learn more about the issue; you see I've managed to make the speech dictation provided by MS to work using VB.NET, and it writes everything perfectly; what I'm trying to do is disable the grammar and the vocabulary list, since it is using the English words; what I want to do is basically type out what the user says with out forming up the correct word; or let me say type the combination of phonemes... any ideas? thnx in advance Regards Abraham
-
Gud Day all I've been trying to solve this problem for quite some time now...I hope someone might have a clue or at least a place where I can learn more about the issue; you see I've managed to make the speech dictation provided by MS to work using VB.NET, and it writes everything perfectly; what I'm trying to do is disable the grammar and the vocabulary list, since it is using the English words; what I want to do is basically type out what the user says with out forming up the correct word; or let me say type the combination of phonemes... any ideas? thnx in advance Regards Abraham
AbrahamMOKDAD wrote:
you see I've managed to make the speech dictation provided by MS to work
Which translates sound into known words.
AbrahamMOKDAD wrote:
or let me say type the combination of phonemes...
Sounds aren't mapped to a phoneme by the SAPI. Phonemes can't be interpreted, and can't be mapped to a series of characters. A single letter has a different sound in a different language, and there is no version that tries to map "generic sounds" to a generic alphabet. An alternative would be to record the sounds yourself (DirectInput) and use AI-algo to classify it (to map the sound to a/a serie of chars) The latter option is quite a challenge, I suppose. I hope that someone else posts an answer that I missed :)
I are Troll :)
-
AbrahamMOKDAD wrote:
you see I've managed to make the speech dictation provided by MS to work
Which translates sound into known words.
AbrahamMOKDAD wrote:
or let me say type the combination of phonemes...
Sounds aren't mapped to a phoneme by the SAPI. Phonemes can't be interpreted, and can't be mapped to a series of characters. A single letter has a different sound in a different language, and there is no version that tries to map "generic sounds" to a generic alphabet. An alternative would be to record the sounds yourself (DirectInput) and use AI-algo to classify it (to map the sound to a/a serie of chars) The latter option is quite a challenge, I suppose. I hope that someone else posts an answer that I missed :)
I are Troll :)
Thnx Eddy :thumbsup: ; and if you may; is it possible using SAPI to dictate the english letters; let us say if the user says "parlon" it would be written as said or close to it; not necessarily makes sense. any thoughts on that?
-
Thnx Eddy :thumbsup: ; and if you may; is it possible using SAPI to dictate the english letters; let us say if the user says "parlon" it would be written as said or close to it; not necessarily makes sense. any thoughts on that?
AbrahamMOKDAD wrote:
nd if you may; is it possible using SAPI to dictate the english letters;
Yup, but then you'd be spelling each word, e.g. "Hotel Echo Lima Lima Oscar" for "Hello".
AbrahamMOKDAD wrote:
let us say if the user says "parlon" it would be written as said or close to it; not necessarily makes sense.
That's what happens when you try an English word; it tries to map the sounds to match up to the sounds of a known word. You could start at the .NET classes for speech recognition, in the System.Speech.Recognition[^] namespace. There's also a cool example right here on CodeProject[^].
I are Troll :)
-
AbrahamMOKDAD wrote:
nd if you may; is it possible using SAPI to dictate the english letters;
Yup, but then you'd be spelling each word, e.g. "Hotel Echo Lima Lima Oscar" for "Hello".
AbrahamMOKDAD wrote:
let us say if the user says "parlon" it would be written as said or close to it; not necessarily makes sense.
That's what happens when you try an English word; it tries to map the sounds to match up to the sounds of a known word. You could start at the .NET classes for speech recognition, in the System.Speech.Recognition[^] namespace. There's also a cool example right here on CodeProject[^].
I are Troll :)
Thanx Eddy ^^