Open siva3378 opened 7 years ago
Confirmed using the following code:
// add command
artyom.on(["hello"]).then((e) => {
console.log("This shouldn't be executed, however it seems it's executed.")
});
// Start recognition
artyom.initialize({
lang:"en-GB",// A lot of languages are supported. Read the docs !
continuous:true,// recognize 1 command and stop listening !
listen:true, // Start recognizing
debug:true, // Show everything in the console
speed:1 // talk normally
});
// Talk
window.artyom.say("Hello, execute hello command, hello. hello. hello", {
onStart() {
window.artyom.dontObey();
},
onEnd() {
window.artyom.obey();
}
});
I'll see what i can do.
+1
Ehi.. i'm solved in this way:
var initObj = {
speed: 1,
lang: "it-IT",
continuous: false,
debug: true,
listen: true // Start to listen commands !
};
artyom.say(msg, {
onStart: function () {
},
onEnd: function () {
artyom.fatality(); // Stop artyom
setTimeout(function(){
artyom.initialize(initObj).then(function () { // re-initialize artyom
......... // do something
});
},250);
}
});
Any update to stop listening while artyom is speaking ?
^ any update? If not this library is ... :/
Well, after one year I decided to take a look to this problem :v. As someway I don't consider this a bug, as the most suitable environment to use Artyom will be with earphones with Microphone, I decided to not modify the library for this nor just add a new property in the initialization.
Instead the solution, using the latest version of artyom 1.0.6 because that's the only version that exposes the webkitspeechrecognition object internally used by artyom is invoking the abort
method of the ArtyomWebkitSpeechRecognition
object of your instance of Artyom in the onEnd
callback of the method say:
let Assistant = new Artyom();
Assistant.say("Hello, how are you?", {
onEnd (){
// Abort the speech recognition when artyom stops talking !
// This prevents any "hello" command from being executed indefinitely
Assistant.ArtyomWebkitSpeechRecognition.abort();
}
});
This solve the mentioned behaviour, when you add a command, for example "hello" and artyom triggers the command that speaks "hello". Artyom will recognize the text spoken by himself and will trigger the command "hello" over and over again. To solve this problem, use the previous snippet e.g:
let Assistant = new Artyom();
Assistant.addCommands({
indexes: ["Hello", "How are you"],
action: () => {
// Then if the command is executed and you say hello again
// It will enter into an endless loop because artyom is hearing itself
// again and again ...
// To prevent this, abort the current queue of the speech recognition instance
// used by Artyom with the method abort on the onEnd callback of artyom.say:
Assistant.say("Hello", {
onEnd (){
// Abort the speech recognition when artyom stops talking !
// Then, the command won't be triggered when artyom says hello !
Assistant.ArtyomWebkitSpeechRecognition.abort();
}
});
}
});
// When artyom says this:
Assistant.say("Hello, how are you?");
// The Hello command will be triggered
It's worth to say, that this feature relies as well on the execution of the onEnd callback. Once the speech synthesis finishes, it's ready to recognize commands. After some tests it works properly, however if someone tries as well and fails, please let me know it.
Okey. Thank you. Is not the best solution (because we need to add Assistant.ArtyomWebkitSpeechRecognition.abort(); in almost all "Assistant.say" but works.
Thank you 👍
Thanks @sdkcarlos , will give this a try later
We can use global variables
artyom.when("SPEECH_SYNTHESIS_END",function(){ artyom.ArtyomWebkitSpeechRecognition.abort(); });
I'm interested in the following
There are no problems with the PC. But when I use on the phone I have an intersection when performing functions
I use global variables to disable and enable the microphone, which would avoid the intersection, approximately so
artyom.when("SPEECH_SYNTHESIS_START",function(err){ if(recognizing){ artyom.fatality(); } });
artyom.when("SPEECH_SYNTHESIS_END",function(err){ artyom.ArtyomWebkitSpeechRecognition.abort(); if(recognizing){ setTimeout(function(){ artyom.initialize(initObj).then(function () { // re-initialize artyom console.log("Started"); // do something }); },0); artyom.clearGarbageCollection(); } });
None of the above mentioned solutions worked for me. However, using a callback on an event was the way to go.
artyom.when("SPEECH_SYNTHESIS_START", function(){
if (artyom.isRecognizing()){
artyom.dontObey();
}
});
artyom.when("SPEECH_SYNTHESIS_END", function(){
if(!artyom.isRecognizing() || !artyom.isSpeaking()){
artyom.obey();
}
});
That works like a charm! Artyom is a good sdk. Good work @sdkcarlos!
well, to solve the problem, what I did was modify the function, execute by adding the following conditional:
`//if artyom is synthesizing voice it will only be able to accept command specified in the initialized
if (_this.ArtyomProperties.speaking){ if (voz.indexOf(_this.ArtyomProperties.interruptkeyword) != 0){ _this.debug("Artyom requires with a interruptkeyword \"" + _this.ArtyomProperties.interruptkeyword + "\" but interrupt was not pronounced.", "warn"); return; } }`
of course add the following property to the ArtyomProperties:
interruptkeyword: 'interrupt'
and it is conditional on the function of initialize:
if (config.hasOwnProperty ("interruptkeyword")) { this.ArtyomProperties.interruptkeyword = config.interruptkeyword; }
Using the solution of @ZvirUSka seems to do the trick, without the need of defining a callback on each artyom.say
:
artyom.when("SPEECH_SYNTHESIS_END", () => {
artyom.ArtyomWebkitSpeechRecognition.abort();
});
This also works on phones for me (tested with Chrome 86 on Android)
While artyom speaks, it is also recognizing the text spoken by itself. I tried to stop/start obeying while it speaks, but it is not working as expected. Is there any way to stop listening while artyom speaks?