Following HTML shows empty array in console on first click:
<!DOCTYPE html>
<html>
<head>
<script>
function test(){
console.log(window.speechSynthesis.getVoices())
}
</script>
</head>
<body>
Test
</body>
</html>
In second click you will get the expected list.
If you add onload event to call this function (<body onload="test()">), then you can get correct result on first click. Note that the first call on onload still doesn't work properly. It returns empty on page load but works afterward.
Questions:
Since it might be a bug in beta version, I gave up on "Why" questions.
Now, the question is if you want to access window.speechSynthesis on page load:
What is the best hack for this issue?
How can you make sure it will load speechSynthesis, on page load?
Background and tests:
I was testing the new features in Web Speech API, then I got to this problem in my code:
<script type="text/javascript">
$(document).ready(function(){
// Browser support messages. (You might need Chrome 33.0 Beta)
if (!('speechSynthesis' in window)) {
alert("You don't have speechSynthesis");
}
var voices = window.speechSynthesis.getVoices();
console.log(voices) // []
$("#test").on('click', function(){
var voices = window.speechSynthesis.getVoices();
console.log(voices); // [SpeechSynthesisVoice, ...]
});
});
</script>
<a id="test" href="#">click here if 'ready()' didn't work</a>
My question was: why does window.speechSynthesis.getVoices() return empty array, after page is loaded and onready function is triggered? As you can see if you click on the link, same function returns an array of available voices of Chrome by onclick triger?
It seems Chrome loads window.speechSynthesis after the page load!
The problem is not in ready event. If I remove the line var voice=... from ready function, for first click it shows empty list in console. But the second click works fine.
It seems window.speechSynthesis needs more time to load after first call. You need to call it twice! But also, you need to wait and let it load before second call on window.speechSynthesis. For example, following code shows two empty arrays in console if you run it for first time:
// First speechSynthesis call
var voices = window.speechSynthesis.getVoices();
console.log(voices);
// Second speechSynthesis call
voices = window.speechSynthesis.getVoices();
console.log(voices);
According to Web Speech API Errata (E11 2013-10-17), the voice list is loaded async to the page. An onvoiceschanged event is fired when they are loaded.
voiceschanged: Fired when the contents of the SpeechSynthesisVoiceList, that the getVoices method will return, have changed. Examples include: server-side synthesis where the list is determined asynchronously, or when client-side voices are installed/uninstalled.
So, the trick is to set your voice from the callback for that event listener:
// wait on voices to be loaded before fetching list
window.speechSynthesis.onvoiceschanged = function() {
window.speechSynthesis.getVoices();
...
};
You can use a setInterval to wait until the voices are loaded before using them however you need and then clearing the setInterval:
var timer = setInterval(function() {
var voices = speechSynthesis.getVoices();
console.log(voices);
if (voices.length !== 0) {
var msg = new SpeechSynthesisUtterance(/*some string here*/);
msg.voice = voices[/*some number here to choose from array*/];
speechSynthesis.speak(msg);
clearInterval(timer);
}
}, 200);
$("#test").on('click', timer);
After studying the behavior on Google Chrome and Firefox, this is what can get all voices:
Since it involves something asynchronous, it might be best done with a promise:
const allVoicesObtained = new Promise(function(resolve, reject) {
let voices = window.speechSynthesis.getVoices();
if (voices.length !== 0) {
resolve(voices);
} else {
window.speechSynthesis.addEventListener("voiceschanged", function() {
voices = window.speechSynthesis.getVoices();
resolve(voices);
});
}
});
allVoicesObtained.then(voices => console.log("All voices:", voices));
Note:
When the event voiceschanged fires, we need to call .getVoices() again. The original array won't be populated with content.
On Google Chrome, we don't have to call getVoices() initially. We only need to listen on the event, and it will then happen. On Firefox, listening is not enough, you have to call getVoices() and then listen on the event voiceschanged, and set the array using getVoices() once you get notified.
Using a promise makes the code more clean. Everything related to getting voices are in this promise code. If you don't use a promise but instead put this code in your speech routine, it is quite messy.
You can write a voiceObtained promise to resolve to a voice you want, and then your function to say something can just do: voiceObtained.then(voice => { }) and inside that handler, call the window.speechSynthesis.speak() to speak something. Or you can even write a promise speechReady("hello world").then(speech => { window.speechSynthesis.speak(speech) }) to say something.
heres the answer
function synthVoice(text) {
const awaitVoices = new Promise(resolve=>
window.speechSynthesis.onvoiceschanged = resolve)
.then(()=> {
const synth = window.speechSynthesis;
var voices = synth.getVoices();
console.log(voices)
const utterance = new SpeechSynthesisUtterance();
utterance.voice = voices[3];
utterance.text = text;
synth.speak(utterance);
});
}
At first i used onvoiceschanged , but it kept firing even after the voices was loaded, so my goal was to avoid onvoiceschanged at all cost.
This is what i came up with. It seems to work so far, will update if it breaks.
loadVoicesWhenAvailable();
function loadVoicesWhenAvailable() {
voices = synth.getVoices();
if (voices.length !== 0) {
console.log("start loading voices");
LoadVoices();
}
else {
setTimeout(function () { loadVoicesWhenAvailable(); }, 10)
}
}
setInterval solution by Salman Oskooi was perfect
Please see https://jsfiddle.net/exrx8e1y/
function myFunction() {
dtlarea=document.getElementById("details");
//dtlarea.style.display="none";
dtltxt="";
var mytimer = setInterval(function() {
var voices = speechSynthesis.getVoices();
//console.log(voices);
if (voices.length !== 0) {
var msg = new SpeechSynthesisUtterance();
msg.rate = document.getElementById("rate").value; // 0.1 to 10
msg.pitch = document.getElementById("pitch").value; //0 to 2
msg.volume = document.getElementById("volume").value; // 0 to 1
msg.text = document.getElementById("sampletext").value;
msg.lang = document.getElementById("lang").value; //'hi-IN';
for(var i=0;i<voices.length;i++){
dtltxt+=voices[i].lang+' '+voices[i].name+'\n';
if(voices[i].lang==msg.lang) {
msg.voice = voices[i]; // Note: some voices don't support altering params
msg.voiceURI = voices[i].voiceURI;
// break;
}
}
msg.onend = function(e) {
console.log('Finished in ' + event.elapsedTime + ' seconds.');
dtlarea.value=dtltxt;
};
speechSynthesis.speak(msg);
clearInterval(mytimer);
}
}, 1000);
}
This works fine on Chrome for MAC, Linux(Ubuntu), Windows and Android
Android has non-standard en_GB wile others have en-GB as language code
Also you will see that same language(lang) has multiple names
On Mac Chrome you get en-GB Daniel besides en-GB Google UK English Female and n-GB Google UK English Male
en-GB Daniel (Mac and iOS)
en-GB Google UK English Female
en-GB Google UK English Male
en_GB English United Kingdom
hi-IN Google हिन्दी
hi-IN Lekha (Mac and iOS)
hi_IN Hindi India
Another way to ensure voices are loaded before you need them is to bind their loading state to a promise, and then dispatch your speech commands from a then:
const awaitVoices = new Promise(done => speechSynthesis.onvoiceschanged = done);
function listVoices() {
awaitVoices.then(()=> {
let voices = speechSynthesis.getVoices();
console.log(voices);
});
}
When you call listVoices, it will either wait for the voices to load first, or dispatch your operation on the next tick.
I used this code to load voices successfully:
<select id="voices"></select>
...
function loadVoices() {
populateVoiceList();
if (speechSynthesis.onvoiceschanged !== undefined) {
speechSynthesis.onvoiceschanged = populateVoiceList;
}
}
function populateVoiceList() {
var allVoices = speechSynthesis.getVoices();
allVoices.forEach(function(voice, index) {
var option = $('<option>').val(index).html(voice.name).prop("selected", voice.default);
$('#voices').append(option);
});
if (allVoices.length > 0 && speechSynthesis.onvoiceschanged !== undefined) {
// unregister event listener (it is fired multiple times)
speechSynthesis.onvoiceschanged = null;
}
}
I found the 'onvoiceschanged' code from this article: https://hacks.mozilla.org/2016/01/firefox-and-the-web-speech-api/
Note: requires JQuery.
Works in Firefox/Safari and Chrome (and in Google Apps Script too - but only in the HTML).
async function speak(txt) {
await initVoices();
const u = new SpeechSynthesisUtterance(txt);
u.voice = speechSynthesis.getVoices()[3];
speechSynthesis.speak(u);
}
function initVoices() {
return new Promise(function (res, rej){
speechSynthesis.getVoices();
if (window.speechSynthesis.onvoiceschanged) {
res();
} else {
window.speechSynthesis.onvoiceschanged = () => res();
}
});
}
While the accepted answer works great but if you're using SPA and not loading full-page, on navigating between links, the voices will not be available.
This will run on a full-page load
window.speechSynthesis.onvoiceschanged
For SPA, it wouldn't run.
You can check if it's undefined, run it, or else, get it from the window object.
An example that works:
let voices = [];
if(window.speechSynthesis.onvoiceschanged == undefined){
window.speechSynthesis.onvoiceschanged = () => {
voices = window.speechSynthesis.getVoices();
}
}else{
voices = window.speechSynthesis.getVoices();
}
// console.log("voices", voices);
I had to do my own research for this to make sure I understood it properly, so just sharing (feel free to edit).
My goal is to:
Get a list of voices available on my device
Populate a select element with those voices (after a particular page loads)
Use easy to understand code
The basic functionality is demonstrated in MDN's official live demo of:
https://github.com/mdn/web-speech-api/tree/master/speak-easy-synthesis
but I wanted to understand it better.
To break the topic down...
SpeechSynthesis
The SpeechSynthesis interface of the Web Speech API is the controller
interface for the speech service; this can be used to retrieve
information about the synthesis voices available on the device, start
and pause speech, and other commands besides.
Source
onvoiceschanged
The onvoiceschanged property of the SpeechSynthesis interface
represents an event handler that will run when the list of
SpeechSynthesisVoice objects that would be returned by the
SpeechSynthesis.getVoices() method has changed (when the voiceschanged
event fires.)
Source
Example A
If my application merely has:
var synth = window.speechSynthesis;
console.log(synth);
console.log(synth.onvoiceschanged);
Chrome developer tools console will show:
Example B
If I change the code to:
var synth = window.speechSynthesis;
console.log("BEFORE");
console.log(synth);
console.log(synth.onvoiceschanged);
console.log("AFTER");
var voices = synth.getVoices();
console.log(voices);
console.log(synth);
console.log(synth.onvoiceschanged);
The before and after states are the same, and voices is an empty array.
Solution
Although i'm not confident implementing Promises, the following worked for me:
Defining the function
var synth = window.speechSynthesis;
// declare so that values are accessible globally
var voices = [];
function set_up_speech() {
return new Promise(function(resolve, reject) {
// get the voices
var voices = synth.getVoices();
// get reference to select element
var $select_topic_speaking_voice = $("#select_topic_speaking_voice");
// for each voice, generate select option html and append to select
for (var i = 0; i < voices.length; i++) {
var option = $("<option></option>");
var suffix = "";
// if it is the default voice, add suffix text
if (voices[i].default) {
suffix = " -- DEFAULT";
}
// create the option text
var option_text = voices[i].name + " (" + voices[i].lang + suffix + ")";
// add the option text
option.text(option_text);
// add option attributes
option.attr("data-lang", voices[i].lang);
option.attr("data-name", voices[i].name);
// append option to select element
$select_topic_speaking_voice.append(option);
}
// resolve the voices value
resolve(voices)
});
}
Calling the function
// in your handler, populate the select element
if (page_title === "something") {
set_up_speech()
}
Android Chrome - turn off data saver. It was helpfull for me.(Chrome 71.0.3578.99)
// wait until the voices load
window.speechSynthesis.onvoiceschanged = function() {
window.speechSynthesis.getVoices();
};
let voices = speechSynthesis.getVoices();
let gotVoices = false;
if (voices.length) {
resolve(voices, message);
} else {
speechSynthesis.onvoiceschanged = () => {
if (!gotVoices) {
voices = speechSynthesis.getVoices();
gotVoices = true;
if (voices.length) resolve(voices, message);
}
};
}
function resolve(voices, message) {
var synth = window.speechSynthesis;
let utter = new SpeechSynthesisUtterance();
utter.lang = 'en-US';
utter.voice = voices[65];
utter.text = message;
utter.volume = 100.0;
synth.speak(utter);
}
Works for Edge, Chrome and Safari - doesn't repeat the sentences.
I brought this up in my last post but since it was off topic from the original question I'm posting it separately. I'm having trouble with getting my transmitted audio to play back through Web Audio the same way it would sound in a media player. I have tried 2 different transmission protocols, binaryjs and socketio, and neither make a difference when trying to play through Web Audio. To rule out the transportation of the audio data being the issue I created an example that sends the data back to the server after it's received from the client and dumps the return to stdout. Piping that into VLC results in a listening experience that you would expect to hear.
To hear the results when playing through vlc, which sounds the way it should, run the example at https://github.com/grkblood13/web-audio-stream/tree/master/vlc using the following command:
$ node webaudio_vlc_svr.js | vlc -
For whatever reason though when I try to play this same audio data through Web Audio it fails miserably. The results are random noises with large gaps of silence in between.
What is wrong with the following code that is making the playback sound so bad?
window.AudioContext = window.AudioContext || window.webkitAudioContext;
var context = new AudioContext();
var delayTime = 0;
var init = 0;
var audioStack = [];
client.on('stream', function(stream, meta){
stream.on('data', function(data) {
context.decodeAudioData(data, function(buffer) {
audioStack.push(buffer);
if (audioStack.length > 10 && init == 0) { init++; playBuffer(); }
}, function(err) {
console.log("err(decodeAudioData): "+err);
});
});
});
function playBuffer() {
var buffer = audioStack.shift();
setTimeout( function() {
var source = context.createBufferSource();
source.buffer = buffer;
source.connect(context.destination);
source.start(context.currentTime);
delayTime=source.buffer.duration*1000; // Make the next buffer wait the length of the last buffer before being played
playBuffer();
}, delayTime);
}
Full source: https://github.com/grkblood13/web-audio-stream/tree/master/binaryjs
You really can't just call source.start(audioContext.currentTime) like that.
setTimeout() has a long and imprecise latency - other main-thread stuff can be going on, so your setTimeout() calls can be delayed by milliseconds, even tens of milliseconds (by garbage collection, JS execution, layout...) Your code is trying to immediately play audio - which needs to be started within about 0.02ms accuracy to not glitch - on a timer that has tens of milliseconds of imprecision.
The whole point of the web audio system is that the audio scheduler works in a separate high-priority thread, and you can pre-schedule audio (starts, stops, and audioparam changes) at very high accuracy. You should rewrite your system to:
1) track when the first block was scheduled in audiocontext time - and DON'T schedule the first block immediately, give some latency so your network can hopefully keep up.
2) schedule each successive block received in the future based on its "next block" timing.
e.g. (note I haven't tested this code, this is off the top of my head):
window.AudioContext = window.AudioContext || window.webkitAudioContext;
var context = new AudioContext();
var delayTime = 0;
var init = 0;
var audioStack = [];
var nextTime = 0;
client.on('stream', function(stream, meta){
stream.on('data', function(data) {
context.decodeAudioData(data, function(buffer) {
audioStack.push(buffer);
if ((init!=0) || (audioStack.length > 10)) { // make sure we put at least 10 chunks in the buffer before starting
init++;
scheduleBuffers();
}
}, function(err) {
console.log("err(decodeAudioData): "+err);
});
});
});
function scheduleBuffers() {
while ( audioStack.length) {
var buffer = audioStack.shift();
var source = context.createBufferSource();
source.buffer = buffer;
source.connect(context.destination);
if (nextTime == 0)
nextTime = context.currentTime + 0.05; /// add 50ms latency to work well across systems - tune this if you like
source.start(nextTime);
nextTime+=source.buffer.duration; // Make the next buffer wait the length of the last buffer before being played
};
}
i started using this browser(chrome) feature.
i ve written a JS based on this , but the problem is even , it recognises the speech only once and ends . its not going continuously, i need to press the button again and again to start speech recognition . tell me where i should tweak . i ve set "recognition.continuous=true" still not helping ?
var recognition = new webkitSpeechRecognition();
recognition.continuous = true;
recognition.interimResults = true;
recognition.onstart = function() {
console.log("Recognition started");
};
recognition.onresult = function(event){
console.log(event.results);
};
recognition.onerror = function(e) {
console.log("Error");
};
recognition.onend = function() {
console.log("Speech recognition ended");
};
function start_speech() {
recognition.lang = 'en-IN'; // 'en-US' works too, as do many others
recognition.start();
}
I call "start_speech" from a button ! thats it
I know this is an old thread, but I had this problem too. I found that, even with the continuous flag set, if there are pauses in the input speech, a "no-speech" error gets thrown (triggers the onerror event) and the engine is shut down. I just added code in the onend to restart the engine:
recognition.onend = function() {
recognition.start();
};
The next problem you might get is, every time the engine restarts, the user has to re-grant permission to let the browser use the microphone. The only solution at this time seems to be to make sure you connect to your site over HTTPS (source: http://updates.html5rocks.com/2013/01/Voice-Driven-Web-Apps-Introduction-to-the-Web-Speech-API bottom of post in bold)
Perhaps the typo on this line:
recognition.continuos = true;
Should equal:
recognition.continuous = true;
recognition.onend = function(){
recognition.start();
// sets off a beep/noise each time it is accessed from a cell phone (Andoid).
// does NOT if accessed from a desktop (Windows using Chrome).
};
I am trying to get the following code to work on chrom by using setVersion (as onupgradeneeded is not available yet).
The IDBVersionChangeRequest is filled with IDBDatabaseException. And the onsuccess function could not be called. I need to create an ObjectStore within the onsuccess function.
specifically this line: request = browserDatabase._db.setVersion(browserDatabase._dbVersion.toString());
Below is my code. Any help would be greatly appreciated...
browserDatabase._db = null;
browserDatabase._dbVersion = 4;
browserDatabase._dbName = "mediaStorageDB";
browserDatabase._storeName = "myStore";
var request = indexedDB.open(browserDatabase._dbName);
// database exist
request.onsuccess = function(e)
{
browserDatabase._db = e.target.result;
// this is specifically for chrome, because it does not support onupgradeneeded
if (browserDatabase._dbVersion != browserDatabase._db.version)
{
request = browserDatabase._db.setVersion(browserDatabase._dbVersion.toString());
request.onerror = function(e) { alert("error") };
request.onblocked = function(e)
{
b = 11; // for some reason the code goes here...
}
request.onsuccess = function(e)
{
browserDatabase._db.createObjectStore(browserDatabase._storeName, {autoIncrement: true});
}
}
}
In your code sample you say you come in to the onblocked callback. The only way you can get in this callback is when you have still open transactions/connections to your db. (aside the one you are working in.) This means you will have to close all other transactions/connections before you can call the setVersion.
When wired things happen to IndexedDB, I "Clear data from hosted apps", quit Chrome windows and take a cup of coffee. After that everything work fine. :-D
browserDatabase._dbVersion < browserDatabase._db.version. Downgrading is not possible. dbVersion = 4 should not be consider lightly. You might open other tab with dbVersion = 5, or browser may be waining your response elsewhere or itself updating. All these are not worth to trace the reasons behind.
As it turns out the answer to my prior question does not work. The problem is that I am overloading the server with requests. The process needs a throttle so that subsequent requests incur a little slowdown. Here is the code in question. The important segment of code that needs a timer or setInterval is that prefaced by the alert "Profile Rejected" although it would be acceptable to slow them both down. Any suggestions?
if (greetThisOne==true && !bGreeted)
{
//alert ("Requesting Message Page");
console.log="Message Page Requested";
chrome.extension.sendRequest({cmd: "openMessage", url: messageLink, keyWordsFound: keyWordList, greeted: bGreeted});
}
else
{
//alert("Profile Rejected");
console.log="Profile Rejected";
chrome.extension.sendRequest({cmd: "profileRejected", url: messageLink, keyWordsFound: keyWordList, greeted: bGreeted});
}
You would need to implement some queue in the background page. For example:
var rejectedProfiles = [];
processRejectedProfiles();
chrome.extension.onRequest.addListener(function(request, sender, sendResponse) {
if(request.cmd == "profileRejected") {
//add to the end of queue
rejectedProfiles.push({url: request.url, other: request.parameters});
}
sendResponse({});
});
function processRejectedProfiles() {
if(rejectedProfiles.length > 0) {
//get the oldest element in queue
var profile = rejectedProfiles.shift();
//process profile
...
}
//process next entry in the queue in 3 seconds
setTimeout(processRejectedProfiles, 3000);
}
This way you will be processing one profile at a time, with a provided delay.