Change speaking immediately #7
Comments
|
I'm not actively supporting this anymore, but it does seem like you have the proper approach. In my demos, stopping an existing voice and playing a new one is not an issue, but that's based on a person clicking a button and pressing another one. There's a natural delay between the two voices. So I wonder if the browser needs some sort of pause in there, so it can kind of reset itself, Maybe set a window.timeout of 500ms before starting the new voice? |
|
I'm calling out seconds remaining in a timer, so it's critical that the
voice isn't noticeably delayed. I've tried writing a function so that
during the final seconds, the voice is cut short (by setting stop on a
timeout). While this works a little better it still isn't getting the job
done completely—for very short speech, stop() seems to not cut off
immediately, or it prevents speech from starting at all.
I haven't tried altering the speech rate to see if that makes a difference,
but knowing operating system voices are different I'd hate to rely on that
kind of tweaking.
…On Thu, Dec 6, 2018 at 10:42 AM Adam Coti ***@***.***> wrote:
I'm not actively supporting this anymore, but it does seem like you have
the proper approach. In my demos, stopping an existing voice and playing a
new one is not an issue, but that's based on a person clicking a button and
pressing another one. There's a natural delay between the two voices.
So I wonder if the browser needs some sort of pause in there, so it can
kind of reset itself, Maybe set a window.timeout of 500ms before starting
the new voice?
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
<#7 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AG1l9-XypKvnBkbT-Ip67bD99ZoNFgMAks5u2TrhgaJpZM4ZGpQE>
.
|
|
At this point, you may be bumping up against browser limitations on how they implement speech synthesis. If I think of some other approach you can try, I'll let you know. Please keep me posted on how your testing proceeds. |
|
Thanks for your help. I think it is browser limitations as well, because different browsers on different devices gave slightly different results. I wasn't able to resolve this for very short duration speech as in my use case. As a workaround, I switched to playing an audio file to mark the final seconds. The audio file will play in parallel with speech synthesis if needed. |
|
I ran into the same issue, it definitely seems to be a problem with the speechSynthesis api. A minimal self contained example is the following: <html>
<body>
<div><button onclick="playFirst()">Play First</button></div>
<div><button onclick="playSecond()">Play Second</button></div>
<script type="text/javascript">
const myText = "the first text";
const myText2 = "the second text!";
function playFirst() {
speechSynthesis.speak(new SpeechSynthesisUtterance(myText));
}
function playSecond() {
speechSynthesis.cancel();
speechSynthesis.speak(new SpeechSynthesisUtterance(myText2));
}
</script>
</body>
</html>If you click the first button, then the second whilst the text from the first is still being spoken, the speech stops, and the second string is not spoken. |
|
Hmm, it seems like the above snippet works in firefox, but not chrome. |
|
One thing I want to point out: there is a Chrome fix at lines 66-69 in the annotated "articulate.js" file. This was necessary at the time when the speech synthesis API was relatively new, but is also two years old. Perhaps this "fix" is no longer needed or is, somehow, affecting what you are trying to do. |
The voice won't stop speaking the current section if something else is told to start. I need to have the option to switch right away and force something else to speak. I tried writing a function this way:
$().articulate('stop');
$(obj).articulate('speak');
...but it only starts speaking if the browser wasn't already speaking. Is there another way to accomplish this?
The text was updated successfully, but these errors were encountered: