Permalink
15be08c Dec 23, 2016
@TalAter @evancohen
206 lines (160 sloc) 9.27 KB

Frequently Asked Questions

What languages are supported?

Language support is up to each browser. While there isn't an official list of supported languages in Chrome, here is a list based on anecdotal evidence.

  • Afrikaans af
  • Basque eu
  • Bulgarian bg
  • Catalan ca
  • Arabic (Egypt) ar-EG
  • Arabic (Jordan) ar-JO
  • Arabic (Kuwait) ar-KW
  • Arabic (Lebanon) ar-LB
  • Arabic (Qatar) ar-QA
  • Arabic (UAE) ar-AE
  • Arabic (Morocco) ar-MA
  • Arabic (Iraq) ar-IQ
  • Arabic (Algeria) ar-DZ
  • Arabic (Bahrain) ar-BH
  • Arabic (Lybia) ar-LY
  • Arabic (Oman) ar-OM
  • Arabic (Saudi Arabia) ar-SA
  • Arabic (Tunisia) ar-TN
  • Arabic (Yemen) ar-YE
  • Czech cs
  • Dutch nl-NL
  • English (Australia) en-AU
  • English (Canada) en-CA
  • English (India) en-IN
  • English (New Zealand) en-NZ
  • English (South Africa) en-ZA
  • English(UK) en-GB
  • English(US) en-US
  • Finnish fi
  • French fr-FR
  • Galician gl
  • German de-DE
  • Hebrew he
  • Hungarian hu
  • Icelandic is
  • Italian it-IT
  • Indonesian id
  • Japanese ja
  • Korean ko
  • Latin la
  • Mandarin Chinese zh-CN
  • Traditional Taiwan zh-TW
  • Simplified China zh-CN ?
  • Simplified Hong Kong zh-HK
  • Yue Chinese (Traditional Hong Kong) zh-yue
  • Malaysian ms-MY
  • Norwegian no-NO
  • Polish pl
  • Pig Latin xx-piglatin
  • Portuguese pt-PT
  • Portuguese (Brasil) pt-BR
  • Romanian ro-RO
  • Russian ru
  • Serbian sr-SP
  • Slovak sk
  • Spanish (Argentina) es-AR
  • Spanish (Bolivia) es-BO
  • Spanish (Chile) es-CL
  • Spanish (Colombia) es-CO
  • Spanish (Costa Rica) es-CR
  • Spanish (Dominican Republic) es-DO
  • Spanish (Ecuador) es-EC
  • Spanish (El Salvador) es-SV
  • Spanish (Guatemala) es-GT
  • Spanish (Honduras) es-HN
  • Spanish (Mexico) es-MX
  • Spanish (Nicaragua) es-NI
  • Spanish (Panama) es-PA
  • Spanish (Paraguay) es-PY
  • Spanish (Peru) es-PE
  • Spanish (Puerto Rico) es-PR
  • Spanish (Spain) es-ES
  • Spanish (US) es-US
  • Spanish (Uruguay) es-UY
  • Spanish (Venezuela) es-VE
  • Swedish sv-SE
  • Turkish tr
  • Zulu zu

Why does the browser repeatedly ask for permission to use the microphone?

Chrome's speech recognition behaves differently based on the protocol used:

  • https:// Asks for permission once and remembers the choice.

  • http:// Asks for permission repeatedly on every page load. Results are also returned significantly slower in HTTP.

For a great user experience, don't compromise on anything less than HTTPS (available free with CloudFlare and Let's Encrypt).

What can I do to make speech recognition results return faster?

First, remember that because the actual speech-to-text processing is done in the cloud, a faster connection can mean faster results.

Second, when the speech recognition is in continuous mode, results are returned slower (the browser waits after you finish talking to see if there's anything else you'd like to add).

Turning continuous mode off tends to make the browser return recognized results much faster.

To start annyang in non-continuous mode, you can pass continuous: false in the options object that annyang.start() accepts. You will most likely want to also turn on autoRestart if you do that. You can read more about both options in the annyang API Docs

For example:

annyang.start({ autoRestart: true, continuous: false });

Note that these settings are already the default if you are using HTTPS. If you are using HTTP, continuous mode will be turned on by default (resulting in slower recognition) to prevent repeated security notices.

How can I contribute to annyang's development?

There are three main ways for you to help. Check out the CONTRIBUTING guide for more details.

Why does Speech Recognition repeatedly starts and stops?

The most common reason for this is because you have opened more than one tab or window that uses Speech Recognition in your browser at the same time (e.g. if you open annyang's homepage in one tab, and the Speech Recognition app you are developing in another).

When a browser detects that one tab has started Speech Recognition, it aborts all other Speech Recognition processes in other tabs. annyang detects when it is aborted by an external process and restarts itself. If you have two windows aborting each other, and restarting themselves you may experience Speech Recognition starting and stopping over and over again.

Another possible reason for this might be that you are offline.

Can annyang work offline?

No. annyang relies on the browser's own speech recognition engine. In Chrome, this engine performs the recognition in the cloud.

Can annyang be used to capture the full text spoken by the user?

Yes. You can listen to the result event which is triggered whenever speech is recognized. This event will fire with a list of possible phrases the user may have said, regardless of whether any of them matched an annyang command or not. You can even do this without registering any commands:

annyang.addCallback('result', function(phrases) {
  console.log("I think the user said: ", phrases[0]);
  console.log("But then again, it could be any of the following: ", phrases);
});

Alternatively, you may choose to only capture what the user said when it matches an annyang command (resultMatch), or when it does not match a command (resultNoMatch).

annyang.addCallback('resultMatch', function(userSaid, commandText, phrases) {
  console.log(userSaid); // sample output: 'hello'
  console.log(commandText); // sample output: 'hello (there)'
  console.log(phrases); // sample output: ['hello', 'halo', 'yellow', 'polo', 'hello kitty']
});

annyang.addCallback('resultNoMatch', function(phrases) {
  console.log("I think the user said: ", phrases[0]);
  console.log("But then again, it could be any of the following: ", phrases);
});

Can I detect when the user starts and stops speaking?

Yes. Sometimes.

You can detect when a sound is first detected by the microphone with the soundstart event. Unfortunately, due to a bug in Chrome, this event will only fire once in every speech recognition session. If you are in non-continuous mode and annyang is restarting after every sentence recognized (the default in HTTPS), this will not be a problem. Because speech recognition will abort and restart, soundstart will fire again correctly.

The following code will detect when a user starts and stops speaking.

annyang.addCallback('soundstart', function() {
  console.log('sound detected');
});

annyang.addCallback('result', function() {
  console.log('sound stopped');
});

Note: The soundstart event is only available in annyang v2.6.0 and up.

Can annyang be used in Chromium or Electron?

Yes, however you must create your own Chromium keys and are limited to 50 requests/day. To do this you'll need to provide your own keys at runtime by following the instructions for Acquiring Keys in the Chromium developer docs.

Can annyang be used in Cordova?

Yes. In order to use webKitSpeechRecognition you will need to use Crosswalk and the Cordova Media Plugin. These can be added to an existing cordova project with the following commands:

cordova plugin add cordova-plugin-crosswalk-webview
cordova plugin add cordova-plugin-media

Known issues:

  • This has only been verified to work on Android.
  • This will not work on a device without a SpeechRecognizer (like the Kindle Fire) unless you sideload the Google Now Launcher.
  • Your app will ding with the native speech recognition sound, you have no control over this. See #194.