Word Factory performs a support check before loading Listen & Choose. It looks for the browser’s speechSynthesis interface, the SpeechSynthesisUtterance constructor and at least one voice whose language begins with en. If the voice list is initially empty, the page waits up to four seconds for the browser’s voiceschanged event.

This gate matters because showing a listening question without playable speech would turn the round into guessing. If the check fails, the iframe is not loaded and the launch status explains that an English voice is required.

Why voice availability can differ

The Web Speech API exposes voices made available by the browser and operating system. The Web Speech API specification defines the synthesis interface, while MDN’s getVoices documentation explains that the returned list represents voices available on the current device. As a result, two current browsers can pronounce the same word with different accents or voice quality.

Some browsers populate the voice list after the page has loaded. The voiceschanged event exists for this delayed availability. Word Factory listens for that event before deciding that speech is unavailable.

Troubleshooting checklist

  1. Confirm that the device is not muted and that the browser tab is allowed to play sound.
  2. Press the game’s Play button. Speech may require a user action before audio begins.
  3. If the page reports no English voice, check the operating system’s language or accessibility settings and install an English text-to-speech voice where that option is available.
  4. Close and reopen the browser after adding a voice so its voice list is rebuilt.
  5. Try a current version of another mainstream browser on the same device. This separates a browser configuration problem from a device voice problem.
  6. Use another Word Factory game if no English voice is exposed. Word Search, Missing Letters and the other visual modes do not require speech.

When the game opens but no sound plays

A passed support gate proves that an English voice is listed; it does not prove that the tab, output device or system volume is audible. Check whether sound is routed to headphones, a monitor or a disconnected Bluetooth device. Press the in-game speaker once after the question appears, because some environments allow speech only after a user gesture. Then test ordinary media in the same browser profile. If other audio works but speech remains silent, restart the browser so its synthesis service and voice list are rebuilt.

When the wrong language or accent is used

The gate accepts any exposed voice whose language metadata starts with English. A device may choose a different English region than the learner expects. Review the operating system’s preferred speech language and default voice. Word Factory does not download a private voice or promise a particular accent, so managed devices should be configured consistently before a shared listening activity.

Worked example: the button stays on the launch page

A learner presses “Play Listen & Choose,” but the status changes to “An English speech voice is required.” The correct response is not to refresh repeatedly. First, confirm that the operating system has an English speech voice. After installing or enabling one, restart the browser and press Play again. If the same message appears, try another current browser. When the gate passes, the game loads and the speaker control can replay the current word.

Best classroom practice: test the exact browser profile and headphones before a session. A successful check on the teacher’s laptop does not prove that a managed student device exposes the same voices.

What the game sends

The game asks the browser to speak a word through its synthesis API; it does not request microphone access and it does not perform speech recognition. Voice implementations may be local or may depend on services supplied by the browser or operating system. Word Factory cannot determine from the API alone how a particular installed voice is produced, so users with strict network or privacy requirements should consult their device and browser documentation.

The support check reads only the available voice metadata needed to find an English language entry. The game does not ask the learner to record or upload speech.

If speech begins and then stops after the tab is hidden, return to the tab and use the replay control. Browsers may pause or cancel queued speech when a page loses focus, the device sleeps or another application takes over audio. This is a playback-state issue rather than evidence that the answer or level data was lost.

Known limitations and alternatives

Pronunciation, accent, pacing and volume can vary. The game therefore tests recognition of the spoken form provided by the current device, not a single standardized recording. It is useful practice, but it should not be treated as a pronunciation assessment.

If consistent audio is essential, use the same managed device image and browser for every learner. Otherwise treat voice variation as additional listening exposure. For silent environments or unsupported devices, choose a visual game from the game-selection guide.

How this page was prepared. The Word Factory editorial team reviewed the current local browser build and the linked game rules on September 26, 2026. Interface behavior described here was checked against the current HTML, JavaScript and Phaser implementation. Corrections can be sent through the contact page; see the editorial policy for review and correction standards.