Note

• When using the voice recognition feature on Google Translate (or any other, for that matter), the user speaks into the microphone, and after he stops talking, he will have to wait for a few seconds for the Web Speech API to return the spoken words into text; whereas in Chrome browser, the words will instantly appear into the input box while the user speaks.

Affected versions

• Nightly v72.0a1

Affected platforms

• all and Fenix

Steps to reproduce

1. Open Nightly v72.0a1 from 2019-11-15
2. Open about:config
3. Flip to following prefs si "true":
   media.webspeech.recognition.enable
   media.webspeech.recognition.force_enable
4. Go to https://translate.google.ro/
5. Click the microphone icon/button from the right of the search input box and speak into the microphone.

Expected result

• The official expected result is unknown to me, but I suggest it would better work like Chrome does: words are being written into the input box while the user speaks OR just made faster so the user does not get irritated with its performance.

Actual result

• The user speaks into the microphone, and after he stops talking, he will have to wait for a few seconds for the Web Speech API to return the spoken words into text.

Additional notes

• This issue is present al all platforms, including Fenix v72.0a1 191115 on Android.
• This issue does not occur on the Chrome browser.
• The presence of the "media.webspeech.service.endpoint" pref with the "https://dev.speaktome.nonprod.cloudops.mozgcp.net/" value does not seem to influence the reproduction of the issue.