Comment 9 for bug 1792300

Speech synthesis processing is not shipped in Chromium source code.

Ordinarily, in order to use speech synthesis at Chromium a speech synthesis engine needs to be installed, for example, espeak or espeak-ng, and speech-dispatch needs to be installed

  $ sudo apt install python3-speechd

then launch Chromium with

  --enable-speech-dispatcher

flag.

However, no voices are loaded at snap version of Chromium at

Chromium 85.0.4183.102 (Official Build) snap (64-bit)
Revision ffe848af6a5df4fa127e2929331116b7f9f1cb30-refs/branch-heads/4183@{#1770}
OS Linux

Chromium from https://download-chromium.appspot.com/

Chromium 87.0.4268.0 (Developer Build) custom (64-bit)
Revision b2793c4720e3daf929778440ff945eadf202b200-refs/heads/master@{#808630}
OS Linux
JavaScript V8 8.7.142

also has serious issues with Web Speech API implementation.

It should be possible to run

  $ spd-conf

then edit the speech synthesis engine configuration file to modify variations, which should be reflected in speechSynthesis.getVoices(), however, that also does not currently work for non-snap Chromium, where snap Chromium loads no voices at all.

Also, the queue to speech-dispatcher is severely broken at non-snap Chromium, requiring executing speechSynthesis.cancel() to cycle through the code that worked within the last few releases at https://guest271314.github.io/SSMLParser/, where the code outputs the expected result at Firefox 80 and Nightly 82.